Episode summary
Naveen Rao: 4D Computing, AI's Energy Wall & Beating Biology
All-In Podcast
· · 23 min · Hosts: Chamath Palihapitiya · Guests: Naveen Rao
Summary
Naveen Rao argued that AI is running into an energy wall and that the industry’s next bottleneck is no longer chips or rack space, but power. He estimated Google alone processes 3.2 quadrillion tokens a month, and used that scale to argue that AI services could soon consume far more electricity than the current data-center system can supply. In his view, about 50% of the cost of serving a token is energy, so the central business problem is how to monetize every watt far more efficiently than today’s hardware allows.
His answer is a new kind of computer built around physical dynamics rather than the usual von Neumann split between memory and compute. Rao said Unconventional AI is drawing on neuroscience, dynamical systems, and sparse physical circuits to create “4D computing,” where time becomes part of the machine. He pointed to the brain’s efficiency, especially animal brains, as proof that intelligence can be far more power-efficient than GPUs.
Rao said the company built its first physical dynamical computer in five months, taped out the chip on June 1, and has already generated images on it. He claimed the system is roughly 500 nanojoules per image, far below a GPU, and said the team expects a full product in about two years, with customers porting model layers rather than rewriting everything from scratch.
Key takeaways
- Rao said AI compute is hitting an energy constraint, with Google’s 3.2 quadrillion monthly tokens implying roughly 12 gigawatts of power demand.
- He argued that about 50% of token-serving cost is energy, so the winning hardware will be the one that monetizes every watt better.
- Unconventional AI’s thesis is “4D computing,” a dynamical-computer approach that fuses memory and compute instead of moving data back and forth.
- Rao said the company built its first physical dynamical computer in five months and taped out the chip on June 1.
- He expects a full rack product in about two years and says existing models will work after porting at the model layer, not the operations layer.
Notable quotes
“I think AI is one of the most transformational technologies that humanity's ever created and will enable us to get to that next level of evolution”
“we're going to run out of energy pretty fast in like 3 years or so is my estimate”
“about 50% of the cost of serving a token. So every time you try something on chat GBT 50% of that cost is energy”
“This is actually the first physical dynamical computer ever built”
“We want to make something better and enable you know compute everywhere including compute in new robotic um forms”
How the episode opens
Naveen Ralph, co-founder and CEO of Unconventional AI, which is an AI chip startup. — Best known for building and selling two deep tech companies. Naveen is kind of definitionally outlier founder. When I came there, we had about a $20 million business and it was, you know, 7 or 800 million when I left.
I don't think you really understand something until you can build it. — Just because something is tried does not mean it's wrong. — I'm the opposite of an AI goomer. I think AI is the next evolution of humanity. We need innovation on the hardware substrate to actually build true intelligence.
Please welcome Naveen Ralph. Hey everyone, great to be here. Um, you know, switching gears a little bit to AI now, which you may have heard a little bit about.
Um, it's super exciting to be at this conference specifically because as was said in the uh intro, I'm the opposite of a doomer. I think AI is one of the most transformational technologies that humanity's ever created and will enable us to get to that next level of evolution which I'm here for and uh this is sort of the anti-doomer conference. So, let's go.
Um, so before we get going, I'll tell you a little bit about myself. Um, you know, I'm it's kind of weird. I'm really right where I wanted to be my whole life.
Uh this was me at about five or six years old, something like that. We had a computer very early on, so I'll date myself, but uh this was in 1978. We got a computer.
This is probably in the early 80s. Uh learned to program when I was a little kid. I just thought it was like a puzzle.
Um you know, learned to I became an electrical engineer.
From the episode’s captions: roughly the first two minutes, lightly cleaned. Keep watching on YouTube →
Topics
- ai energy costs
- ai compute and power contracts
- 4d computing
- dynamical systems
- model deployment and product building
Watch the episode
Watch “Naveen Rao: 4D Computing, AI's Energy Wall & Beating Biology” on YouTube
Summary, takeaways and quote selection by Octagon. Quotes are short excerpts; the full conversation is in the episode.
More from All-In Podcast
-
Freedom vs. Safety: Axon CEO Rick Smith on the Future of Policing -
Midterm Blood Bath, $4T AI Sell Off, Data Center Freeze, What’s Next for Trump -
All-In Podcast REACTS to Democrat Sweep! -
Anthropic S-1 Shocker, Greatest Business Ever? Google’s $500B Wipeout, Midterm Predictions -
WHOOP CEO, Will Ahmed: 50B Hours of Health Data, a 4 Million Member Data Moat, & the FDA Fight -
AI Cyber Panic, Ken Griffin’s Bubble Warning, The Fall of France?
All-In Podcast, this episode and its artwork belong to All-In Podcast, LLC. Octagon is not affiliated with the show; this page is an independent summary.
