Midterm Blood Bath, $4T AI Sell Off, Data Center Freeze, What’s Next for Trump
Chamath PalihapitiyaJason CalacanisDavid SacksDavid Friedberg
- The hosts frame the Democratic sweep as an affordability revolt, but disagree over whether ending the Iran war can reverse it. Sacks links inflation’s rise from 2.4% to 3.7% to fuel and transportation costs, arguing that ending the war could take Brent from as high as $117 toward $70 and reopen the path to rate cuts. Calacanis calls getting out of the war “my number one, my number two, and my number three”; Chamath and Calacanis see a broader economic failure.
- Chamath abandons his belief that asset-led growth is reaching wage earners and makes economic fairness the 2028 electoral test. Asset owners prospered through cheap capital, higher rates, tariffs and subsidies, while paycheck earners remained squeezed: “I have to update my priors.” His prediction is continued electoral reversals against incumbents unless either party delivers meaningful relief.
- Friedberg sees government spending as the structural affordability problem that Washington refused to confront. He points to DOGE’s collapse, a $7 trillion annual budget and $2 trillion deficits, arguing that spending inflates assets while eroding workers’ purchasing power. “The more the government spends, the more expensive things get”; he worries Congress and the sitting president will not reach a clean budget agreement, leaving living costs at risk in 2027–28.
- The reported $4 trillion AI sell-off turns data-center permitting into a major risk for infrastructure spending and IPO expectations. Calacanis says OpenAI’s Polymarket odds of going public fell from 95% to 71%, alongside valuation expectations moving toward roughly $1 trillion. Sacks’s base case preserves meaningful construction spending in 2027, but he worries the 2028 pipeline could close or shrink under “conditions of great uncertainty.”
- Chamath argues that constrained compute could strengthen infrastructure owners while trapping frontier-model companies and neoclouds. Existing hyperscalers could reserve capacity for themselves or charge renters more; he singles out the SpaceX–Anthropic agreement’s mutual 90-day termination right as a potential repricing trigger. Calacanis suggests a negotiated midpoint, Chamath points to Elon’s relationship with Anthropic, and Friedberg says, “This is a spreadsheet and the spreadsheet will dictate the answer.”
- The AI-policy debate divides over how much control government will seek and who regulation would ultimately empower. Sacks assigns a major Democratic legislative push a 100% probability and predicts a bargain exchanging regulatory protection for government power. Friedberg warns about compute becoming a regulated utility; Calacanis puts an extreme state-control scenario around 1%, while Chamath argues supply restrictions would mainly entrench hyperscalers.
- Calacanis predicts a two-year shift toward offshore data centers and local models. He estimates $40,000–$50,000 in annual AI subscriptions creates roughly $2 million of value at his firm, and sees adoption plus agents driving vastly greater demand. He is considering $400,000 of desktop hardware for local-first inference: “We want to have unmetered intelligence.”
- Friedberg’s scientific counterweight is an OpenAI complexity-proof claim that he acknowledges remains unverified. He describes a 175-page paper challenging the exponential time hypothesis, completed in less than 10 days, with some reports claiming hours. He presents it as evidence of AI’s research potential while explicitly saying it is “not fully verified by third parties yet.”
1. The Republican defeat becomes an argument about affordability
- Calacanis reports Democrats taking both chambers and winning at least 29 governor races: 54–56 Senate seats and 250–260 House seats, with Alaska and Mississippi unresolved. Texas, Kansas, Iowa and South Carolina flip after Trump won them by double digits two years earlier. Sacks calls it the biggest midterm defeat since Watergate: “There’s no sugarcoating it.”
- Sacks’s chain runs from war through fuel, transport and fertilizer to lost purchasing power. Gas moves from around $3 to above $4, costing households, he says, $200 more monthly; January’s 2.4% inflation becomes 3.7%, rate-cut hopes become a hike, and mortgages reach roughly 7.5%. His prewar 5% GDP-growth expectation draws “Aggressive. Unbelievable. Unrealistic” from Calacanis. The failure is “this war and the economic consequences that follow from the war.”
- His admission—“I was wrong”—concerns the defeat’s severity. He cites 4.1% unemployment, first-half growth around 2.1%–2.5% and third-quarter GDPNow tracking at 3.7%. He expected something resembling Biden’s inflation-heavy “red spritz” midterms. Instead, inflation eroded the first Trump year’s roughly 3.1% gain in real after-tax household income; voters judged their own purchasing power.
- War also fractures the coalition, Sacks says, costing Republicans Carlson, Rogan, Theo Von and Megyn Kelly: “they are our media.” Sacks cites a 36-point young-male swing; he also cites 34% calling Iran crucial and 40% a major voting factor. JD Vance’s post called for a reset centered on ending the war, tackling living costs, putting America first and changing the relationship with Israel.
2. Friedberg traces the backlash to an abandoned spending overhaul
- Calacanis revisits his post-2024 prediction of rising socialist movements despite deregulation and AI-driven growth: some companies and communities would win, others lose. Friedberg says DOGE promised government efficiency and positive returns on spending. When Musk left on June 5, 2025, Friedberg considered that effort finished and wrote, “China just won.”
- Friedberg’s March 2025 Washington visit supplies the disappointment: Musk was struggling, while members of both parties showed little interest in spending efficiency or government’s share of GDP. July’s One Big Beautiful Bill reduced some entitlement costs without changing fundamental spending. Friedberg’s tally remains a $7 trillion annual budget and $2 trillion deficit; he says Congress was never on board with the necessary overhaul.
- His causal claim is categorical: “The more the government spends, the more expensive things get.” He links COVID’s doubling of spending to asset inflation rewarding owners while squeezing paycheck earners. With this Congress and the sitting president, he worries a clean 2027–28 budget agreement will be difficult; continued $2 trillion deficits could worsen costs. Watch the bond market, he says. Voters seeking relief can produce socialist policies without choosing socialism ideologically.
- AI becomes the beneficiary people resent when growth bypasses them. Friedberg cites Gallup figures of 93% Chinese optimism about national benefits and 93% about daily-life improvements, against 36% American optimism, fourth from bottom—below Afghanistan and slightly above Malawi and Bangladesh. He allows skepticism about Chinese polling. His point is receptivity to an “oligarch narrative”; whether that backlash persists remains uncertain.
3. Chamath rejects the assumption that asset gains reach wage earners
- Chamath says, “I got it totally wrong,” identifying a blind spot rooted in his own interests. Asset owners benefited from cheap capital at zero rates and pricing power at high rates, across tariffs and subsidies; wage earners lacked that resilience. “I thought trickle-down economics could still work.” His update: “It’s clearly not happening. It’s not true. I have to update my priors.”
- His evidence against an ideological reading is failed spending and gerrymandering. Money poured into Texas and Ohio without delivering the intended wins; of 15 Republican-redrawn seats, he says Democrats won 11 and Republicans only one. He sees centrist affordability candidates prevailing, explicitly disputes Friedberg’s socialism framing and says Iran, Israel and AI were not the central explanation.
- Since COVID, Chamath sees a swing bloc granting mandates to Biden, Trump and then Democrats, demanding economic repairs. His 2028 call is economic populism built on “reasonable standards of economic fairness,” from either party. Incumbents who do nothing will keep losing. Democrats must now legislate that fairness or concentrate on impeaching Trump and risk reversal if Republicans supply the economic agenda.
4. Calacanis challenges the war-only explanation with an earlier decline
- Calacanis points to Trump’s net approval falling from roughly +5% to −28%, with deterioration predating Iran. He says polls and prediction markets were right. Liberation Day’s tariffs raised costs without balancing the budget; he recalls speaking with the Trump-supporting CEO of Cuts Clothing during a discussion of the tariffs. Trump should have worked through Congress rather than pursuing the route that lost at the Supreme Court.
- Summer 2025 brings Calacanis’s next breaks: ICE enforcement by masked agents without visible badges or body cameras, and Trump reneging on sending out the Epstein files, he argues. Those episodes alienated podcast audiences before the war; grifting allegations and inflation deepened the damage. “Trump didn’t make people feel like he was working for them.” He credits the ideas while criticizing chaotic execution.
- Sacks counters with Senate-control markets: Republican odds remained strong through 2025, then deteriorated after the war began on February 28, 2026. Brent’s 10% election-day fall, without a change in the Strait of Hormuz, illustrates his recovery mechanism. If the war ends, he expects crude—which has been as high as $117—to fall toward $70. Lower inflation could permit rate cuts and restore Trump’s approval into the 40s.
- Calacanis uses 2024’s egg-price controversy as an example: eggs were “a window into a broad problem,” and gasoline can serve the same purpose. Fixing one visible price may miss structural hardship. Sacks agrees affordability is central but sees healthy employment and growth as grounds for recovery without restructuring the economy. If the war continues, however, he expects Republicans to keep bleeding support.
5. Trump’s proposed reset combines economic bargains with an end to war
- Calacanis proposes a controversial reset: pardon Trump’s family, Trump himself and administration officials immediately, then tell Democrats, “Let’s work. Let’s make a deal.” He wants equity-backed loans above $10 million taxed at 50% of the capital-gains rate, credit for repayment and the balance collected when shares sell. The aim is revenue and an answer to claims that wealthy borrowers pay no tax.
- His remaining proposals cover healthcare, a phased $15 minimum wage over four years, banning stock trading in office and business with foreign countries, and incentivizing housing construction. He estimates 10 million–15 million workers earn below $15, making the wage measure manageable and a signal that the bottom half matters. Solving two or three of those five issues, alongside ending the war, could redeem the presidency.
- Chamath first suggests zero federal taxes below “probably $150,000,” later proposing $125,000 or under; he calls the revenue loss “a nothing burger” and would close loopholes. He pairs ending federal underwriting of loans with getting rid of student debt. Preserve border closure through an economic bargain, he argues. Tariffs remain conditional: if promised investment materializes and revives domestic manufacturing, their historical record could improve as gains compound.
- Sacks contrasts course correction with escalation. Clinton lost the House in 1994 and won in 1996; Obama lost seven Senate and 63 House seats in 2010 and won in 2012. Bush lost in 2006 over Iraq, chose the surge, and Republicans lost again in 2008. Sacks sees renewed escalation repeating that failure; Calacanis calls ending the war “my number one, my number two, and my number three.”
6. The AI sell-off exposes a dispute over local benefits
- Calacanis describes nearly $4 trillion erased across Nvidia, SpaceX, Tesla, AMD and Micron: a dollar record, explicitly not a percentage record, followed by a partial rebound. Polymarket’s OpenAI IPO odds fall from 95% to 71%; he relays Wall Street Journal valuation expectations near $1 trillion versus $1.5 trillion–$2 trillion and earlier $3 trillion–$4 trillion talk. Anthropic’s IPO faces a darker backdrop. He has been buying SpaceX and Tesla “through this whole insanity.”
