ChatGPT uses ~10x more energy than a Google search, and AI demand is surging. EnCharge is tackling the energy crisis with analog in-memory computing, cutting power use by up to 20x vs. today’s chips. No more data shuttling. Just fast, efficient AI.

AI could consume 500 TWh/year by 2027. That's almost like adding a new France to the grid! At #WebSummitVancouver on May 29, our CEO Naveen Verma explores how rethinking chip architecture can make AI more accessible and sustainable. Join us: vancouver.websummit.com

At this year’s Web Summit in Vancouver, our CEO Naveen Verma & @CerebrasSystems' Naor Penso explored why current chip architectures can’t sustainably handle AI’s computational demands and the critical need for specialized hardware in a session moderated by the The Guardian's Dara Kerr.

Traditional chips quickly hit a wall when trying to deliver AI within the confines of a laptop. But, as highlighted by EE Times | Electronic Engineering Times, we’ve cracked the code with our flagship EN100 AI accelerator. eetimes.com/encharge-picks…

With AI increasingly taking over critical workflows and vital activities, security is a fundamental component of AI’s utility. At Web Summit Vancouver, our CEO Naveen Verma and @CerebrasSystems' Naor Penso discussed the impact of security in all aspects of #AI architecture.

💡 Empathy doesn’t mean going easy on people. Naveen Verma EnCharge AI , Princeton professor turned startup CEO , has learned that being an empathetic leader sometimes means making hard calls and making them quickly. In this clip, he shares why compassion in leadership still

Proud to see our co-founder & CEO Naveen Verma take the stage as a featured innovator at #CPI2025. Thank you Princeton University for spotlighting our mission to enable a new era of advanced AI at the edge.

Today’s GPUs waste 95% of inference energy moving data between memory and compute. We're eliminating that commute with analog in-memory compute that processes data where it lives, for 20x efficiency gains and a dramatic drop in data center power demand. bain.com/insights/can-t…

The U.S. is losing ground on the AI-defined battlefield, not for lack of models, but for lack of ultra-efficient chips to run them at the tactical edge. The tech is here. Our analog IMC makes this real. What’s missing isn’t innovation, it’s deployment. csis.org/blogs/strategi…

The Sept 27th All-In Podcast covered a paper in Nature Computational Science claiming 7,000× speedups and 70,000× energy efficiency gains. Caveats? It’s not that simple, and it’s not that new. Read more: open.substack.com/pub/aiafterhou…