Skip to content
a16za16z

Why AI Demand Is Outrunning Compute Supply

a16z’s David George sits down with Gavin Baker to unpack the state of the AI boom, why demand for intelligence may still be dramatically underestimated, and why the outcome doesn't necessarily have to be winner-take-all. David and Gavin explore the possibility that frontier labs, open-source models, applications, clouds, and NVIDIA can all capture significant value as AI adoption expands. They dig into the economics of the infrastructure buildout, why compute investments can have unusually fast payback periods, and what happens when today's relatively small group of heavy AI users expands to hundreds of millions of people. They also debate the risk of an AI bubble versus an AI shortage, the backlash against data centers, orbital compute, the rise of multi-model architectures, and NVIDIA's position at the center of the AI supply chain. Gavin makes the case that the AI buildout could help reindustrialize America, while David explores whether the bigger near-term risk is not overbuilding, but failing to build enough. Timestamps: 00:00 - Intro 01:06 - Finding the Bear Case: Why Gavin Can't Find One 08:06 - How's This All Gonna Go Wrong? An "And" Thing, Not an "Or" 09:09 - Will Labs Reinvest All Their Profits Into Training Forever? 14:44 - The Demand Side: 30 Million Heavy Users & the Diffusion Question 19:16 - From Reactive Coding to Fully Autonomous Agents 30:18 - What Happens If There's a Massive Supply Shortage? 34:21 - Orbital Data Centers: The SpaceX Compute Play 40:00 - Starlink's $2 Trillion Market & the Heads-You-Win Compute Bet 44:04 - The Most Futuristic SpaceX Idea: Asteroid Mining 54:01 - Who Becomes the Abstraction Layer of Intelligence? 57:57 - Harvey, Cursor & Vertical AI Winners 01:00:26 - Jensen, Nvidia & the Central Bank of AI 01:12:16 - How Chip Deal Structures Reveal True Customer Preference Resources: Follow Gavin Baker on X: https://x.com/GavinSBaker Follow David George on X: https://x.com/DavidGeorge83 Stay Updated: If you enjoyed this episode, be sure to like, subscribe, and share with your friends! Find a16z on X: https://twitter.com/a16z Find a16z on LinkedIn: https://www.linkedin.com/company/a16z Listen to the a16z Show on Spotify: https://open.spotify.com/show/5bC65RDvs3oxnLyqqvkUYX Listen to the a16z Show on Apple Podcasts: https://podcasts.apple.com/us/podcast/a16z-podcast/id842818711 Follow our host: https://x.com/eriktorenberg Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details please see http://a16z.com/disclosures.

Gavin BakerguestDavid Georgehost
Aug 31, 20261h 14mWatch on YouTube ↗

At a glance

WHAT IT’S REALLY ABOUT

AI adoption is exploding faster than power, chips, and data centers

  1. Gavin Baker and David George argue that AI demand is accelerating across labs, open source, and applications even as public AI equities have drawn down, making the real risk persistent compute undersupply rather than a near-term bubble pop.
  2. They contend the AI buildout has unusually strong economics—often sub‑one‑year paybacks—supported by prepayments, spot-market monetization, extended useful lives, and sophisticated project financing.
  3. They describe a demand-side story still in its infancy: token spend is concentrated among a small set of heavy users today, but diffusion to 1.5B knowledge workers plus agentic automation could multiply consumption dramatically.
  4. They warn that regulation and local opposition to data centers could create ‘compute inequality’ and slow societal benefits, while emphasizing data centers’ role in reindustrializing parts of the U.S. through jobs and tax bases.
  5. They explore ‘future capacity’ bets—especially SpaceX orbital compute enabled by Starship reusability—and predict a multi-model enterprise future where the key battle is who becomes the abstraction layer for intelligence across organizations.

IDEAS WORTH REMEMBERING

5 ideas

AI is more likely to be supply-constrained than overbuilt through the late 2020s.

They argue the market is still early in real adoption (tens of millions of heavy users vs ~1.5B knowledge workers), while physical constraints (power, chips, copper, permitting, politics) limit buildouts—creating a sustained shortage dynamic rather than an imminent glut.

Public-market AI financials may be unusually volatile because revenue is partly a compute-allocation choice.

Unlike prior tech waves where serving more users didn’t force a direct infrastructure trade-off, frontier labs can materially swing reported revenue by reallocating scarce power/compute between inference (monetized) and training/research (future capability).

Compute infrastructure can justify massive capital deployment because payback periods are exceptionally short.

They cite disclosures and unit-economics heuristics suggesting gigawatt-scale deployments can pay back in under a year, aided by customer prepayments, spot pricing, and improving utilization/asset life—plus the ability to finance large portions of the build.

Autonomous agents imply potentially ‘endless’ token consumption, not just incremental productivity gains.

As companies move from “reactive” copilots (summaries, assist) toward autonomous agents that plan and execute workflows, token usage becomes closer to an open-ended operational input—driving much higher ongoing demand.

The AI industry’s biggest near-term risk is political/regulatory friction, not technical feasibility.

They claim data centers can dramatically increase local tax revenue and jobs, and that some popular critiques (especially water use) are overstated; they also warn that blocking buildouts could produce ‘compute inequality’ where only large firms/wealthy users can afford frontier access.

WORDS WORTH SAVING

5 quotes

When the history of the 21st century is written, you know, there was like the Victorian age. I think this will be like the age of Elon and Jensen because they are fundamentally altering the fabric of human society and civilization.

Gavin Baker

This is not an or thing, it's an and thing, right? Like, this is an and thing.

David George

I just, in my career as an investor, there haven't been that many opportunities where you have companies that could deploy tens, hundreds of billions of dollars and get sub one year paybacks.

Gavin Baker

The only person who can, the only group that can tell the AI, AI industry's truth is the AI industry. They need to just start telling the truth.

Gavin Baker

An increasing fraction of the world's compute is gonna be in orbit.

Gavin Baker

Compute supply constraints vs overbuild cyclesTraining vs inference allocation trade-offsToken-demand diffusion and heavy-user concentrationAutonomous agents and rising token consumptionData centers, regulation, and local economic impactOrbital data centers and Starship reusability economicsNvidia ecosystem, financing, and ‘central bank’ role

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.