Ben Thompson

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.

Andrew Sharp × Ben Thompson

NVIDIA’s financing platform with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR pitches AI GPUs as long-lived infrastructure for patient capital.Yet A100 pricing at CoreWeave may reflect stranded air-cooled facilities, while abstraction above CUDA and potential overcapacity leave GPU economics, timing, funding air pockets and NVIDIA’s moat unresolved.

Andrew Sharp × Ben Thompson

Apple’s mid-cycle Mac price increases suggest it missed the memory shock in both procurement and pricing, while delayed Siri may arrive in 2026 with 2024-era capabilities.As major suppliers prioritize HBM, DRAM scarcity could let Chinese producers climb the learning curve profitably, with export controls potentially accelerating that competition.Vibe coding makes highly bespoke software viable: an AI assistant catalogs household objects, recognizes roughly 95% from photos, and links them to locations through QR codes.

Andrew Sharp × Ben Thompson

Fast inference retains a premium while humans supervise agents, but longer autonomous runs shift the bottleneck toward KV-cache capacity and tiered memory.Cerebras and Groq show the premium persists for voice and consumer responsiveness, but off-chip memory can make performance “totally plummet.”That favors slower, cheaper commodity infrastructure and potentially China, while challenging NVIDIA’s integrated inference economics without displacing its training advantage.

Andrew Sharp × Ben Thompson

AWS’s 28% growth and expanding margins reinforce its cost-led cloud strategy, combining custom silicon, service breadth and lock-in to turn infrastructure efficiency into pricing power.Large-scale training still favors NVIDIA’s tightly connected data-center architecture, but cheaper, more efficiently utilized inference could make Trainium and AWS’s commodity economics a stronger AI catalyst.

Andrew Sharp × Ben Thompson

AI agents shift the AWS contest from infrastructure cost to capability, favoring Nvidia/OpenAI or Google’s integrated stack over cheaper Trainium and open-source models.Amazon may retain legacy workloads yet lose new AI customers, while Apple’s strong results delay strategic change; Nvidia scarcity, installed-base inertia, and leadership culture remain the decisive risks.

Andrew Sharp × Ben Thompson

TSMC’s customer-first model helped build its moat but may have left 5 nm and 3 nm underpriced as leading-edge fabs became shorter-lived and cost “well into the 30 billions.”Planned capex of $52–56 billion this year still may not provide enough 2028–29 capacity, transferring supply risk to NVIDIA, Microsoft, Google and other buyers unless Intel and Samsung become credible second sources.

Andrew Sharp × Ben Thompson

Groq’s compiler-first, SRAM-based architecture delivers extremely fast inference but only 256 megabytes of memory per chip, creating a sharp trade-off between latency-sensitive applications and context-intensive workloads.Nvidia’s licensing and hiring arrangement could turn that niche into a software- and supply-chain-enabled platform, though the non-acquisition structure also highlights an antitrust regime that may make consequential deals easier to avoid reviewing.

Andrew Sharp × Ben Thompson

Substrate’s particle-accelerator-powered X-ray lithography claims could halve chip-manufacturing costs, but consistent sources and compatibility with doping remain unproven.Because lithography anchors the fab’s surrounding processes, replacing ASML’s tool may require rebuilding etching, coating, and doping integration from first principles.Bubble-era capital is financing this 1% possibility alongside thermodynamic-chip experiments, while ASML and TSMC’s co-evolution keeps the incumbent system deeply locked in.

Andrew Sharp × Ben Thompson

OpenAI’s roughly $1 trillion of infrastructure commitments aim to make ChatGPT the agentic layer across devices, email, calendars, and even sleep monitoring for as many as 8 billion assistants.Ben now sees the sprawling strategy as rational amid “astronomical, massive growth,” while compute shortages, subscription expectations, and AMD diversification remain tests of financing, execution, and platform leverage.