Unconventional AI
Funding History
| Seed · 12/2025 | $4.5B |
|---|
About
Unconventional AI designs analog and mixed-signal computer chips intended to run AI models using far less energy than conventional digital processors, targeting model types such as diffusion, flow, and energy-based models whose probabilistic behavior maps naturally onto analog physics.
- totalFunding
- $475M
- postMoneyValuation
- $4.5B
- founded
- 2025
- headquarters
- San Francisco, California
Overview
Unconventional AI was founded in 2025 by Naveen Rao, Michael Carbin, Sara Achour, and MeeLan Lee to rethink the physical hardware that AI models run on. Rather than simulating neural network math on general-purpose digital chips, the company designs analog and mixed-signal circuits — using oscillators, thermodynamic effects, and spiking-neuron-style dynamics — that store and manipulate probability distributions directly in silicon physics.
The company closed a $475 million seed round in December 2025 at a $4.5 billion valuation, one of the largest seed rounds on record, only a couple of months after launching. Leadership has said the round is a first installment toward a larger financing goal of up to $1 billion as the company builds out multiple candidate computing architectures.
In June 2026 the company released Un-0, an image-generation model built on a software simulation of its oscillator-based architecture, intended to show the approach can match state-of-the-art diffusion models like Stable Diffusion while pointing toward large future power savings once run on dedicated chips.
Private securities are speculative, illiquid, and involve risk of complete loss of principal. You should be prepared to hold private securities for an extended period; there is no assurance of a secondary market, an IPO, or other liquidity event. You are fully responsible for conducting your own due diligence. Pricing is indicative — nothing here is an offer of securities.