Sebastian Mallaby

Sebastian Mallaby is the Paul A. Volcker Senior Fellow for International Economics at the Council on Foreign Relations, a two-time Pulitzer Prize finalist, and the author of six books — including More Money Than God (hedge funds) and The Power Law (venture capital), both of which have become investment classics. He is a former Economist and Washington Post journalist who now writes and broadcasts on where markets, technology, and geopolitics collide.

His latest book, The Infinity Machine: Demis Hassabis, DeepMind, and the Quest for Superintelligence, is a reported history of the modern AI race told through its most scientifically-driven lab. Reporting it — including internal DeepMind documents on the safety-oversight fights, published over the general counsel’s objection — makes Mallaby an unusually well-sourced sceptic: he treats AI both as a genuine technological revolution and as a field where the loudest company may be its weakest business. He describes himself as someone who has ‘spent years studying the history of AI’.

Core positions

An OpenAI bubble, not an AI bubble. Mallaby’s signature 2026 distinction: the sector’s demand — compute, semiconductors, data centres — is real and rests on fast, genuine progress since ChatGPT; but OpenAI the company is a coin-flip, structurally out-spending an under-monetising user base and likely, on his standing bet, to run out of money within 18 months and be absorbed by a cash-rich acquirer.

Government stakes are ‘cheating capitalism’. He reads OpenAI’s floated ~5% federal-stake proposal as an end-run around the level playing field, using the Intel precedent as a warning rather than a model — once Washington picks winners and steers contracts, ‘you don’t have a market anymore’. He grants a national-security carve-out for chips (Taiwan concentration) but denies it for a strategically-replaceable model lab.

Compete with China, but negotiate. A 2022 supporter of chip export controls who now doubts they work, Mallaby argues from a Cold War frame — mutually-assured-destruction-style competition alongside a non-proliferation regime — that the real danger is near-frontier Chinese models released open-weight, and that only a US–China deal can contain it.

Safety cannot be self-imposed. The lesson of his DeepMind reporting: a lab in a competitive posture cannot bind itself to safety restraints; only governments enforcing restraint on all players at once can, which for open-weight models means including Beijing.

Appearances beyond the wiki

Mallaby is a recurring voice on the AI-economics circuit; these 2026 appearances (not yet ingested) are tracked here as leads:

In the wiki

The Spillover — his own weekly show, co-hosted with Rebecca Patterson:

• The AI Bubble, the Productivity Paradox, and India's AI Summit — the OpenAI-bubble-not-AI-bubble thesis in its earlier form, plus the productivity paradox and stalled AI governance (February 2026)
• U.S.-China AI Race Escalates, Chip Bans Aren't Working, and a Lesson from Nuclear Proliferation — debating export-control architect Chris McGuire: the chip choke point, the cloud loophole, distillation, and a possible AI non-proliferation treaty (March 2026)
• Private Credit's Black Box and Why It's Not 2008 But Still Risky — the credit-market wobble, mark-to-model opacity, and why it rhymes with but is not 2008 (March 2026)
• The Modern Oppenheimer, Demis Hassabis, and the Quest for Superintelligence — Patterson interviews him on his Hassabis book, then a listener mailbag on the dollar, AI economics, and drones (April 2026)
• The Dollar Hedge Debate, Gold's Track Record vs. Bitcoin's Potential — a staged debate in which he takes the Bitcoin side (and concedes gold won) (July 2026)

As a guest:

• Sebastian Mallaby on OpenAI's Cash Crunch, the AI Bubble Debate, and the China AI Race — his runway thesis, the government-stake critique, and the China turn (Prof G Markets, July 2026)

Book:

• The Infinity Machine — his 2026 history of the AI race told through Demis Hassabis and DeepMind

Concepts:

• Compute Export Controls — the US chip-control strategy he backed in 2022 and now second-guesses
• Token Economics — the AI unit-cost problem central to his OpenAI case
• Narrative Valuation — the founder-story dynamics he sees in OpenAI’s fundraising