- Sacks’s Texas example connects electoral economics to infrastructure: the incoming governor rejects culture-war questions and links data centers to utility bills and water use. Calacanis adds that the oligarchs are getting tax breaks while residents get higher bills. Chamath concedes the bargaining problem: should seven or eight powerful companies need tax relief for $30 billion–$50 billion investments generating $345 billion–$500 billion? Incoming governors may demand directly measurable benefits before agreeing.
- Chamath’s illustration is SpaceX’s proposed 6-gigawatt Arizona facility. Presenters Riley and Katie promise electricity costs down 40%, property taxes down 20%, no water use and revenue to retire a county bond, yet the audience shouts over them. He calls it “an emotional issue that is not tethered to the facts”: economic division that technical and fiscal promises alone fail to defuse.
7. Scarce compute could transfer leverage from model companies to owners
- Chamath’s call favors hyperscalers with existing capacity—Amazon, Meta, Microsoft, Google and SpaceX—over neocloud renters and compute-starved frontier labs. Anthropic and OpenAI improve on 60–90-day loops, but lack the combined fortress balance sheets, profitability and quasi-infinite compute that would make them “home free.” His warning is that the data-center renter class is “very much at risk.”
- Owners can keep renting or reserve incremental capacity for themselves. Google could prioritize Gemini while rivals face higher prices, deteriorating user experience and slower iteration. Chamath points to Google taking over a renter’s data center; Friedberg says the hyperscalers have critical ingredients, specifically power. Calacanis adds that concentrating compute on training could determine who produces the best next model.
- Chamath watches the SpaceX–Anthropic agreement’s mutual 90-day termination right. If SpaceX exits and reprices scarce capacity, he argues, that revenue could rise two or three times and Anthropic would have little choice but to pay. Other hyperscalers would reassess yields; neoclouds could spend remaining debt and equity securing whatever compute they can, because being shut out risks bankruptcy.
- Calacanis suggests Musk could negotiate a middle ground. Chamath points to Elon’s close working relationship with Anthropic; Sacks cites acrimony between Elon and Sam Altman and argues that relationships will inform the decision. Friedberg says, “This is a spreadsheet, and the spreadsheet will dictate the answer.” Sacks’s point is that if constraints push GPU-hour and megawatt-month prices sufficiently high, even long-term offtake can become negative expected value. Capacity could go to Google or Meta; the argument does not require selling it to OpenAI.
8. AI regulation divides the hosts over state power and practical incentives
- Friedberg calls AI “the Ring of Sauron”: both political sides want its power. Even a million data centers running open-source models could weaken centralized authority without an oligarch. His prediction, allowing “I could be wrong,” is regulation of compute and treatment of data centers like utilities. Chips and software are harder to police; compute access offers a practical control point.
- Sacks frames a model as code and software, making regulation a free-speech question. He argues that existing laws cover harmful uses, such as theft and illegal weapons. His Second Amendment analogy concerns unequal capability: a state with more powerful AI gains power without the same check. He says the concern is that the state might have access to capabilities that citizens and businesses do not.
- Sacks gives a major Democratic AI-legislation push a 100% probability. His forecast for its shape is a federal AI department preapproving frontier models, with misinformation rules becoming censorship. The quid pro quo gives government more power and the closed-model duopoly regulatory protection. He predicts surveillance and censorship enhanced by superintelligence; this is his interpretation of regulation’s destination.
- Calacanis first floats 10% as a handicap for an extreme scenario, then says the chance of something like it is about 1%. Friedberg says the concern is not binary, but gives no probability. Chamath cites California legislation vetoed by Gavin Newsom and a patchwork of state laws. Chamath disputes the priority and political logic of nationalization: entrenching an oligopoly contradicts voters’ economic demands. Infrastructure restrictions, he argues, ultimately empower hyperscalers, because model companies cannot obtain protection from shortages of electricity and money merely through regulatory capture.
9. Permitting uncertainty threatens the construction pipeline after 2027
- Sacks separates federal AI-safety rules from state infrastructure permits. Texas’s incoming governor has said no new permits will be issued before legislation, Sacks says, without unequivocally protecting existing construction. A partly built facility may still need air-pollution or other late approvals. He allows reasonable or unreasonable implementation: “I’m not saying that’s necessarily going to happen,” but stranded half-finished projects are possible.
- His base case preserves meaningful capex in 2027 as fully permitted projects finish; by 2028, the pipeline could close or shrink substantially. Both parties attacked politicians for approving data centers, leaving ambitious officials little incentive to approve more over the next four years. Gigawatt-scale projects costing around $50 billion require certainty before developers risk injunctions, labor action or community opposition halfway through construction.
- Sacks says a slowdown need not be a conspiracy, but believes Democrats could cause the damage and blame Trump. Calacanis sees an opposing incentive: tanking growth creates an electoral vulnerability, whereas negotiating electricity, schools and local benefits rewards constituents. Their disagreement is whether governors chiefly want better terms or permitting politics itself blocks capacity.
10. Calacanis expects demand to escape through offshore and local compute
- At his 20-person venture firm, Calacanis estimates annual AI subscriptions of $40,000–$50,000, mainly Grok and Claude, create roughly $2 million in value and about $1 million in incremental revenue. Automating investment workflows and media production increases the firm’s operating velocity.
- He estimates only 10%–20% of corporate workers use AI daily, leaving five to ten times more potential users. Agents abstract away coding and setup while using 50 times as much compute. His resulting demand call is 500–1,000 times today’s need.
- His firm is considering $20,000 Mac Studios with 512 GB, totaling $400,000 for 20 people, running local models before falling back to frontier services. He says their current use is only 70%–80% of what frontier models can do. The objective: “We want to have unmetered intelligence.”
- His two-year prediction is offshore data centers plus desktop models. Canada, Mexico, Singapore and Malaysia can host capacity; ending the war could reopen Saudi Arabia, the UAE and Qatar. Fiber and Starlink support serving users across borders, he argues. America could become a net token importer, losing local tax revenue and construction jobs he describes as paying $300,000–$400,000.
11. Friedberg’s scientific optimism rests on a proof still under review
- Friedberg starts with solving a problem quickly versus checking its answer quickly. Sudoku is his example: finding the solution can be hard while checking it is easy. He uses that distinction to introduce P versus NP and the exponential time hypothesis, proposed in 1999 for 3-SAT.
- He says a 175-page OpenAI paper provides a subexponential-time approach and “disproved” that hypothesis. The work took less than 10 days end to end, with some reports claiming hours, against a 27-year-old question. He qualifies the claim: “It’s not fully verified by third parties yet.” Others are still evaluating its novel approach.
- Friedberg interprets the result as keeping P = NP possible and, in his framing, the prospect that every problem could be solved by computer. He calls it “one of these moments of humility” for mathematicians. Sacks returns to his broader wager: AI could deliver enormous benefits, and China will pursue them if America does not.
Full transcript
All right, everybody. Welcome back to the number one podcast in the world. It’s the All-In podcast. I’d like to start by saying that All-In recognizes we’re on the sacred land in Austin, Texas, of the Coahuiltecan, the Comanche, the Kickapoo, and the Tonkawa. We appreciate their sacrifices in bringing us here.
I just talked to our HR department. We’ll be evaluating a fifth bestie, Brad: too white, too male. So we’re looking for alternative DEI fifth besties to include in the round table.
Wow.
Some big changes coming to the—
DEI is back, is what you’re saying.
DEI is back. Sacks, do you have anything for the docket this week?
I have an announcement.
You have an announcement. Okay.
I’m transitioning from a douchebag to an—
Let your winners ride. Rain Man David.
We open-sourced it to the fans, and they’ve just gone crazy with it. Love you.
Sacks, where do you want to start? Do you want to start with—
I think there’s some really important news around SaaS out there. Maybe Salesforce launched a couple of new features. Definitely top-of-the-docket material.
I’d like to start with Science Corner, words nobody in our fan base has ever said.
I’m ready to go. You guys want to talk about the Exponential Time Hypothesis?
I think that’s what the audience wants. They want the SaaS update from Sacks and the science update, and then we can get our market review and wrap the show up.
No, in all seriousness, there was an election this week. We had an amazing livestream, with over a million views so far. Crazy, Chamath. I don’t know if you knew this, but 400,000 people tuned in live, making us, I believe, the number one election-night livestream.
Not true. It was 500,000. Just to put it in perspective, the far-right streamers got about 87,000. The far-left streamers got 250,000. We got 500,000, which was bigger than most cable, as well. We were the biggest—bigger than, I think, most, if not all, cable, as well. Apparently, America wants hot takes from tech bros. Who would have thought?
Who would have thought it? And, by the way, that was with only 3 of our All-Stars on the court. Sacks had a board meeting Thursday night and couldn’t make it Tuesday night. He was busy.
Yeah, I was busy that night. I think I was passed out at the bar.
Sacks is the only guy who goes to the bar with his own bottle of tequila. He’s like, “Can you put this behind the bar and serve it to me at $50 a glass?”
Special thanks to Ted Cruz, Rahm Emanuel, and Shayne Coplan of Polymarket. Man, did he have a great night. I think Polymarket nailed it. 99% of their prediction markets turned out to have it perfectly. The hilarious fifth bestie, Tim Dillon; Taylor Lorenz—based. Taylor Lorenz was fantastic and insightful. Brad Gerstner did a solid job and brought some charts, and Ben Shapiro from the right.
1. Midterms postmortem: What led to a generational GOP collapse?
It was a really great panel, and today we’re going to take a pretty systematic approach to what happened this week. We’re going to start with a postmortem on the midterms, and then we’re going to talk about the second half of Trump’s term, the next 2 years. I’m sure data centers will come up in our recap and postmortem, but we’re also going to do a deep dive into the AI trade potentially cracking.
Let’s start with our postmortem on what just happened in the midterms. Obviously, the Democrats swept both chambers of Congress, and they won at least 29 governor races. The Democrats will wind up with 54, 55, or 56 seats, depending on how Alaska and Mississippi wind up. They have a lot of votes still coming in in the House. The Democrats will have at least 250 seats, maybe even 260.
The bigger story is where the Republicans lost deeply red states that haven’t gone blue in generations, including the great state of Texas. Please don’t blame me. I’ve only been here for 3 years, and you might be surprised at what my voting ticket looked like. Kansas, Iowa, and South Carolina went blue for the first time in decades. Trump won all of those states by double digits just 2 years ago. The GOP lost ground in every voter group, even white men.
