News

AI Researchers Walk Away From Bezos-Backed Prometheus to Build Physical-World AI

AI researchers Anima Anandkumar and Benedikt Jenik have launched a new company, Accelerated Understanding, after choosing not to join Bezos-backed Project Prometheus. Their venture aims to build large-scale models of the physical world, with potential applications in chip design, robotics, weather forecasting and energy.

Written By : Soham Halder
Reviewed By : Pranchal Srivastava

Two artificial intelligence researchers who were once approached to lead Jeff Bezos-backed Project Prometheus have chosen a different path, launching their own company to develop technology focused on understanding the physical world rather than language.

Anima Anandkumar and Benedikt Jenik, the co-founders of Accelerated Understanding Inc, unveiled their new venture on Tuesday. The company said its system can process as many as 5 trillion pieces of data in a single prompt, which is roughly five million times the typical capacity of flagship models from companies such as Anthropic and Google.

Different Approach to AI

Unlike popular systems such as ChatGPT, which are largely built around processing and generating language, Accelerated Understanding is developing technology designed to model physical phenomena across space and time.

The company's approach uses neural operators, a technology Anandkumar helped advance during her earlier research. Rather than relying on the Transformer architecture widely used in today's leading language models, the system is designed to work with physics-based information.

Anandkumar, a professor at Caltech, described the concept as a shift from a language-focused view of intelligence towards a nature-centric approach.

Why Does the Physical World Matters?

The founders believe their technology could have applications well beyond scientific research. Potential commercial uses include chip design, robotics, extreme-weather forecasting and energy exploration.

For semiconductor companies, for example, the technology could help model how materials and temperatures affect chip performance. In weather forecasting, it could potentially help predict complex physical events without relying on separate models for every scenario.

The company plans to initially focus on enterprise customers, rather than launching a consumer-facing product. 

"The language-centric view of intelligence is humans at the center. Putting physics at the center is a nature-centric view," said Anandkumar, a Caltech professor of computing and mathematical sciences.

Why They Turned Down Bezos-Backed Prometheus?

The decision to build independently came despite an attractive proposal from Project Prometheus, which Bezos co-founded with investor and biotech entrepreneur Vik Bajaj. According to reports, Anandkumar and Jenik were offered a combined 35% stake in Prometheus, along with an annual salary package that could have risen to $2 million. The proposal also outlined more than $2 billion in committed financing through Series B.

The pair ultimately continued building their own company. Prometheus later raised $12 billion in a Series B round and is pursuing technology to automate the manufacturing of complex physical systems.

Also Read: Humanoids, AMRs, and Physical AI: Future of Robotics in Real Environments

From Nvidia to New AI Bet

Anandkumar's background includes five years as a director at Nvidia, where she worked on research involving GPUs and advanced AI applications. Nvidia CEO Jensen Huang encouraged her to pursue the broader idea, she told new agencies.

With Accelerated Understanding now entering the market, the company is betting that the next major step in AI could come not from processing more text, but from better understanding the laws governing the physical world.

Join our WhatsApp Channel to get the latest news, exclusives and videos on WhatsApp

Crypto News Today: Bitcoin Inflows, XRP Active Addresses Jumps, Zcash Reached 8-year High

11 Crypto Trading Content Marketing Strategies That Build Trust and Drive Traffic

Ethereum vs. Tokenized Assets: Why Institutional Adoption Could Strengthen ETH's Role in Finance

XRP Supply Dilution is Rising While XRPL Revenue Falls: What it Means for XRP's Long-Term Value

Bitcoin vs. Sovereign Debt: Can Fiscal Stress Strengthen BTC's Role as a Hedge?