Interconnect for the
AI era.
Vega Links is building next-generation high-speed interconnect beyond 200G for the AI datacenter — spanning the full 0–50m reach, from inside the package to across the rack.
The AI buildout runs on interconnect.
We are building its next generation.
Inside every AI rack is an exploding web of high-speed links. Keeping that fabric fast, reliable, and efficient is one of the hardest — and most valuable — problems in the datacenter.
A defining bottleneck
AI compute scales only as fast as the fabric that connects it. Every new generation of accelerators multiplies the interconnect inside the rack — and bandwidth, reach, and power are now the limiters.
See the market →Built for what comes after 200G
The industry has scaled copper from 28G to 224G, but the next step strains reach and power budgets. We are engineering the interconnect that carries the ecosystem through its next generations.
Our approach →A team that has shipped this before
Founded by a wireline-interconnect architect, with a team and advisors drawn from the companies that defined high-speed SerDes and optical DSP.
Meet the team →We are hiring, and talking to investors.
If you build at the edge of analog, mixed-signal, and DSP — or you back the teams that do — we should talk.