So, let’s go around the horn. Sacks, people want to hear from you most of all, I guess. What went wrong here? What are the lessons the Republican Party can learn? And what did the Democrats do right in the last 18 months leading up to the midterms?
This was shocking. This was a blowout. It was, I believe, the biggest midterm loss since the Watergate midterm loss. This was obviously a big defeat for the Republicans. There’s no sugarcoating it.
What’s the reason for that? I believe the vice president, JD Vance, explained it very well: it was this war. The only thing I would add to that is, it was this war and the economic consequences that followed from the war. What am I talking about? The price of gas was around $3 before the war; it went over $4 and higher. The price of diesel went even higher. That factors into the cost of food and all the goods that have to be transported. Fertilizer costs went up, too.
You saw the inflation rate at the beginning of the year, back in January before the war, down to 2.4%. We finally had that problem licked, just about. Remember, it was 9% under Biden. In the first year of the Trump administration, we brought it down to 2.4%. The market was expecting multiple rate cuts this year, and we were forecasting 5% GDP growth, which I don’t think was overly bullish based on what we were seeing at that time, combined with the expectation of rate cuts.
Aggressive. Unbelievable. Unrealistic.
Not unrealistic.
Yes. Okay.
By the way, the prediction markets back in January were showing a 70% chance of Republicans holding the Senate, and the House was perceived to be a toss-up. I think that if this election had happened back in January, it would have been a pretty successful one for the Republicans.
What happened since then? Well, you got inflation going from 2.4% to 3.7%, again, mostly driven by the cost of gas and diesel and all these other products where you have to pay a higher cost of transportation. You end up not having any rate cuts. You end up getting a rate hike that brought the cost of mortgages to roughly 7.5%. That overall fed this narrative around affordability.
It’s the war plus the perception of the lack of affordability that really created this tidal wave. I have to admit, I thought—I didn’t think this would be a great election for Republicans, but I thought it wouldn’t be this bad. I was wrong.
You said it was a spritz. It would be a spritz.
I thought it could be a spritz. That was based on a couple of things. Number one is that the last few elections, the pollsters have all been wrong, and Trump has always managed to defy the odds. I thought that this polling data must fall into that pattern.
The second thing is that the economy is not that bad. In fact, it’s quite good. Just today, we had a new jobs report. It was a very robust jobs report. The number of new jobs created is being revised up. You’ve got unemployment ticking down to almost a record-low 4.1%. You’ve still got pretty healthy GDP growth: in the first half of the year, it was around 2.1% to 2.5%, and in the third quarter, the GDPNow tracking is 3.7%.
Given that you’ve got great employment and solid GDP growth, I thought the fact that inflation was too high wouldn’t result in this type of tidal wave. I thought it’d be more like the Biden midterms, where he had extremely high inflation, but the economy itself was still pretty decent, and it ended up being more of a red spritz as opposed to a red wave. So that was my pattern recognition. It was wrong.
I think one of the reasons it was wrong here is that the public is in a pretty ornery mood. They’re very concerned about affordability. It doesn’t matter how good the economy actually is; it matters how good people perceive it to be for them. In the first year of the Trump administration, we made enormous progress in delivering real post-tax household income. It rose by about 3.1%. But because of the inflation numbers over the last 6 months, you’re starting to chip away at that, right? As long as inflation is growing faster than your wages, you’re losing purchasing power in real terms. I think the electorate was getting ornery about that.
And just one other thing I want to say about it: there’s one other reason the war was so bad, which is that it fractured the Republican coalition and the Republican base. J-Cal, you’ve pointed this out. We lost all the podcast bros. Obviously, it was Tucker Carlson, Joe Rogan, Theo Von, and that whole set, Megyn Kelly.
90% left.
And I think one of the reasons Trump produced such a resounding victory in 2024 is that he appealed to that whole podcasting circuit. Frankly, we should have done a better job giving them all front-row seats in the White House press room. We should have stayed connected to all of them, because they are our media.
By the way, they don’t work for us. They’re willing to defect. They have their own opinions.
They have their own opinions.
But we have to pay attention. I think our side has to pay a lot of attention to them, because without them we don’t have any media that supports us. I think losing them was a big deal. It fueled the sense—and I think a lot of those people showed up and did a protest vote because they wanted to send the message that they want out of this war.
Friedberg, just a couple of quotes for you to react to. Ted Cruz called it a Watergate-level repudiation. Chamath Palihapitiya called it “close to an extinction-level event for the current GOP.” JD Vance wrote a very long post, which was pretty surprising to everybody. There’s some reporting that it was approved, maybe not approved; we’ll get all that out later. What’s important is what he said: he’s in line with you that the Iran war was the driving force behind the loss, and that the GOP has to rebuild trust with voters on 4 fronts: ending the war, focusing on the cost of living, putting America first, and, maybe a little controversially, changing our relationship with Israel. What are your thoughts on JD’s post and the other reactions? And why did this turn out to be a tsunami?
I don’t know if it’s as much of a positioning and marketing issue as it is something fundamental. I’ll go back to the clip from the prediction show right after Trump was elected.
Okay.
Can we pull that up, Nick?
I think the party line is that socialism was defeated in this election cycle, and that there was a resounding vote from the American populace against socialism. My contrarian belief is that we’ll see a dramatic rise in socialist movements in 2025 in the United States.
Okay.
And I think we’re going to see, as Gavin pointed out, an acceleration of progress. We’re going to see an unleashing of economic growth because of deregulation and AI. But I do think some industries and companies are going to be huge winners, and some companies are going to be huge losers. There are going to be parts of the economy that are big winners and parts that are big losers.
I just wanted to dial it there, Friedberg. I was very accurate for the first year of the presidency.
If you zoom out a little bit, the macro condition was that, after the election, there was this idea of DOGE: that government spending would be reduced, which I believe fundamentally would have affected the cost of living. We can argue about that; people disagree and dismiss that point. I do think government spending is the fundamental driver of the cost of living. The more the government spends, the more expensive things get. We can go through the analysis on what happened in France. Government spending as a percentage of GDP is one metric; how many people are employed by or depend on government checks is another. But as the government gets involved in more of the economy, things get more expensive.
When DOGE was coming around, the idea of government efficiency and making sure that every dollar the government spent had a positive ROI for the people was a great idea. We visited D.C. in March 2025 and met with Elon. We talked about DOGE, and he was struggling—not just internally, but with Congress and with the public perception of what was being done there. I came away from that visit, where I also met with multiple members of Congress, Republicans and Democrats, who didn’t give a shit about improving government spending efficiency or about total government spending as a percentage of GDP. It wasn’t a priority. On June 5, 2025, Elon left, and that was the end of DOGE.
The idea that DOGE was going to continue was over. I said, “China just won,” because, to me, the day this administration ended that effort was the day it became clear that we were going to have a runaway cost of living and a challenge for progress for the majority of Americans. The end result is people voting for more socialist principles, socialist policies, and socialist politicians. My tweet on June 5 that China had just won was posted the day it was announced that DOGE was basically coming to an end.
Then the One Big Beautiful Bill passed in July 2025. It had some entitlement cost reductions, but government spending hadn’t been fundamentally reduced. We’re still running a $7 trillion-a-year government budget with a $2 trillion-a-year deficit. I don’t think Congress was ever on board. It’s a structural, fundamental problem, and no one’s on board with fixing it. I think that was key to what happened here.
For the majority of Americans, they don’t feel like they’re getting ahead. They don’t feel like they can afford things. They don’t feel like health care is accessible, and they’re struggling. It wasn’t addressed in the way it should have been. We can all disagree on policy matters, but that’s my view. As a result, people still felt that way, and that’s what drove the vote.
On AI, I think there’s also this idea of an oligarchy: if I’m not getting ahead, someone else is, because “the economy is growing” and “the stock market is growing.” Who’s getting all that growth? I’m not. The AI guys are, the tech oligarchs are, and so on. I think that made AI a target. But I also think there was some external influence on AI. I’ll pull up this chart I shared in the group chat yesterday: an incredible Gallup study showing AI optimism and pessimism in countries around the world.
China is the most optimistic country about AI. Ninety-three percent of Chinese citizens think AI will mostly help people in their country, and 93% think AI will make their daily lives better. We can believe or not believe how Gallup does its surveys in China. The U.S. is fourth from the bottom. We’re worse than Afghanistan and only slightly better than Malawi and Bangladesh. We’re at 36% of Americans believing that AI will help.
I think the idea that if I’m not getting ahead, someone else is—and that the AI guys, because we create AI in this country, are getting ahead—made it really easy to attack with this oligarch narrative. There’s a group of people benefiting and making a lot of money, and then there’s a large percentage of people who aren’t. I think that vector was really effective. Whether it persists is TBD, and I think we should talk a little about what’s next later in the show, because AI is going to play a major role in how politics play out over the next year.
Those are 2 incredibly concise and candid recaps of what happened there. Chamath, as a side programming note, great job on Rick Rubin’s podcast this week. If you haven’t checked it out, Chamath was on Rick Rubin’s podcast, and it was awesome. Thank you, Chamath.
Did you sit barefoot in his backyard with him, Chamath, or—
Yeah. By the way, when you go to Rick Rubin’s recording studio, he walks you around and says, “Oh, Chamath, here’s Bob Dylan’s old touring van. That’s where we recorded Adele’s album.” Then he says, “This is where Kanye and I laid down this album. And here’s where Jay-Z and I…” It is the craziest place, but it puts you in a totally centered state of mind, I’ve got to say. We sat around in his studio barefoot, sitting across from each other, just talking. Then they put it out.
We had the greatest stock market highs of our lifetimes and the lowest unemployment of our lifetimes, and we still had the tsunami. Chamath, when you look back on it, what do you think happened here? You said you were in a moment of great candor and self-reflection during the livestream. You said, “Hey, listen, I had a blind spot here.”
I got it totally wrong.
Okay, unpack.
I’ll go through all the ways I got it wrong, but I want to first acknowledge what my blind spot is. Like most people’s blind spots, I can speak only for myself: it’s seeing things through the lens of my own lived experience and, very narrowly, what’s in my best interests. I think we had a move away from the culture wars. That was very much in my best interest, but I think it was in everybody else’s best interest as well. I don’t think that was on the ballot.
What I don’t think I understood clearly enough, and what I think this election has made extremely clear, is that since COVID, we’ve moved past political ideology. I want to say this again: I do not think Democrats or Republicans matter. That’s my real takeaway after this election. There’s a central bloc of people who vote on the predominant issues of our time. It was a gargantuan win for Biden. Then it was a gargantuan win for Trump. And then it was a gargantuan win for the Democrats in these midterms.
