Key facts
- Volantis raised $88 million in venture capital.
- The company is developing technology to connect AI chips and memory chips using laser light.
- This technology aims to allow 220 memory chips around a GPU, compared to eight with current methods.
- Volantis plans to deliver a chip next year to speed up AI coding.
- The funding round was led by Lachy Groom and Abstract Ventures.
San Francisco-based semiconductor startup Volantis announced on Thursday that it has secured $88 million in venture capital funding. The company aims to address a critical bottleneck in AI chip performance by developing a novel technology that enhances communication between computing and memory chips.
Currently, AI chips from companies like Nvidia and AMD are limited by the number of high-bandwidth memory chips they can connect to, due to the reach of electrical wires. Volantis' approach utilizes vertical-cavity surface-emitting lasers (VCSELs), the same technology found in features like Apple's facial recognition, to transmit data. This method is expected to allow for significantly more memory chips—up to 220 per GPU—compared to the eight currently supported by Nvidia's best offerings.
Volantis CEO and co-founder Tapa Ghosh stated that by integrating relatively known components, the company hopes to circumvent supply chain issues and deliver a chip next year that can accelerate tasks such as AI coding. The funding round was led by Lachy Groom and Abstract Ventures, with participation from investors including John Doerr, VXI Capital, Triatomic, and Susa Ventures, as well as angel investors like Dwarkesh Patel and Naveen Rao.