🔮 Exponential View’s year in review
Our best thinking from a year of acceleration
2025 was the year the compute economy became real.
Massive infrastructure commitments. Data centres acquiring power plants. AI models that reason. The central question went from “what can AI do?” to “how do we build the energy systems, governance frameworks, and coordination mechanisms to absorb what’s coming?”
That question – how intelligence, energy, and coordination collide – is the organising logic of Exponential View. Nearly every meaningful story we tracked this year sat somewhere on that fault line. You couldn’t understand one without the others.
Here’s a curated map of our most durable work from 2025 – your fastest way to catch up and revisit the ideas that matter.
Begin with the stars (★): the pieces most likely to outrun the news cycle by far. Choose two or three that fit your current challenges to make the most of today’s review.
And if you’ve been reading the free editions, wondering whether to go deeper, now is the time – it’s only going to get weirder!
Big picture / thesis pieces
★ The next 24 months in AI (Dec 3) A unified view of AI’s critical pressure points – energy, compute, talent – and what to watch. This is the framework piece.
The great value inversion (Sep 20) “Has value broken?” Electricity at negative prices, credentials losing worth, abundance creating crisis. This is distinctive EV thinking as the economic grammar is out of date.
★ Is AI a bubble? (Sep 17) + the live dashboard The five-gauge framework for assessing boom vs. bubble. Circulated through boardrooms and policy circles.
★ Energy is technology (Sep 17 2024) Our evergreen primer on how energy went from being a commodity to being a technology and why this changes everything.
The shift to computation (Nov 7) Zooming out to understand what the infinite compute demand means for the rest of the economy.
China / Geopolitics
The Chinese Sputnik that demands our attention (Dec 31 2024) We were the first to contextualize the DeepSeek surprise in December 2024.
72 hours in China (2 Jul) A firsthand look at daily life, technology, and pace inside contemporary China.
Paul Krugman and Azeem on the world in 2025 (Jan 31) The conversation that framed the year: tariffs, trade wars, the new information ecosystem, AI and debt.
Kimi K2 is the model that should worry Silicon Valley (Jul 15) “AI has its Vostok 1 moment.” The strategic analysis of what Chinese open-source means for US incumbents.
★ Why China builds while the US debates, with Dan Wang (Oct 1) Why America’s constraints are largely internal, and why China’s advantage lies in coordinated execution.
Under-the-radar
How Europe outsourced its future to fear (Nov 19) Europe is being held back by a precautionary reflex that treats inaction as the safest choice, even as the costs of standing still rise sharply.
★ The $100 trillion productivity puzzle (Jun 19) The capability-absorption gap: labs racing ahead faster than firms can absorb. Why macro-productivity numbers remain flat despite AI breakthroughs.
Did OpenAI’s $100 billion path just get narrower? (Dec 4) Sam Altman’s internal pivot and what it tells us about the AI market right now.
Unpicking OpenAI’s real revenues (Nov 14) How the misreading of cash flows is spooking investors.
Learning & accountability
★ Seven lessons from building with AI (Apr 16) Practical takeaways from automating our work with AI, and a stack of 50+ tools that we have used.
What I was right (and wrong) about (Jul 4) Marking your own work. AI capability, solar records, EVs, workforce impact. What held up and what didn’t.
★ The cheat codes of technological progress (May 31) What still governs technological advance. Revisiting the ‘laws’ of progress. How, when and why they work (or don’t).
The exponential thinker’s reading list: 10 timeless essays (Sep 10) A collection of essays that explain how and why technologies reshape the world.