What are these people voting for? I think what they’re shouting now—and what they were trying to signal, and are now shouting—is a broad-based belief that the economy has completely decoupled from reality and is broken. The economic reality in America is broken. There’s a small percentage of people—the 4 of us on this pod and all our friends—who are the asset owners of America. We’ve benefited in every weather condition.
When rates were zero, we benefited because we could build our businesses with very cheap capital. When rates were high, we benefited because we had pricing power. In every condition in between, whether there were tariffs or not, whether there were subsidies or not, we’ve managed to continue accumulating and growing assets. But we are few and far between. Then there’s everybody else, who are wage earners, and the structural capacity of the American economy to support and grow wage earners is broken.
It’s clear. Why do I say that? I don’t think—and I disagree with Friedberg—that this has anything to do with socialism. I don’t think the far left won particularly more than they were expected to, and I don’t think the far right won particularly more than was expected. This is a complete and total repudiation of the state of the economy. The economy is broken.
I have a few charts to support my point of view. Number one: if this was ideological, and ideology mattered, money in politics would have had a more direct impact on the election. But when you look at the amount of money that was spent, all of this money was wasted. All the money that went into Texas was wasted. All the money that went into Senate races in Ohio was wasted. The better bet was to look at who money was being spent on. If you had taken the opposite candidate, who had no money, you would have been overwhelmingly more right. So money follows ideology. Money did not work in this election. Ergo, in my opinion, ideology did not matter that much.
Next chart. Redistricting and gerrymandering also follow ideology. You try to set up the geographies in a way that lets you gain more seats. Look what happened: of the 15 Republican seats that were redrawn so they could capture all of them, Democrats won 11. Republicans only won one. So again, ideological gerrymandering also totally failed. We have ideological spending in politics failing, and ideological gerrymandering failing.
Next chart. What you had were centrists, normal people. These were not people running around talking about communism. They were people who focused very narrowly on one thing and one thing only: affordability and the functioning of the United States economy. They spoke to a large bloc of people who were both Democrats and Republicans, initially supported Biden, overwhelmingly supported Trump, and then overwhelmingly went to the Democrats, largely to vote their displeasure with the economy.
That’s it. I don’t believe Iran really had much to do with it. I don’t believe Israel had much to do with it. I don’t believe AI had much to do with it. This is American society asking for a reset on the economic principles on which we operate. I think the obvious thing for the ambitious candidate on the left or the right to do is take these three charts together and have one very precise takeaway: the economic populist platform will be the one you can take to the presidency in 2028.
The path is really straightforward. If you look at what happened to Harris and then at how aggressively all of these states moved toward what I think is economic centrism, what you have is this: neither side needs to worry about whether they’re Democrats or Republicans. If you embrace reasonable standards of economic fairness, you will win.
So, I got this wrong. I thought trickle-down economics could still work. I thought that when asset owners built assets and invested in the economy, those things could get to everybody else. It’s clearly not happening. It’s not true. I have to update my priors. My new framework is that we need to figure out how to stop all the loopholes. We need to figure out how to do some of the things we’ve talked about before: stop the federal underwriting of loans, combine that with getting rid of student debt, and take all federal taxes beneath a certain band—probably $150,000—to zero. There are all kinds of things that can be done now to allow the wage-earning class to feel more ownership of the United States economy.
Until it happens, my prediction is not that Democrats or Republicans will win, but that they will violently keep going back and forth away from the status quo. If you are in power and you do nothing, they will vote against you. So now the Democrats have a really interesting challenge: what will you do with the majority in both houses? Are you actually going to try to legislate economic fairness, or are you going to spend your time distracted with impeaching Trump? Because then, if the Republican candidate embraces the economic populist platform, you’ll find yourself losing the way the Republicans lost in the midterms.
All right. Incredibly well explained there, Chamath, and I agree. Now is not the time to dunk. Now’s the time to dissect all this. Who could have ever seen this coming? Nick, roll the tape.
This is why they will lose the midterms. The Democrats are going to shellack the Republicans in the midterms. Trump sold people that prices would go down. We’re back up to 3% inflation, folks. Not only is it not going down, it’s going up. Trump has failed the middle class. That’s the bottom line here. It’s basically going to result in the Democrats doing a clean sweep in the midterms. Trump promised not to start a war. He did. He promised to solve inflation. He didn’t. The end.
Well, a stopped clock is right twice a day, J-Cal.
Absolutely. I did cheat out, but let me—
My take on that is that you were not right until the war began. That last clip that you did, you were vindicated, right? Because it’s the war and inflation. Those are the two things, specifically the affordability problem that results when you have this inflation.
Look, the price of gas: everyone goes to the pump to fill up every week or two, and the average household is paying $200 a month more in gas costs. Then they go to their grocery store, and everything is much more expensive because the cost of agricultural products has gone way up, and so on down the line. At the end of the day, what matters to people is their pocketbook. Do you have more purchasing power in real, post-tax, post-inflation terms? That’s how you measure if you’re better off.
The Trump administration delivered that in the first year. We had a 3.1% increase in real household income. It was negative during the Biden years and positive during Trump 45. What happens? The war begins. The price of gas and diesel and all the rest goes up, and purchasing power starts to go in reverse. On top of it, you lose the young males, who are again the fans of all these podcasts. That’s where they get their news from. Young males flipped their vote by 36 points in this midterm relative to the 2024 election.
So again, in my view, it’s very simple. It comes back to the war and its economic consequences. Now, here’s the good news: if we get out of the war, we can make rapid progress on this. All the other economic numbers are still good. Job growth—the job report today was great. Economic growth is still good. There are lots of very positive economic numbers. The issue is fundamentally the price of oil, the price of gas, the price of diesel, and the way these things flow through the economy. If the war ends quickly, those prices will drop very quickly. Inflation will come back down. That will give the Fed permission to cut rates so mortgage costs come back down, and we can get back on track.
2. Market reaction: Data center fears, $4T in market cap shed, AI ban potential
Sacks, you’re right. The endless war and endless inflation were the straw on the camel’s back. But I think you’re absolutely wrong when it comes to what happened previously. All of the polls were correct. Everybody on the right said the polls were wrong. All the prediction markets were correct, and everybody on the right was saying that those were wrong and were being manipulated.
For a year-plus on this podcast, I pointed out this chart. The Silver Bulletin absolutely nailed this cycle. They did a great job. Nate Silver is phenomenal at what he does. If you look at it, it’s really worth looking at. It’s not just for Liberation Day. People forget: Trump started the term at a 5% plus net approval rating. He ended it at -28. This is like an over-30-point swing.
Liberation Day, as I pointed out when it happened, was far too chaotic. I spoke to the CEO of Cuts Clothing, if you remember; we had this discussion on the podcast, and he’s a Trump loyalist. People were arguing at that time, “These tariffs are going to be great. These are awesome, and no prices would be added. The American public didn’t have to pay for it, and this was going to balance the budget.” Not only did it not balance the budget, it did spike inflation. President Trump should have just worked with Congress. Remember, he had to go to the Supreme Court. He lost there, and this is a weak link in President Trump’s game.
Then over the summer, we had some pretty hot discussions here about ICE enforcement. I said, “Listen, no Americans want to see this—not people I know in law enforcement, and not people I know who employ people who have been here for 20 years and, yes, are technically illegal aliens. Nobody wants to see masked agents without body cameras beating people up and without their badges on.” That was a huge mistake.
He had a massive drop and then started to hit a -20% net rating. This is all before the Iran war started. In the summer of 2025, with those ICE actions, there were also the Epstein files. That’s where you lost all the podcast bros. They did not want to have Trump renege on sending the Epstein files out. And, of course, you get all the way to the war, the endless allegations of grifting, and the endless inflation, and that then just plummeted from 20 to -23.
I just want to state for the record here: I’m an independent. I vote 55% Democrat in my life, 45% Republican. When we started this podcast, it was probably 70/30. I have moved significantly toward the right because of the lunacy of the left. But what happened here is that the Trump 2.0 agenda, I think, got abandoned, and there were tons of distractions. President Trump has the right ideas, but often his execution is extremely variable. That’s just a very generous way of saying it’s chaotic.
If you have that chaotic moment and then everything is about the data centers, everything is about our industry, everything is about helping us, and people laugh and mock the idea of raising the minimum wage, it doesn’t feel fair to people. The K-shaped economy is what you were invoking there. I think you made a very accurate point: people are swinging and hoping somebody will care about the bottom half of non-equity holders.
That’s exactly what it is. Trump didn’t make people feel like he was working for them. That’s why they abandoned him. The grifting allegations, whether they’re true or not, are all going to come out. But that’s what we need to talk about.
3. How can Trump salvage his presidency?
That leads us to our second topic: what the next 24 months are going to be about. Sacks, you wanted to get to the next 18 to 24 months. Is there a path to redemption for President Trump? Is there a way to turn this presidency around and maybe save his legacy?
First of all, let me put up some charts, since you guys are doing the chart game and I didn’t know we were going to play the chart game today. Nick, can you throw up these charts?
The best chart?
Yeah, Nick, can you throw up these charts? Okay, look, JC, there are all sorts of issues out there that people are going to point to. I know you’ve got yours, and I’m not saying that every single issue has been executed perfectly. I do think President Trump deserves enormous credit for completely sealing the border and solving that immigration issue, just as an example.
I agree with that.
But look, again, the question is: Why was this not a normal midterm result? Look at Trump’s approval rating versus the price of gas. It maps perfectly. Trump’s net approval maps to the price of gas.
The next chart: This was the prediction market for which party will win the Senate in 2026. What you see is that Republicans were expected to win by a very large margin throughout the entire first year of the Trump presidency, through all the things you were talking about, J-Cal. The tariff rollout did not affect this whatsoever. Liberation Day was in April. Not even a blip here.
And to be clear, not a blip on the Senate balance.
Yeah, yeah.
I get it. But I’m using this as a proxy for what just happened, where we lost, what, 8 or 9 Senate seats?
Yes, but I’m using this approval rating I showed, right? Where does it start to change? March of 2026. What happened? When did the war begin? February 28, 2026. Very clear. That’s where everything starts to change in terms of losing the Senate.
Now just ask voters: Why did you vote? Thinking about your vote for Congress, how important to you was the war in Iran? “Crucial,” 34%; “a major factor,” 40%. Basically, 74% of the electorate said it was either crucial or a major factor. And of those, they broke massively for blue.
