Inference Will Be $10TRN Market: What No One Sees | Software Investors are F***ed | Naveen Rao
Summary
The inference market will reach $10 trillion within 10 years, but energy—not software—is the real bottleneck. Building one gigawatt costs $50B and takes years; the US has only 3-5 years of excess grid capacity left before hitting a hard infrastructure ceiling that will constrain AI deployment.
Key Takeaways
- Energy infrastructure is the critical constraint, not chip availability. One gigawatt costs $50B to build—more expensive and complex than a semiconductor fab ($20-30B)—and the US can realistically build only ~10 gigawatts annually ($500B investment required).
- The US has 2-3 years before exhausting excess grid capacity. Current US power grid runs at ~900 GW average with 1.2-1.3 TW peak capacity; once this buffer depletes, AI growth becomes bottlenecked by infrastructure build rate, not demand.
- China built energy infrastructure 6x faster than the US over 20 years and surpassed total US capacity in 2023. However, centralized planning advantages don't translate to innovation; the US should leverage its edge in frontier tech while fixing policy barriers to infrastructure deployment.
- Hardware, not software, is where value accumulates over the next decade. Pure software investors face structural headwinds as inference becomes commoditized and energy/hardware constraints dominate economics and defensibility.
- Inference will be a $10 trillion market within 10 years with infinite demand, making energy availability the primary constraint on market size capture. Winners will control access to power and compute, not software layers.
Related topics
Transcript Excerpt
We're going to be out of energy in the next few years. This is a blip. It's not some kind of bubble being burst. We don't treat the model as sacrian. And we're energy first in our mindset. Naveen Ralph is a serial entrepreneur. I mean, he sold his last business to data bricks for $1.3 billion. Mosaic ML. Now he's back with unconventional AI, which has raised hundreds of millions of dollars. Their last round close to $5 billion in valuation. Navine is one of the most articulate and clear thinkers across such a wide range of very complicated topics. He was fantastic and this show was a joy to do. Intelligence to me will have infinite demand. I actually think the hardware is the place to be over the next 10 years. I would say if you're a pure software investor, you're kind of [ __ ] The infer…
More from 20VC
- The New VC Game | CoreWeave Sell-Off Creates Panic | Revenue Multiples Out of Control
- The Future of Inference Market & Model Routing: Commodity or Increasing in Value? | LangChain CEO
- Atlassian CEO: Enterprises Won't Trust Chinese AI | Why The Jury’s Still Out on Specialised Models
- How to Build a Sales Machine | Max Freeman, Ramp
- Crusoe CEO: Why Everyone Gets GPU Depreciation & AI Energy Costs Wrong