Staera Semiconductor Exits Stealth With $4.6M Pre-Seed

Staera Semiconductor has emerged from stealth with $4.6 million in pre-seed funding to build optical interconnects directly inside silicon chips, betting that AI computing's bottleneck is now data movement, not computation. Openspace Capital led the round, with SGInnovate, Black Kite Capital, 28 Holdings, In Group and Decimus Technologies participating. The company, formerly named New Silicon Corporation, is targeting commercial product availability in 2030.

The claim at the center of the raise is bold but simple: replace copper wiring between chips with light, and build the entire optical link into the chip itself. "AI's bottleneck is no longer how fast chips compute, but how fast and far they can move data," said Dr. Andrew Kim, CEO of Staera Semiconductor. "Copper performance plummets over the millimeter-to-meter distances that matter most in AI accelerators, and bolt-on optics add power and complexity. We build the light into the chip, the entire optical link inside a chip, so the connection gets smaller, cooler and faster with each generation."

What problem is Lightwire solving?

AI accelerators keep gaining raw compute, but performance now depends heavily on how quickly and cheaply data moves between processors, memory and other components. Conventional copper interconnects face limits on signal integrity, power consumption and distance. Optical links offer an alternative, yet today's optical interconnects typically use separately manufactured light sources and other components attached to semiconductor packages — added complexity that Staera wants to eliminate.

Its Lightwire technology combines CMOS circuits, gallium nitride LEDs and silicon nitride waveguides, fabricated together using processes the company says are compatible with existing semiconductor manufacturing facilities. Lightwire is designed to operate within bridge or interposer chips that connect processors and memory, positioning it to slot into current advanced packaging approaches rather than replace them.

What do the numbers say — and what don't they say?

The performance targets are aggressive. Staera is aiming for energy consumption below 0.2 picojoules per bit, compared with roughly five picojoules per bit for certain co-packaged optical technologies — a claimed 25-fold improvement. The architecture is also designed to support more than 600 optical links per millimeter of chip edge. The company estimates that a system moving 24 terabytes of data per second could need under 50 watts through Lightwire, versus nearly one kilowatt with alternative co-packaged optical systems — the difference between a light bulb and a space heater, per-household framing aside.

Investors and readers should weigh one caveat the source states plainly: these figures are design targets and internal performance comparisons, not independently established commercial deployment results. Initial links support four to eight gigabits per second each, and Staera also envisions future high-bandwidth memory configurations that would raise aggregate throughput substantially. In plain words: the physics looks promising on paper, and the market will not know until silicon ships.

How credible is the manufacturing story?

The company is not starting from zero. Staera has worked on integrating gallium nitride LEDs into silicon through commercial foundries since 2016 — nearly a decade of process development. It claims manufacturing yields of 98% to 99%. That figure is self-reported and unaudited, but a decade-long foundry track record is more than most pre-seed hardware startups can show.

The technology has also drawn external recognition, albeit of the competition variety. In May 2026, Staera won the startup competition at the IEEE Electronic Components and Technology Conference and finished runner-up in the Lam Capital Venture Competition.

What happens next?

The funding will support continued technical development, hiring, and engagement with semiconductor customers and supply chain partners. "This funding takes Lightwire from the lab to customers, and the Staera name marks that move," Kim said.

Hian Goh, co-founder of Openspace Capital, framed the bet in infrastructure terms: "Staera is addressing one of AI infrastructure's most urgent challenges: moving vast amounts of data between chips without a corresponding rise in power, heat and complexity. We backed Staera because of the strength of its technology, the depth of its team and its potential to define a new category in semiconductor connectivity."

The so-what is straightforward. If AI data center power demand keeps climbing, connectivity efficiency becomes a procurement decision, not an engineering nicety — and a company that can prove sub-0.2-picojoule links at foundry yields by 2030 would enter a market desperate for exactly that.