This doesn’t just include Democrats, obviously. It includes a lot of independents, and even, I would say, younger Republican voters. There’s a huge divide in the party. You’ve got the Fox News group, kind of the over-50 or even over-60 crowd. I think they’re still supportive of the war. But younger voters, young males, are highly against it. Again, it ties back to the podcast bros. There’s abundant data to show what I’ve been saying.
Now, your question was, can Trump fix this? I think absolutely. The top 3 things he can do to fix it—
Just get out of the war.
Okay. End the war.
That’s my number one, my number two, and my number three.
Because what will happen? The price of Brent crude will drop precipitously. On the day of the election, the price of Brent crude dropped 10%. Is that because something changed in the Strait of Hormuz? No. It’s because the market started to price in the mere possibility that the war would end sooner, and we got a 10% cut in the price of Brent crude.
It’ll go from, you know, it’s been as high as $117, to $70 if this war ends. Gas prices will drop, diesel will drop, the price of food will drop, and then the underlying strong economic growth, the GDP growth and the employment picture, which are still robust, will come into focus. I think Trump’s popularity will rebound, maybe back into the 40s.
Okay.
That’s the thing we need to do. And if we don’t do that—look, if this war continues, hold on, let me finish this one point. If the war continues, the Republicans will continue to bleed popularity.
I don’t think this is just the war in Iran. It is, inasmuch as it adds to a broken economy. Since COVID, the economy has been structurally broken, and I just think there’s a huge voting bloc that keeps swinging back and forth to either side, asking them to solve the problems and giving people mandates to solve a problem. That’s why I think somebody needs to understand economic populism. Do you buy that argument, or do you think it’s more narrowly just the war in Iran?
I don’t want to make it overly abstract. To me, the issue is affordability, and I think the electorate really cares about that. It’s the central pocketbook issue.
I agree.
You remember the 2024 election. I agree with you that economic affordability can be an abstract term, but I think the voting electorate is smart enough to use a message as a vessel. I remember in ’24, the vessel—which, again, I didn’t hear it loud enough, but if I look back, maybe I’m just imposing my narrative bias on it—was the whole controversy around the price of eggs. Trump hammered Biden and Harris on the price of eggs.
Yes, over and over. Yeah.
It wasn’t the price of eggs. I mean, yes, it was, but it was a window into a broad problem and an idea. I think gas is a problem, but I also think it’s a window into a broad class of problems. People focus their energy on something because it’s easier to message around repeatedly.
But the data centers—that’s what—look, I was listening to Gina Hinojosa. She’s the new governor-elect of Texas. Frankly, it’s a race I wasn’t even paying attention to because I didn’t think anyone was going to win. A very successful three-term governor in Texas presided over an economic boom. That’s why we moved to this state, and we thought—I mean, honestly, I think it was the biggest shocker of the whole night.
Anyway, I started watching some interviews with her to catch up on what I had obviously missed. I watched an interview she did with the New York Times where the interviewer kept trying to draw her out on social issues, on what you might call woke issues, identity politics, and she said, “Listen, I’m not participating in your culture war.” Then she laid out the case for economic populism as she saw it. She connected it to data centers: the reason data centers are bad is that they’re making your utility bills go up, right? You’re paying more for electricity, and the data centers are using up your water.
I think that’s why the data center issue was resonant. There is anxiety about AI and the potential job loss that could come from it. But I think the big reason people hate the data centers so much is they perceive that it’s driving up their utility bills, and the oligarchs are getting rich off the data centers while they’re paying more.
The oligarchs are getting tax breaks, and you’re getting higher utility bills. That goes to Chamath’s point: people feel they’re unheard, and they’re going to take the only action they can, which is going to the voting booth.
Friedberg, you’ve heard my breakdown. You heard Chamath’s breakdown and then Sacks with a little bit of a rebuttal. Let’s do an alternate history here. Friedberg, if the war had not started, what would the midterms have looked like? And what do you think happens in the next 18 months of Trump’s presidency? Is there a way to turn around his legacy? If there is, what 2 or 3 pieces of advice would you give this administration to focus on in their remaining 18 to 24 months of power?
I kind of agree with the notion that the AI data center hysteria and the hatred for the Iran war are both rooted in this idea that you haven’t solved my economic problems first. I don’t think they’re the primary issue. I think they’re a reflection of the consequence of not focusing on my issues first, my economic cost problem first. So I think that’s still the biggest issue.
I think 2 things are about to happen in ’27. One is I think it’s going to be very hard, with this Congress and the sitting president, to come to a resolution on a clean budget. As a result, I think things are going to continue as they are, and I worry about the cost of things continuing to climb.
By the way, I’ll just reflect on Chamath’s point. COVID was an accelerating moment for this unaffordability problem for people who don’t own assets in the U.S. The reason is we doubled government spending. That’s what happened. By doubling government spending, everything inflates, including assets, which is why people who are asset owners benefit, and people who are earners for their labor lose because the cost of everything goes up.
If you’re making money from a paycheck every week, you can’t afford your healthcare, your education, your housing bill, your food, or your gas. I do think the Iran war was acutely off target because, at the beginning, the declaration was, “We’re not going to do any new wars,” and then a big war was started. So it was an easy target.
I do think this budget question is a big one. What’s going to happen with spending in 2027 and 2028? As we talked about a couple of weeks ago, I think the bond market will tell you where they think that’s headed, because if we continue this $2 trillion-a-year deficit spending, things are going to remain expensive. They’re going to get worse.
And it’s not about people choosing socialism as an ideology. To counter Chamath’s comment earlier, I’m not saying people are choosing socialism as an ideology or as a set of principles. Socialist outcomes arise from the policies people think are going to solve their cost issues. People think the best way to solve the cost of things is to have the government do more, and then you get more socialist policies. It’s not that people are saying, “I want to be a socialist.” That’s where you end up in this slippery-slope, spiraling problem, and I worry about how we’re going to resolve the budget going into 2027 and 2028.
I do think AI is going to continue to be a really big political hot point. AI is almost like the Ring of Sauron. Both sides want the Ring of Sauron. It’s this ultimate tool, this ultimate power. It is a risk to the state. It is fundamentally something that takes power out of the government’s hands and distributes it.
Let’s assume there isn’t a company making AI and there isn’t a company that owns all the data centers. Assume that there are a million data centers and a million variants of open-source AI. In a world like that, there is no oligarchy making money off AI. Let’s assume that’s the case for a moment and that we can get to that end state.
In that end state, the tools of AI pull a lot of power and influence away from the government, from the state. As that happens, I think the actors who want the state to be central—to have central command, central planning, central authority, and ultimately a central power structure—are going to react violently against that. They’re going to want to have the AI.
I actually think the calls to stop the development of AI are calls for them to have absolute control over AI, because it’s way too powerful for them not to control it. They want the AI. I think a lot of the policy we’re going to end up seeing—and this is where, if you play the game theory out, I still think, though I could be wrong, that we end up in a place where the government says, “We want to regulate compute.”
They’re not going to regulate the chips or the models, because those are too hard to catch up on. You’re never going to be able to regulate the development of software. It’s very hard to regulate the deployment of software, but the utilization of that software is fundamentally constrained by how much access you get to compute.
I think data centers end up being treated like utilities in the future. I do think the government is going to try to move down this path. They’re going to realize it’s the best path to get what I just described: the ability to control and manage who gets access to this technology.
Yeah. At a minimum, they’re going to get a great tax base. I think a lot of the data center stuff is performative. They just want to get a better deal. If you look at how Gina from Texas looked at data centers, she’s looking at it from a community-first perspective. If you look at how President Trump looked at it, it was from the companies’ and the economy’s—the GDP’s—perspective first.
But they both wind up in the same place: Don’t disturb your neighbors, don’t pollute, be a good citizen, and maybe splash a little cash there.
By the way, in terms of what Trump can do, I have a take some people might find a little shocking. I’m not a big fan of pardons, but they’re the law. They’re built into the system. Both parties use them liberally, and every administration gets more aggressive in using them. So until we change that, I think Trump should pardon everybody this week. Pardon his family, pardon himself, pardon every single person in the administration, because he’s going to do it anyway. Then that takes endless investigations by the Democrats off the table, or neuters them to a certain extent.
Then he should say to them, “Let’s work. Let’s make a deal.” The deal he should make, number one, is to do some tax reform on capital gains. Listen, we all make our money from capital gains, and I’m assuming most of the people in our circle leverage margin loans—in other words, take loans against our equity holdings. We should just say anything over $10 million gets taxed at 50% of the capital gains rate. If you pay the loan down, you get that as a credit. If you sell shares, you pay the rest of the capital gains. Get that out there and smooth it out. It’s a revenue stream, and it makes it feel like, hey, when rich people take these loans, at least some tax revenue comes in. You get rid of the “Bernie Sanders, these people haven’t paid taxes, and this year they paid zero” argument.
Then we’ve got to work on healthcare. I’ll leave it up to other folks with more experience to say what that should be. But Trump should say, “Hey, let’s work on healthcare.” I’m going to keep harping on the minimum wage. There are 2 ways to do it, obviously. You could say no taxes for people under a certain revenue or income threshold, as you sort of evoked earlier, Chamath. Or, just raise the minimum wage by $2 a year for 4 years and get everybody up to $15 an hour.
I know people will call me a libtard for saying this, even though I’m a moderate independent. I think it would be good for the aspirational class. It would counter this narrative that we don’t care about the bottom half of society. In fact, only 10 or 15 million people are impacted by a minimum wage under $15 an hour, so it’s an easy thing to do.
I’d go with number four: a ban on trading stocks in office and doing business with foreign countries. Rahm Emanuel has this proposal. People on the right have it, too. So I think we should get some consensus on that from everybody.
Finally, I think Trump should come to them and say, “Hey, listen. We know housing is what’s frustrating people a whole bunch. Let’s come up with a plan to incentivize the building of new homes.” He should go right into deal mode. He should go right into populist Trump, which everybody loves. If he solves 2 or 3 of these 5 things, I think it would redeem his presidency. And obviously, to your point, Sacks, he’s just got to end the war. He’s just got to end the war.
However it’s possible, I think the Democrats may force his hand. Then he might say, “Well, the Democrats forced me to do it.” That gives him this golden bridge to cross.
Look, before we totally throw the baby out with the bathwater, I think it’s important to realize that Reagan’s presidency has only grown in stature over time because the structural changes Reagan made became more obvious in their totality the further you got away from them.
I think the reality is, even though the Democrats may not want to admit this, an economic populist will have to understand that keeping the borders closed is incredibly important. The base of people you need to vote for you want more of the pie, so sharing more of the pie with people streaming across the border is just mathematically a bad decision.
