Archive · 26 Aug 2026
Anima Anandkumar Bets on Physics-Driven AI Breakthrough
Anima Anandkumar is a computer science professor at Caltech who is working on a different kind of artificial intelligence.
Instead of focusing mainly on words, her company wants AI to understand how the physical world works.
The physical world follows rules involving numbers, movement, and scientific laws.
Her company, Accelerated Understanding, is using a technology called neural operators for this goal.
These systems may learn about one physical problem and use that knowledge to help solve others.
The company says its system handled 5 trillion data points in one prompt.
Anandkumar previously worked at Amazon Web Services and NVIDIA.
She and Jenik reportedly chose to build their own company instead of joining Jeff Bezos-backed Project Prometheus.
Anima Anandkumar, a Caltech computing professor, is developing physics-focused AI rather than another language model.
She previously worked at Amazon Web Services and NVIDIA, where she led research into advanced AI applications using GPUs.
Her company, Accelerated Understanding Inc., argues that language-based AI cannot fully represent continuous variables, mathematical relationships, and physical laws.
The company is developing neural operators designed to learn from physical systems and apply that knowledge across different problems.
Accelerated Understanding claims its system processed 5 trillion data points in a single prompt, while Anandkumar and Jenik reportedly declined an offer from Jeff Bezos-backed Project Prometheus.
- Who
- Anima Anandkumar, her company Accelerated Understanding Inc., and collaborator Jenik.
- What
- They are developing physics-focused AI based on neural operators rather than building another language-centered chatbot.
- When
- The articles do not specify when Accelerated Understanding Inc. was founded or when its claimed data-processing demonstration occurred.
- Where
- Anandkumar is a professor at the California Institute of Technology in California; her earlier education and work included institutions in India and the United States.
- Why
- Anandkumar believes language-based AI has limitations because physical systems involve continuous variables, mathematical relationships, and physical laws.
- Current role
- Bren Professor of Computing and Mathematical Sciences at the California Institute of Technology.
- Education
- Bachelor’s degree in electrical engineering from IIT Madras in 2004 and PhD from Cornell University in 2009, followed by postdoctoral research at MIT.
- Previous positions
- Professor at UC Irvine, principal scientist at Amazon Web Services, and senior director of AI research at NVIDIA.
- Company
- Accelerated Understanding Inc.
- Core technology
- Neural operators, intended to help AI learn from physical systems.
- Reported processing claim
- Accelerated Understanding says its system processed 5 trillion data points in a single prompt.
- Alternative opportunity
- Anandkumar and Jenik reportedly declined a lucrative offer to join Jeff Bezos-backed Project Prometheus.