Volantis Secured $88 Million for Optical Chip Interconnects
The San Francisco startup aims to break AI memory bottlenecks by using lasers to connect hundreds of memory chips.
Updated on Oct. 1, 2026 in Semiconductors
Live Poll
Do you believe new AI chip developments will significantly improve the performance of your consumer electronics?
San Francisco-based semiconductor startup Volantis has raised $88 million in venture capital to develop new hardware for AI workloads. The firm is researching a laser-based communication system designed to overcome current limits in data transfer between computing and memory chips.
Why it matters
Current AI hardware faces severe connectivity limitations, as bandwidth constraints between processors and memory often throttle performance. Volantis is attempting to solve this by replacing traditional electrical connections with optical hardware to increase data throughput.
The startup claims its technology enables up to 220 memory chips to be connected to a single GPU, a massive increase over the eight memory chips supported by current standard Nvidia hardware.
The players
Volantis
A San Francisco-based semiconductor startup researching laser-based optical interconnects for high-performance computing.
Lachy Groom
A venture investor who co-led the $88 million funding round for Volantis.
Abstract Ventures
A venture capital firm focused on early-stage technology investments that co-led the Volantis financing.
Nvidia
A global semiconductor designer currently dominating the AI hardware market with GPU architectures that typically support eight memory chips.
The details
Volantis utilizes VCSELs — vertical-cavity surface-emitting lasers, a type of semiconductor laser diode that emits light vertically from the chip surface — to transmit data between computing and memory silicon. By swapping electrical traces for optical links, the company aims to move data faster than traditional copper pathways allow. This development remains in the research and prototyping stage as the firm scales its optical architecture.
Timeline
October 1, 2026: Volantis announced the $88 million funding round.
2027: The company expects to deliver a chip capable of accelerating AI coding tasks.
The Tech Race
Volantis is competing directly with established electrical interface standards that currently limit how much memory can communicate with an AI processor. By testing optical alternatives, the company is attempting to surpass the data-shuttling ceiling inherent in current Nvidia GPU designs.
The technology is currently in development and will not affect end-user hardware or AI software performance until at least 2027. Developers and hardware engineers should watch for upcoming technical disclosures regarding power efficiency and compatibility with standard GPU sockets.
The takeaway
Volantis is tackling the primary bottleneck in modern AI compute by scaling memory connectivity via light. Watch for the company's 2027 product delivery to see if optical interconnects can move from theoretical research into viable AI hardware performance benchmarks.
Further reading
For broader trends in chip architecture, see Semiconductors.
Live Poll
Do you believe new AI chip developments will significantly improve the performance of your consumer electronics?