Then you have tariffs, which, if you partner them with the amount of money people have committed, could revive domestic manufacturing if that money follows through. The further you get away from it and the more it’s allowed to compound, the more positive that’s going to look.
I don’t think it’s fair to say that they lost the midterm, but I don’t think it means history will look at his presidency as failed. Right now, I think people will say that some of these policies are obviously good, like closing the border; some are a question mark; and some, like Iran, you’re going to have to get out of. The further you get away from it, the clearer it’ll be.
So what should he do? Jason, to your point, I think he should figure out ways to support the parts of his agenda that will endure. What’s a great trade-off to make? Keep the border closed in return for some economic changes. Maybe the Democrats want a higher minimum wage. Maybe what they should do—the simpler thing, I think—is wipe out federal taxes for people earning $125,000 or under. It’s a nothing burger. It doesn’t really matter in terms of the revenue it generates. Close a few loopholes.
But trade it off for parts of his core strategy, because that’s what will endure. People will know the president had the courage to close the border when nobody else would, or when Biden just let people stream in. And the populist economic president who sits on top of those decisions in 2028, whether it’s a Democrat or a Republican, will have to thank Trump. That’s how the votes will have come to them, and that’s how people will feel like this economic malaise is being fixed.
Okay, Sacks, any final thoughts here? I’ll give you the last word again. Aside from getting out of the war—we all agree that’s number one on the list—what else can President Trump do in the last 18 months that would be productive for the country and for his legacy?
All right.
Look, let me give you 3 examples of previous midterm elections. We’ve had big midterm elections before. In 1994, Bill Clinton lost the House. It was a huge defeat. That’s when Newt Gingrich became speaker.
Clinton still went on to win re-election in 1996, so he made course corrections. Obama lost 7 Senate seats and 63 House seats in 2010. That was the shellacking. He made adjustments and still won re-election in 2012.
Presidents can definitely come back from midterm defeats. It’s pretty obvious: you just have to make corrections. What’s the one that didn’t work out so well? In 2006, George W. Bush lost a pretty bad midterm because of the Iraq War. People were very unhappy with it. This was 3 years after the start of the Iraq War, and Bush wasn’t up for re-election, but the Republicans then lost again in 2008.
Why? Because what did Bush do after the midterms? The surge. He could have adjusted course. Instead, he doubled down. I think this is really what it comes down to. There are a lot of people—neocons, basically—encouraging the president to do some version of the surge, some version of doubling down: some strike, Pickaxe Mountain, Kharg Island, what have you.
If you do that, I think you’re basically going to reenact the whole 2006–2008 playbook. If you course-correct and get out now, I think you’ll have a very successful next 2 years. That’s what it comes down to.
All right, we dipped our toe into data centers. Let’s do a double-click here and dive right in. Stocks got killed earlier this week: Nvidia, SpaceX, Tesla, AMD, Micron—everything across the board took a massive haircut. Close to $4 trillion in market cap was wiped out.
Now, again, Trump has overseen the top-ticking of the market for the last 2 years. Overall, I think the market is still up during his second term, but this was the worst single day in the history of the stock market by total dollars, not percentage. It’s the worst day we’ve had in a little while.
This puts a bit of a dark cloud over Anthropic’s IPO and, of course, OpenAI’s IPO after that. Polymarket, our great partner—thanks to Shayne for coming on Tuesday—says there’s a 71% chance OpenAI goes public this year. It was 95% before the midterms. The Wall Street Journal reported today that the haircut could be baked in now. Instead of $1.5 trillion to $2 trillion and then going to $3 trillion or $4 trillion, as Gavin Baker said on this podcast, we could be looking at a $1 trillion valuation.
Markets have rebounded a bit today. And all of this is based on, Chamath, whether these data centers are going to go forward or not. Are they going to be stopped? What’s going on with governors, specifically? Sacks evoked what happened here with Abbott losing to Gina Hinojosa, who took the time to do a little presser last night and give her position. She went on all the top shows. She’s being pretty clear she wants a great deal, and it doesn’t seem like she’s against them. She’s going to let people build them out until 2028, but it’s unclear what happens after that point.
You’re obviously connected to a lot of the major players in the AI space, Chamath. What’s going to happen with data centers? Can Democrats afford to pop the AI bubble by putting cold water on data centers and slowing them down? That will tank the economy. I think we all understand that. Is that in their best interest? Is this posturing and positioning to get better deals for their local communities, as is their job?
There was an incredible video where the SpaceX team went to Cochise County in Arizona to present their plans for this new 6-gigawatt data center. If I had to characterize that video—it’s on X if you want to find it—it’s these 2 very bright, earnest people from SpaceX trying to explain how they’re going to reduce electricity costs by 40%. They don’t use water. They’ll reduce property taxes by 20%. They’ll generate enough revenue to pay off a bond in the county.
Meanwhile, the audience is just screaming. The 2 people were very poised. Riley and Katie, those are their names. They’re good SpaceX employees, and they should be proud of how they conducted themselves. But they were getting run over.
My takeaway, Jason, is that this is an emotional issue that is not tethered to the facts, as Friedberg says. It represents a totem around the oligarchy and the economic fissure that exists in the U.S. economy. But it also suggests that the deals that have happened are a little gratuitous. Should it really be the case that the most powerful 7 or 8 companies need a tax break to invest $30 billion to $50 billion and then generate $345 billion to $500 billion? Probably not.
They’re doing smart deals because that’s what made sense. But I think what the incoming governors are going to say is, “These things are off the table. Instead, we want directly measurable economic impacts before we even entertain this.” I think what it means is we’re going to see a slowdown. That benefits the hyperscalers because they have existing capacity. I think it’s very bad for the neoclouds because they’re renters, and it’s pretty bad for the foundation models because they’re so short on this capacity and need it to grow.
To the extent that these frontier models are so much better than everybody else’s product, they’re the ones that need this capacity. It sets the hyperscalers up to be kingmakers, and they have this interesting game theory now.
Can you just say who the hyperscalers are, so people understand the difference between them?
Yeah. Amazon, Meta, Microsoft, Google, SpaceX.
Yeah.
What they’ll be able to do is make a decision. Path number one is to continue to rent the way you’re doing it today: the status quo. Rent to some customers, large corporates, right? Rent to the neoclouds, rent to the foundation models. That’s what they’re doing today. Everybody does that.
But then there’s the path less traveled, which all of a sudden becomes really interesting. You could decide to take all of the incremental compute and advantage your own products. Google has a ton of capacity coming online. They could say, “I’m going to give it to Gemini.” What does that do? It actually allows Gemini to catch up in quality relative to the foundation models.
Why? Because if you constrain the amount of assets that can actually generate compute, and a smaller surface area of it is available to the leading companies, those leading companies will have to pay more. The user experience will decay, and then other products can have time to catch up. That’s important because Anthropic and OpenAI are in a very fast recursive loop. They’re on a 60- to 90-day loop in improving their models.
Without an infinite amount of supply coming online, you can slow these folks down. That’s the path less traveled. Then there’s the crazy path.
Crazy path.
You should look for this one very narrow thing. You and I talked about this before on the pod, but SpaceX and Anthropic’s compute deal has a mutual 90-day out. Anthropic can terminate in 90 days, and SpaceX can terminate in 90 days. If SpaceX says, “I’m going to terminate this compute in 90 days,” that is the kingmaking move. Why? It’s a disproportionate amount of the compute that Anthropic needs.
It allows SpaceX to reprice that compute at the spot rate, which is probably double the market rate.
The market rate, right?
The market rate. And it’s probably not just double; it could probably triple. So whatever revenue you see on SpaceX’s balance sheet from Anthropic could 2x to 3x, and Anthropic would have no choice except to immediately capitulate and take the higher price.
It would also show all the other hyperscalers that they are very far away from having optimized economic yield. That forces them to consider: Should I be giving it to myself? Should I be selling it to other people? It then forces the neoclouds to take all the money they’ve raised, whether it’s debt or equity, and say, “I have 2 choices: buy as much compute today as I can at whatever price it’s available, because the other path is assured destruction and bankruptcy for me. I’m just out of the game.”
So what do they do? They’ll take the last gasp and do whatever they have. I think all the game theory, Jason, literally comes down to this one deal and whether Elon decides to rip it, because if he rips it, watch out.
Yeah. And the other 2 factors here, Friedberg—and just some predictions right now—we have the tokens here in the United States. If governors decide they want to put the screws to big tech, hyperscalers, and neoclouds, and they want to slow this down, they’ll need to understand that these data centers could be put in Canada, Mexico, Singapore, Malaysia, and many other locations.
The Middle East is a little bit off the table because of the war. When the war ends, you start putting these data centers in the Middle East, and the job of doing the next large model will not happen, Friedberg, on U.S. soil. It might happen in Saudi Arabia, the UAE, Qatar, Malaysia, or Singapore. We’re going to be a net importer of tokens.
This will be the biggest self-own ever, because all those $300,000 to $400,000 jobs building these data centers are going to go away. All that tax will be collected locally in Malaysia, where Google is putting a huge data center, and the data can make it around the world quite quickly. It doesn’t need to be in the United States, which these governors are going to learn is a very hard lesson in globalization.
What are your thoughts on data centers being banned, or whether this is just posturing? Friedberg, could we see the United States become a net importer of tokens instead of a net exporter?
Well, they might ban the use of AI, too.
Okay. Well, now we’re going to—
Yeah, a very far-out prediction—
—that’s being floated. That’s the conversation right now. Okay, unpack that. Who’s having the conversation? What exactly is the proposal? Because that seems a little far out there. The farther left you go, the more vocal politicians get on this point: no one sees the “benefit” of AI, and they don’t want to see AI developed, deployed, or used.
They want to ban its development and deployment. They want to ban data centers, technology companies from making models, and companies from using the models. I think this rhetoric leads somewhere more constrained than we are today. What are the chances something crazy like that happens? Handicap it for us here. 10%.
California passed legislation, and Gavin Newsom vetoed it. I think it’s much higher than 10%. We’re seeing state-based legislation. Sacks has spent quite a lot of time on this, but there’s a patchwork of hundreds of laws passing all over the country trying to regulate model development, deployment, and utilization; what models are allowed to do; whether the local state government gets to audit and inspect how the models operate; how people are using them; and what they’re outputting.
What do you say, Sacks? Answer his—
Fundamentally, remember: a model is just a bunch of software. It’s code. It’s a point about free speech: can I actually run code on a computer? It’s a restriction on someone reading a book, in a way, because at the end of the day, you’re inputting and outputting digits, ones and zeros.
