Solving the AI Energy and Silicon Complexity Crises with Physics-Based Computing

Description

​As AI’s energy demands reach critical levels with today’s semiconductors, the industry faces unprecedented challenges. Consider this stark reality: a single ChatGPT session consumes 150 times more power than the human brain uses to perform all of its functions. Hyperscalers are dramatically expanding capital investment into cloud infrastructure. Larry Ellison, for example, notes unprecedented demand “we have to build 100 additional cloud data centers…we already got building permits for three nuclear reactors". The semiconductor industry – and hence future of AI – is at a critical crossroads.

​At Normal Computing, we’re building AI to help design the next generation of chips that power drastically more efficient AI – to tackle the AI energy and silicon complexity crises. Physics-based computing represents a fundamental shift in our approach to energy-efficient computation. 

​We invite you to join an exclusive afternoon workshop, featuring leading voices from research, industry, and government agencies for meaningful dialogue on physics-based computing and frontier algorithms research, reducing AI training energy consumption by 1000x, and addressing constraints in advanced semiconductor design, testing, and verification to enable sustainable scaling of computational capabilities.

​Featured Speakers include

• ​Dr. Benjamin Weiner, Fellow at the Advanced Research Projects Agency-Energy (ARPA-E)

• ​Dr. Suraj Bramhavar, Programme Director at the Advanced Research + Invention Agency (ARIA)

• ​Dr. Ahmad Beirami, Google DeepMind

• ​Syrus Ziai, Founder of Eliyan Corporation

• ​Dr. Logan Wright, Professor at Yale University

• ​Dr. Steve Fu, Partner at Celesta Capital

• ​Dr. Patrick Coles, Chief Scientist at Normal Computing

• ​Dr. Gavin Crooks, Senior Research Scientist at Normal Computing

​Space is limited to maintain an atmosphere conducive to substantive discussion and networking.

​Hosted in partnership with NYC Deep Tech Week.