Patrick O'Shaughnessy

Patrick O'Shaughnessy × Neil Movva

Sola Research is repositioning inference around long-running background agents, forecasting workloads shift from 50/50 background and real-time by year-end to 90/10 long term as cheaper tokens unlock unbounded demand.By trading latency for throughput, using overlooked chips and 1MW sites with 95% uptime, Movva targets radically lower costs, while KV-cache waste, idle GPUs, HBM supply, and the durability of frontier labs’ three-to-six-month premium remain key watchpoints.

Patrick O'Shaughnessy × Ben Thompson

AI’s near-term bottleneck may be capital rather than compute or power, as funding shifts from free cash flow and debt toward Google equity and NVIDIA’s $500 billion vehicle.Thompson’s Berkshire analogy makes Search a possible funding engine for AI’s “basically all white-collar work” TAM, while payback periods, hyperscaler chips, and an air-gap risk remain watchpoints as 2028-29 capacity arrives.

Patrick O'Shaughnessy × Eric Vishria

AI is likely to produce an oligopoly plus $100B specialists, not a single winner-take-all lab, while Fireworks shows inference’s hidden moat: roughly 5X speed and multiple-X throughput on the same models and NVIDIA hardware.SaaS incumbents now face “Get to AI or be worth three times revenue,” as migration becomes easier and cost, iteration speed, and transportability matter more, with energy—especially China’s roughly tenfold buildout next year—the key constraint.

Patrick O'Shaughnessy × Vlad Barbalat

Liberty Mutual’s $120B balance sheet combines roughly $70–75B of reserves with growth credit and equity, while permanent mutual capital avoids shareholder pressure and supports 7–10% portfolio targets.Barbalat now questions whether AI makes future cash flows—and therefore multiples—structurally less visible, with four-year software credit appearing safer than 30-year Salesforce or Oracle debt and potentially steeper credit curves ahead.

Patrick O'Shaughnessy × Dara Khosrowshahi

Uber is making AV supply its moat through 30+ partnerships, financing, depots, charging, and insurance; network vehicles are 30% or more busier, improving fleet ROI.The upside is a potential trillion-dollar marketplace, but AI budget overruns, Chinese manufacturing advantages, social backlash, and execution in Uber One and hotels remain risks.

Patrick O'Shaughnessy × Brian Chesky

Brian Chesky expects a consumer AI renaissance within 12–24 months, even as 159 of the last 175 YC companies were enterprise-focused and subscription, advertising, e-commerce, and inference-cost models remain unresolved.Project Hawaii already delivers 600 basis points of lift on $134B, while Airbnb’s person-centered identity and agent strategy could broaden the platform beyond homes through cautious sandbox experimentation.

Patrick O'Shaughnessy × Mitchell Green

Mitchell Green sees public software as a potential risk-adjusted opportunity because distribution, customer success, and switching costs still protect incumbents such as Workday despite AI disruption fears.He expects the AI capex bubble to end badly as models commoditize, with the eventual drawdown creating an entry point, while Lead Edge’s sell discipline and 70% special-situations allocation offer nearer-term liquidity advantages.

Patrick O'Shaughnessy × Shyam Sankar

Shyam Sankar argues that military innovation comes from heretics who build outside bureaucracy, while defense-specialist spending rose from 6% of major weapon-systems spending in the Cold War era to 86% today.His reindustrialization thesis links production capacity to innovation and AI productivity, with hundreds of founders now treated as “plan A” inside the Pentagon and more than $100B invested in their projects, making execution and cost-plus contracting key signals to monitor.

Patrick O'Shaughnessy × Dan Sundheim

Dan Sundheim sees a moment-in-time opportunity in late-stage privates, where some of the largest companies by market cap remain private, and AI-lab investments inform public-equity analysis.The LLM debate now centers on unprecedented capital intensity: four or five models may endure, while hyperscalers could grow faster yet worsen as labs potentially insource compute in 5-10 years; software, scaling laws, and training returns remain watchpoints.

Patrick O'Shaughnessy × Gokul Rajaram

Long-horizon agents that are “resilient to failure” are moving product development bottoms-up, with PMs committing code and PM-to-engineer ratios heading toward 1:20.That shift elevates judgment and self-serve execution while pressuring thin AI applications, seat-priced software, and ad middlemen; durable businesses will need control of data, money, workflows, or networks, with agentic interfaces and outcome-based pricing key signals to monitor.