True sovereignty, yeah.
And remember, we have laws and regulations about what those ones and zeros do. If those ones and zeros take money out of your bank account and put it in mine, I’ve broken the law. I’m a felon, and I’m going to get convicted of that felony. If they’re used to create a weapon, that’s illegal. I’ve broken the law; that’s a felony. You can go down the long list of things people should be concerned about.
So you have to ask yourself: if we already have laws that protect us from the consequences of people using this sort of speech, this sort of software, what are we actually doing? I do think, once again, that the primal, unconscious political impulse is that something threatens the state—threatens government itself. I think that’s a big part of the motivation here. I’ve said for a long time that there’s a power play.
The question—and this is me waxing philosophical for a second—is whether an organization like Anthropic, or whoever, ends up regulatory-aligned with the state, such that the state has unlimited access to that model’s capabilities for development and deployment, while individual citizens and companies do not. If you go back to the modern issue, the Second Amendment issue is no longer about who has the bullets and the guns. It’s about who has this technology. If the state has the technology but citizens and businesses don’t, then the state has supreme power.
Remember, there’s no such thing as a state that doesn’t have central planning and command, which means you’re giving a small group of people an extraordinary amount of power. The whole point of the United States, and the whole point of the Second Amendment, was to always have a check on that power.
Okay.
So if the state ends up getting all of the AI and no one else gets the AI, then you no longer have a check on that power.
I think you’re getting into Dr. Doom scenarios. I think you’re completely wrong. I think there’s about a 1% chance something crazy like this happens. But I like that you exist in the world, Friedberg, to present it, because you’re the backstop to this level of lunacy.
By the way, Jason, call me Dr. Doom all you want, but this is a fundamental question about the intention of this country. The intention of the country was individual liberty.
Yes.
The rights of the citizen, the rights of the individual. It’s not about state power, and it’s not about us. We have a lot of laws that protect citizens from each other.
Mhm. What we’re doing now is creating a different system that extends well beyond the founding principle of the United States, giving a small group of actors an incredible amount of power that they’re going to centrally command and control.
I think that’s something we need to pay attention to. I’m not saying this is happening and the world’s over. That’s why I’m trying to say it’s not binary. As the policies develop, we have to keep in mind what we’re doing.
You’re saying it’s a slippery slope. You’re saying it’s a very slippery slope. Yes. Okay, Sacks. But I do think, just like the P(doom), where people say 10% of people are going to die, I like to put a percentage on it. What are the chances of this happening? Chamath or Sacks, you’re the closest to this, having been the czar. Do you think what Friedberg’s describing is in the realm of possibility? If so, maybe you could put a percentage likelihood on it. And what’s your take on data centers and the potential for the U.S. to be a net importer rather than a net exporter of tokens?
It’s 100% that Democrats are going to make a major push for AI legislation in the next Congress. The question is what that looks like. I don’t think you have to speculate too much to understand where it’s going. I believe Democrats want to create a surveillance and censorship state turbocharged by superintelligence. That’s ultimately the end state.
How do you get there? You create a federal Department of AI. You require that all the so-called frontier models be reviewed and approved by it. Regulators give their notes, and you can’t release a model unless you address their concerns. Their harms will include all the things the government is telling us social media companies should be concerned about, like disinformation, misinformation, and dissident thought—all the things the government pressured social media companies to crack down on and censor during COVID. How quickly we forget all of this. They’re going to reenact that whole regime using AI, and it’s going to be infinitely more powerful because we’re talking about superintelligence here.
I do believe that’s the regime we’re headed toward. It may not be exactly the version Friedberg’s talking about, where only the government can use the tools. I don’t think that’ll be the case.
No, not that binary, Sacks. Sorry, I meant that they’re obviously going to get access to models and capabilities that the public doesn’t. Any distinction between what a central state actor gets and what the general public gets is where that distinction of power arises. It’s what makes it difficult for the founding principles to persist. It would be a quasi-de-facto nationalization of the technology. Chamath, what’s your take on this? I’m going to force everybody and try to pin you guys down a little bit. Do you think this nationalization of the technology—the government having the best version, everybody else in the United States, corporations and individuals, having 60%, 70%, or 80% of that version—is a possibility?
No, I don’t think this is the priority. People will get animated about this issue, but I don’t think it matters. To the extent that people legislate an oligopoly, it’ll further entrench economic power, and I think that’s counter to what these people want. When they connect the dots, I don’t think that’s what’s going to get legislated. I don’t think they have the energy for that. I think at the end of—
What’s going to get legislated is a quid pro quo. The government gets more power, and the duopoly of closed-model companies gets more regulatory capture, or moats around their businesses. That’s going to be the quid pro quo. The government’s going to come in.
The Democratic Congress is going to say, “Look, we have to regulate this. It’s unsafe.” What do you think those companies are going to do? Do you think they’re going to say, “No, you can’t do that”? They’re going to say, “Okay, let’s work together on this.”
I think what’s going to happen is—
—and the government ends up with more power, and those companies end up with regulatory protection.
Chamath, go ahead.
I don’t think so. I think that if anybody legislates AI, it’s going to be broad-based and include a lot of coverage around data centers and AI factories. To the extent that happens, maybe you’ll get some regulatory capture on one end, but you’re going to constrain the supply so severely on the other that the people who’ll really be in control of it are Meta, Amazon, Google, Microsoft, and SpaceX.
You have to remember, the foundational model companies are not home free. What does home free mean? They have a fortress balance sheet, a profitable business model, and quasi-infinite compute for their own ambitions. Right now, they’re renters, and the renter class in data centers, as of Tuesday, is very much at risk. If anything, I think all the renter classes will get co-opted. For example, I don’t know if you guys saw today, but Google just took over a Crusoe data center. They co-opted the renter. They’re like, “You’re gone.”
“We’re going to take it over.”
They did.
Yeah. So why is that happening? It’s because they have all the critical ingredients, specifically the power. I think even a regulatory-capture campaign can only work to the extent that the government violates all these other rules and says, “Here’s all this free electricity. Go build wherever you want, and we don’t care about energy.” There’s no money to pay for that, and there’s no political will to do it. The practical reality is that the Mag 7 or Mag 8 have more power, and when politicians look at that calculus, I don’t think either side wants that.
I did a calculation this week. I was looking at the phenomenal impact agents are having inside our venture capital firm, Launch.
Say the name of your venture capital firm, Launch.com. Where can I go to learn more about that?
To Launch.com.
Thank you.
Thank you very much. I’m just trying to let people know I have a day job beyond podcasting. I invest in 100 companies a year, and we have 20 people, and we’re spending $40,000 to $50,000 a year on these subscriptions, with Grok and Claude being the main ones. I estimate we’re getting $2 million in value from that. I estimate we’re getting about $1 million in incremental revenue, or maybe 15 to 20 times in incremental revenue for the dollars we spend.
The velocity of our investments is going up because we’ve automated everything. The velocity at which we create new media properties is going up because we’ve automated everything. We’re building systems, and only 10% to 20% of people in corporate America are using this technology every day on a regular basis. In other words, there are somewhere between 5 and 10 times as many people who could start using this every day.
When they go agentic, they start using this because it’s so accessible to the average human being. Everything gets abstracted. You don’t have to learn to code or set up your rig or your harness. You just tell it what you want, and it uses 50 times as much. So I think there’s 500 to 1,000 times the need for this. When that happens, this compute is going to go wherever it needs to go.
I can tell you that what we’re doing is only 70% to 80% of what the frontier models are capable of. We’re not solving Navier–Stokes, obviously. We’re on the verge of ordering everybody a Mac Studio with 512 GB for $20,000, which would mean $400,000 in compute on people’s desktops. We’d run local models first and then fall back to them because we don’t want to be throttled. We want to have unmetered intelligence. Offshore data centers will be the trend of the next 2 years. That’s what you’re going to see, along with desktop usage of local models. I’ll make that prediction here.
Open source is obviously the power of that.
I think you’re onto something, Jason. When we’re talking about AI regulation in the next 2 years, I think we need to separate two different things: federal and state. The feds are going to deal with AI safety and creating a federal department of AI, along with all the Orwellian risks around what could turn into censorship, surveillance, and control. That’s one bucket of issues. There’s another bucket around infrastructure, which will be dealt with at the state and local level. Both of them are going to play out.
At the state and local level, I think people are being way too optimistic about what this election means. I’m seeing Gavin Baker say, “Oh, don’t worry. Inosa[?] seems totally reasonable. She’s not threatening any projects.” I’ve watched the interviews with her. I went and watched the interviews Gavin watched, and I don’t see the things she’s saying. She does not say existing projects are protected. She simply says there’ll be no new permits issued until they pass legislation.
Dealing with the issues she’s talking about. I saw an issue she’s talking about.
I’ll explain it. I saw an analyst from SemiAnalysis come out and say, “You do realize that projects under construction right now don’t necessarily have every single permit they need to actually open for business?” In other words, once you’ve constructed the data center, you may still have to get an air pollution permit. There are various kinds of permits that sometimes you don’t get until the end.
If you take what she’s saying seriously and literally—that there will be no new permits issued until this new regulatory regime comes into effect—we could end up with a bunch of stranded, half-finished projects. I’m not saying that’s necessarily going to happen. I’m not criticizing her because I don’t know what she’s going to do. She hasn’t even taken office yet. I’m simply pointing out that this could go a lot of ways. They could do this in a reasonable way or an unreasonable way, so you could end up with a lot of half-finished projects.
My base case is that the projects that are fully permitted will get done in 2027, but I don’t think there are going to be major new projects under this new regime for a long time, for the simple reason that there’s no political incentive for anyone to grant a new permit or approve a new project. What’s their incentive? Think about this: in this last election, both Democrats and Republicans ran campaign attack ads on any politician who approved a data center. If you’re a politician, or just a civil servant who wants to move up the ladder, what’s your incentive to stamp approval on anything? I don’t see it. I don’t see anyone who’s politically ambitious wanting to stick their neck out in any way to approve anything over the next 4 years.
The ones that are already approved and fully permitted will probably get done, but the ones that aren’t approved yet? Good luck. I think it’s going to be very difficult. I think we get by in 2027, and there’ll still be meaningful capex. I get really worried about 2028. The Democrats have a perverse incentive around this, because if that capex tanks in 2028, they’ll blame Trump for it.
Yeah, but they’ve already set the table. They have such an amazing footprint right now; they’re already in the pole position. I think tanking the economy opens up an attack vector for them. They’re much better off saying, “Here’s what we want,” and extracting value from them. We want free electricity. We want schools. We want this. Just negotiate hard. And if you don’t—
Yeah.
