Explore Naveen Rao’s $475M bet with Unconventional AI. Learn how the MosaicML co-founder is rethinking chip design and energy efficiency for deep neural networks.

We are currently using numeric, deterministic digital systems to simulate something—the neural net—that is fundamentally non-numeric and non-deterministic. If we want to keep scaling AI, we can't just buy more GPUs; we have to change how the silicon actually processes information.
Create a cinematic, conversational audio lesson based on the interview with Naveen Rao of Unconventional AI. Follow the user's specific structural and analytical requirements: Lead with the counterargument regarding Nvidia's dominance; treat the company's thesis as a hypothesis; tag claims by confidence levels; and compare digital matrix multiplication to Rao's approach of exploiting silicon physics. Focus on the energy crisis in AI, the Landauer-limit claim (9-10 orders of magnitude gap), and the 'speedrunning research' strategy. Use the user's requested narrative arc: history → AI concentration → energy → physics of intelligence → technical risk. Ensure the style is intellectually rigorous, skeptical, and accessible to a layperson. Include the specified four closing questions and the final philosophical prompt. Attached source: 'The $475M Bet to Build 1,000x More Efficient Compute Naveen Rao, Unconventional AI.txt'



Naveen Rao, the co-founder and CEO of Unconventional AI, is making a $475 million bet that current AI infrastructure is hitting a physical wall regarding energy. Having previously built Nirvana and MosaicML, Rao argues that the industry is relying on foundations not originally designed for the massive power requirements of modern deep neural networks. His new venture focuses on rethinking chip design to handle the unprecedented shift where the majority of global computing is becoming a single workload.
Naveen Rao is a prominent figure in the AI industry known for his deep expertise in chip design and machine learning infrastructure. He successfully founded and sold Nirvana to Intel and later sold MosaicML to Databricks. With his new company, Unconventional AI, Rao is leveraging his experience from the inside of the 'sausage making' of hardware to challenge the current dominance of companies like Nvidia by focusing on the physics of energy efficiency.
Unconventional AI focuses on energy efficiency because Naveen Rao believes we are hitting a physical energy wall as deep neural networks become the dominant computing workload. Rao points out that we are moving toward a world where 80% or more of all computing is dedicated to this single task, which is unprecedented in computing history. This shift requires a departure from traditional CPU-based diversity toward hardware specifically optimized for the physics of neural nets.
From Columbia University alumni built in San Francisco
"Instead of endless scrolling, I just hit play on BeFreed. It saves me so much time."
"I never knew where to start with nonfiction—BeFreed’s book lists turned into podcasts gave me a clear path."
"Perfect balance between learning and entertainment. Finished ‘Thinking, Fast and Slow’ on my commute this week."
"Crazy how much I learned while walking the dog. BeFreed = small habits → big gains."
"Reading used to feel like a chore. Now it’s just part of my lifestyle."
"Feels effortless compared to reading. I’ve finished 6 books this month already."
"BeFreed turned my guilty doomscrolling into something that feels productive and inspiring."
"BeFreed turned my commute into learning time. 20-min podcasts are perfect for finishing books I never had time for."
"BeFreed replaced my podcast queue. Imagine Spotify for books — that’s it. 🙌"
"It is great for me to learn something from the book without reading it."
"The themed book list podcasts help me connect ideas across authors—like a guided audio journey."
"Makes me feel smarter every time before going to work"
From Columbia University alumni built in San Francisco
