American Marconi Wireless News - Opticore Raises $14 Million to Advance Photonic Computing Chips 100x More Energy Efficient Than GPUs

Enabling energy-efficient AI with breakthrough advances in compute efficiency

BERKELEY, CA / ACCESS Newswire / September 10, 2025 / Opticore secured $7.5 million in additional financing to accelerate the development of its breakthrough optical computing technology. This funding builds on the company's pre-seed round last year to bring its total funding to $14.5 million.

Opticore is building a next-generation photonic computing chip that is up to 100 times more energy efficient and offers 25 times the computing density compared to leading GPUs. "Energy-intensity of existing electronics-based AI compute solutions is straining electrical grids, capping the scale-out of data centers. Scaling AI requires fundamentally new approaches to computing efficiency," said Prashant Shukla, partner at Origin Ventures, which co-led Opticore's seed round alongside Jetha Global. Karan Danthi, CIO at Jetha Global, added "we see the silicon photonics roadmap unfolding in stages: first switches and co-packaged optics, then optical interconnect solutions, and ultimately full optical processing."

Co-founder Zaijun Chen explains, "our photonic computing architecture isn't constrained by Moore's Law and can achieve world-class performance while consuming a fraction of the energy budget. We envision a future where data centers won't need their own power plants, and the heaviest compute workloads can be offloaded to photonic processors, shaping the multi-trillion-dollar AI infrastructure buildout."

"The ability to produce better-than-leading-edge CMOS performance on lagging nodes opens up latent capacity in semiconductor fabrication - and all in the United States" adds Opticore's first investor, Alex Turnbull of Sagax Capital. Opticore's financing also included investment from Neotribe Ventures and Thunderbolt Ventures.

Opticore's patented approach leveraging time-multiplexed computing enables Opticore to encode a trillion parameters on a single chip while achieving 100 TOPS/W compute efficiency through parallel computation. Unlike other optical approaches with high energy overhead and limited scalability, Opticore uses dynamic weights for real-time training and inference. The technology is compatible with standard foundry processes, making it highly scalable for commercial production. Following successful demonstration tapeouts this year, the company expects to have its first scaled system demonstration in 2026.

About Opticore:

Opticore's technology is built with exclusive patents from groundbreaking academic work: co-founder Dr. Ryan Hamerly and advisor Prof. Dirk Englund proposed the theoretical basis in Physical Review X (2019), while CEO and co-founder Dr. Chen demonstrated the core principles in Nature Photonics (2023). The latest validation of Opticore's approach is in Science Advances (2025).

For more information about Opticore, please visit https://www.opticore.ai/ and for an outline of the company's technology and vision please see our post "Hurdling the Power Wall With Photonic Computing".

For media inquiries, please contact Adam Cai [email protected]

SOURCE: Opticore

View the original press release on ACCESS Newswire

P.Silva--AMWN