Staera Semiconductor emerged from stealth on Tuesday, November 11, 2026, with a direct answer to the biggest bottleneck in AI infrastructure: moving data between chips. The Singapore startup, formerly known as New Silicon Corporation (NSC), announced a $4.6 million pre-seed round to develop Lightwire, led by Openspace Capital, with participation from SGInnovate, Black Kite Capital, 28 Holdings, In Group and Decimus Technologies.
The company's first product is Lightwire, an optical interconnect built entirely inside the chip. Instead of attaching lasers or LEDs next to the processor, Staera fabricates the light source and the whole optical link on the silicon wafer itself, combining silicon CMOS circuits, GaN (gallium nitride) LEDs and silicon nitride waveguides that carry light between links.
Why light, and why now
According to the company, AI's limit is no longer how fast chips compute, but how fast and far they can move data. Copper's performance falls a thousandfold between one millimeter and one meter — exactly the distances that connect chips, boards and racks. The International Energy Agency (IEA) expects data center electricity use to more than double, from about 415 terawatt-hours in 2024 to around 945 terawatt-hours by 2030, and copper cannot carry more data over longer distances without burning more power.
The shift toward light is already reaching the industry's biggest buyers: Nvidia, which said in 2025 it would “stay with copper for as long as [they] can,” now plans to use both copper and optics in its AI systems. Lightwire was designed for that mix.
Lightwire inside the chip
Lightwire is built inside the bridge or interposer chip that already connects processors and memory in today's AI hardware, so it works with existing packaging. According to Staera, the technology moves data using less than a tenth of the energy of today's best optical links: moving 24 terabytes per second would take under 50 watts with Lightwire, compared with nearly a kilowatt using co-packaged optics. These figures are the company's own projections, with no published independent validation yet.
The company says it has built GaN LEDs into silicon chips in commercial foundries since 2016, with processes that now run at 98 to 99 percent yield. “AI's bottleneck is no longer how fast chips compute, but how fast and far they can move data,” said Staera CEO Dr. Andrew Kim in the announcement.
From the lab to customers
The pre-seed round marks the shift from research to commercial product. Staera will use the funding to advance the technology, expand its team into the US and deepen engagement with customers and supply-chain partners, targeting commercial release in 2030 — a timeline stated by the company itself, not a date confirmed by customers or foundries.
The technology has already drawn industry recognition: in May 2026, the company won the 3rd ECTC Start-up Competition at the IEEE Electronic Components and Technology Conference, and placed runner-up at the 2026 Lam Capital Venture Competition, run by Lam Research in Fremont, California.
Staera began as a spin-off from the Low Energy Electronic Systems (LEES) research group at the Singapore-MIT Alliance for Research and Technology (SMART) — MIT's first research centre outside the United States.
Context
The bet on light for computing also showed up in the Q.ANT and IONOS photonic computing pilot, already covered by Geekiki. The focus differs: while that project uses light to compute, Lightwire uses light to connect — a distinct problem inside the same data center.
Featured image: illustration of Staera's technology, combining CMOS circuits, GaN LEDs and waveguides in a single silicon chip. Image: Staera, via EIN Presswire.
Sources
- EIN Presswire — Staera Semiconductor emerges from stealth with US$4.6 million to break AI's data bottleneck
- Staera Semiconductor — staera.ai
Transparency: this article was produced with the assistance of artificial intelligence, cross-checking public posts on X and sources on the internet. Consult the original sources to verify the details.
By Muse