Yeah. And if they don’t, if the data centers and these governors, whose job it is to fight for their people and get the best deals, just like it’s the companies’ job to get the best location and the best deal for their shareholders, can’t come to an arrangement, these things can be put in Canada, Mexico, or anywhere. We have Starlink and fiber lines. It’s extremely easy. It’s very common to run fiber cables everywhere around the world. It’s an incredibly speed-efficient system we have here. It’s not like 20 years ago, when loading your Gmail in Japan was extremely difficult. It’s very easy and very crisp. We have CDN technology. It’s going to be a no-brainer.
Let alone the prescience of SpaceX, which is severely undervalued as a stock. Tesla—I’ve been buying shares through this whole insanity. They’re going to space and bringing them up there now. Seems like a pretty brilliant idea, doesn’t it, Chamath? If they’re going to block it, all of a sudden, once again, Elon saw around the corner, did he not?
Well, do you remember? I think we talked about this. We said his 90-day out with Anthropic is probably because he negotiated and saw around the corner, not the other way around. People used to think, “Oh my God, this is really bad for SpaceX if they could pull the rip cord.” That’s not going to happen. The question now is, does it happen the other way around? It’s pretty interesting. I think Elon and the Anthropic team have a pretty positive working relationship.
Oh, this has nothing—hold on. I want to be clear. This has nothing to do with their working relationship, or whether Anthropic is a good or bad company, or SpaceX is a good or bad company. This has nothing to do with that. This is a spreadsheet, and the spreadsheet will dictate the answer. The question is, if spot spikes, are you economically forced to do this beyond a certain threshold?
Okay, I’ll take a slightly different side to that. There’s still a lot of acrimony between OpenAI, Sam Altman, and Elon.
Elon has aligned himself very deeply with the Anthropic team. They have an incredible working relationship. I don’t think he just throws it open to spot, even if that was the most ruthless short- or mid-term economic model. Let me finish. No, that’s not what I’m saying. You’re not listening to what I’m saying.
Just let me finish. Then you can counter. You guys are allowed to let me finish my sentences. I know that’s hard to believe.
Not if you miscast what I’m saying.
You can totally respond to it. Just take a beat. Moving to spot might be the right ruthless economic decision. I think Elon’s decision is going to be informed in part by his relationship with Sam Altman, or lack of one, and by Anthropic’s loyalty and success, making them a great long-term partner, especially for the space data center business.
So I could see him, instead of throwing it up to spot, saying, “Hey, let’s meet somewhere in the middle.”
I’m not saying to throw it up to spot. What I’m saying is, when spot moves sufficiently, even a long-term offtake moves so much that existing contracts are negative EV. So you’ve got to move existing contracts to positive EV territory. If you can get out in 90 days and reprice something into positive EV territory, you should.
My point is that beyond a certain point, if a bunch of data centers get slowed down and a bunch of constraints happen, there’s going to be a lot less capacity than people think. The dollar cost per megawatt per month is going to go through the roof. The dollar cost of a GPU hour is going to go through the roof, and the people willing to pay the premium are going to be the ginormous companies with huge balance sheets. You could just sell it to Google or Meta if you wanted to. You don’t have to sell it to OpenAI.
Okay, Sacks. The really interesting part of all this, just to dip into the science corner, is that this infrastructure does 2 things, right? There’s inference, with people using the product every day, and then there’s building the next model. If you want to run away with the next model, cornering compute at any price is a huge advantage.
For Google, which has been behind, if they use all their compute to make their model—or if Microsoft or NVIDIA decide they want to start accelerating their internal models, which they have—this could determine who has the best frontier model. Over to you for the final word.
When you’re talking about infrastructure of this magnitude, these gigawatt power data centers cost around $50 billion. Nobody can make that investment in conditions of uncertainty. The last thing you ever want to do is break ground on a project like that and then, halfway through, get enjoined. You end up with a strike by your workers or some community group that stops you. You really need conditions of certainty before you make an investment like that.
I think we’re headed toward conditions of great uncertainty. You’ve got our biggest data center states where there’s basically been a regime change, and all those governors are indicating that they favor moratoriums until XYZ happens. They can differ a little bit on what that is. It may not be a conspiracy to hurt the Trump economy; it may just be the reality of what’s going on.
I’d expect there will still be meaningful capex in 2027 as a base case, meaning those fully permitted, approved projects can move forward or finish. But by 2028, I’d expect the pipeline will either be closed or significantly reduced, and that will have a big economic impact. I do think Democrats will try to blame the president for that, even though they’ll be solely responsible for it.
4. Science Corner: OpenAI solves Exponential Time Hypothesis
All right. Everybody has been demanding, Friedberg, that we do a science corner on this exponential time hypothesis. I guess a layman’s way of explaining it would be that computers can’t solve important problems quickly. But there seems to be some science and some recent innovations that say maybe this hypothesis is no longer true. Please inform our audience.
The reason for doing this science corner on this OpenAI publication is just to highlight some of the benefits and value that come out of AI. These are pretty substantial breakthroughs that are happening. This one is called the exponential time hypothesis, and it starts with the most fundamental question in computer science: Can a problem be solved with a computer, and if so, how long does it take? That tells you whether it’s actually practically computable.
This is the key point of computer science. We want to create these machines that can solve our problems. There’s a class of problems called P problems, where you can solve them in what’s called polynomial time. The bigger the problem of a particular type, the longer it takes. Polynomial time just means size cubed or size squared. It tells you how many seconds or minutes or hours it might take on a specific computer to solve that problem. As the problem gets bigger, it takes more time.
Then there are what are called NP problems. These are problems where you can very quickly figure out if the answer is correct. It can take a long time to solve a big Sudoku puzzle, but you can check the puzzle in a couple of seconds and see if it’s right. That’s an NP problem. The key question in computer science has been: Do P problems equal NP problems? Meaning, can you actually solve any problem that you can quickly check in a short period of time? If so, then everything is computable. The entirety of the universe is computable. That’s the big question in computer science.
The exponential time hypothesis was proposed in 1999 by a couple of guys. They said that for a specific type of problem called a 3-SAT problem—which we don’t really need to get into—the amount of time it takes is exponential, depending on how big the problem is. The bigger the problem gets, the more it scales exponentially in terms of how long it would take a computer to solve it. That means they’ve demonstrated there’s a class of problems that can never be solved in a short period of time, and therefore that entire class of problems means that P does not equal NP.
What OpenAI did is disprove that, in a 175-page paper. It comes down to this one little formula, which you can pull up on the screen here, Nick. We don’t need to spend a lot of time analyzing it to understand. This formula demonstrates that this 3-SAT problem, this particular class of problems in computer science, can be solved in subexponential time, which means it won’t take an exponential amount of time. So they disproved this big exponential time hypothesis that everyone thought was real. What that means is that P could still equal NP, which means that theoretically every problem in the universe can be solved with a computer.
This solution is being evaluated now. It’s not fully verified by third parties yet; people are still trying to understand it. It’s 175 pages, with a very novel set of approaches that it took to try to get there. They did the whole thing end to end in less than 10 days. Some reports have argued they did it in a few hours. This was a deeply challenging question in computer science, and has been for 27 years: Is this Exponential Time Hypothesis real or not? They just proved that it’s not, and they did it in a few hours.
It may be esoteric for the majority of people listening to this, but it does a couple of things. Number one, it keeps in play the idea that everything in the universe can be analyzed and every problem theoretically could still be solved with a computer. Number two, it’s amazing that OpenAI was able to do this in a few hours and deliver this sort of result that mathematicians and computer scientists have been struggling over for 27 years. This proof is so far holding up. This is amazing, and everyone in computer science, I think, was kind of shocked when this came out.
Absolutely. Yeah.
It was shocking in terms of how it was presented, how they solved it. It was very novel, and again, it’s one of these moments of humility for mathematicians, for computer science, being like, “Holy, I would have never done that.”
And, Sacks, you were just very heated about the Exponential Time Hypothesis in the group chat for the past 3 weeks. So when you hear Friedberg tee this up here this crisply, how does this impact your thoughts on the 2028 election and the future of humanity?
Look, I think AI, superintelligence, whatever you want to call it, is going to create a lot of good things. The fear around it—maybe some fear is justified, because anything that seems powerful can make people afraid. Obviously, the future is unknown, so you can’t prove exactly what’s going to happen. But I tend to think there are going to be so many benefits to this technology.
Frankly, if we don’t do it, China’s going to be the one that benefits. JCal, you’re right that if the U.S. doesn’t build the data centers, they’ll get built in other countries and we’ll be paying to import the tokens. Very dangerous in Canada.
Very dangerous situation.
So dumb.
Yeah.
Crazy. Chamath, I saw you nodding there.
Yeah, you have a background in computer science.
My formal education was in electrical engineering, but I’ve worked around software engineers for so long now. This is an incredibly interesting insight.
All right, another amazing episode of the All-In Podcast. The number one, two, and three podcast in the world in terms of the rankings. I think—wow, what a week we had. We’re number one in the livestream, and obviously, four weeks running, we’ve hit number one. Thank you to the audience, and we truly appreciate you being on this journey with us. You can always expect as candid a discussion in the world as possible. It was very interesting, your discussion with Rick Rubin, Chamath. He was very taken with the podcast, and he’s somebody who understands art and unique moments in the world. I thought it was very interesting that he kept pressing you on, “What’s the magic here?” He’s kind of the sherpa of magic in many ways. I think the magic here is that we’re candid, we break each other’s chops, and we try to move the ball forward with our candor.
You had some very nice things to say about all of us, Chamath, and I just wanted to say thank you. You’re my brother, and I love you.
Thank you. I love you too. What I said to him, basically at the core of it, is that we’re a family. We get together and chop it up once a week, and it’s very rare to have a window into family squabbles.
Yes. I think that’s basically why it works. It’s kind of unique. That’s maybe why the Kardashians work.
You gave such a compliment. I think I said it was the greatest compliment I’ve gotten in my life. You said, “You know what J-Cal is? You know that quote from the scene in The Town where Ben Affleck goes to Jeremy Renner and says, ‘I need you to come with me. We’re going to mess some people up, and you can’t ask me any questions’?” And Jeremy Renner looks up and says, “Whose car are we taking, yours or mine?”
I’ve always felt that way about you.
That was very complimentary. Love you, bestie.
We open-sourced it to the fans, and they’ve just gone crazy with it. Love you.
We should all just get a room and have one big orgy because they’re all just useless. It’s this sexual tension that we just need to release somehow.
Your feet.
We need to get merch.
I’m going all in.