Atoms to Applications: The AI Stack, and AI Investor
Research outfit SemiAnalysis dropped a bombshell earlier this year, calculating the total spend for the AI buildout, from 2024 through to 2029, at bit more than eleven trillion US dollars. A number so big it would be laughable, if two trillion of that hadn't already been spent, with much more on the way.
The technology of artificial intelligence generated a shockwave in capital. So much capital, raised and spent so rapidly, warps everything else in the economy, because all of this is happening at multiple levels, simultaneously: in matter, energy, semiconductors, models, harnesses, and applications.
"Atoms to Applications" is the only frame wide enough to capture the size and scope of the economic transformation underway.
Walking the layers of this new 'stack', seven posts illuminate how this tsunami of capital has changed everything it encounters - while also highlighting things that can't be changed for all the money in the world. It's a map of opportunities - and warnings:
$11,000,000,000,000: The world is set to spend about $11 trillion on AI data centres between 2024 and 2029. That is so much money, so fast, that it is already pushing up memory prices, electricity demand and interest rates across the whole economy.
The Heart of the Matter: AI can grow only as fast as the supply of a few obscure materials such as helium, bromine and rare earths, which come from very few places, take years to scale up, and are at the mercy of wars and trade disputes.
Thought Energy: Each AI answer now uses a tiny fraction of the power it once did, yet demand is rising so fast that power plants and grids have become the main brake on AI. Data centres could strengthen the grid instead of straining it if they were required to bring their own clean energy.
Semiconductors and Data Centres: Nvidia looks like the king of AI, yet real control sits with the handful of firms that make chips and memory, chiefly TSMC, ASML and the memory makers. Their sold-out factories decide who gets computing power and are making phones and PCs dearer for everyone else.
This Year’s Model: AI models are the only part of the stack getting cheaper, because open and smaller models catch up to the frontier within months, and that makes it hard for the big closed labs to justify their huge valuations.
In Harness: A raw AI model only predicts the next word. The real progress is now in the “harness,” the software wrapped around a model that gives it memory, tools and goals, and a good or bad harness can make the same model dramatically better or worse.
Growing Applications: The whole AI buildout depends on applications people will pay for. Coding and finance already justify the spending, while AI agents threaten traditional software businesses and are starting to run whole companies for a fraction of the usual cost.
On the 11th of December, my old friend John Allsopp and I bring that stack into the room at the inaugural AI Investor, a half-day briefing for investors, founders and strategy leaders, organised around one goal: helping you understand where AI is creating and destroying value, and what will shape the market next.
As readers of The Watershed, it's my pleasure to be able to offer you discounted tickets to AI Investor. Just use the code WATERSHED or click on this link. You'll also receive a year of Pro-level subscription to Noops Research - collecting and sharing the most important signals for anyone investing in AI.
Have a read through 'the stack' - and we hope to see you at AI Investor!
Mark