aiexpert
Home / News / Brief
Funding · Aug 20, 2026, 10:04 PM · 3 sources

Olix raises $312M at $3.3B, optical interconnect for token-factory AI inference

UK AI chip startup Olix has raised $312 million in a Series B round at a $3.3 billion valuation, more than tripling its value from the $1 billion post-money it achieved in February. The round was led by Fundomo and includes Arm, Hudson River Trading, and angel investor Reed Hastings (Netflix co-founder), with existing investors increasing commitments. Olix, founded in 2024 by 25-year-old James Dacombe, is developing specialized inference processors with optical interconnects—a radical departure from Nvidia's general-purpose approach.

The core thesis: AI inference is a 'token factory' with hundreds of heterogeneous operations; forcing all stages through one general-purpose chip wastes efficiency. Olix's X-1 platform unrolls models across specialized chips connected by optical links using light instead of copper, and crucially avoids high-bandwidth memory (HBM)—the costliest, most supply-constrained component in modern accelerators. The company claims its DX-1 decode accelerator achieves >10k tokens/sec/user for 100B-parameter models while maintaining superior energy efficiency and dodging HBM supply bottlenecks.

First customer shipments are planned for H2 2027. The funding is strategic: Arm's participation signals ecosystem commitment, the UK government's Sovereign AI fund backed the round, and Stanford's Nick McKeown (SDN pioneer) joined the board. For architects evaluating inference capacity alternatives, Olix represents a genuine architectural bet on disaggregated, photonic-first silicon—not just another GPU clone.

Sources

Everything this brief rests on
  1. 01 Primary source tech.eu
  2. 02 tech.eu tech.eu “The new $3.3bn valuation marks more than a tripling of its just over $1bn valuation when it raised $220m in February this year”
  3. 03 olix.com olix.com “Olix designs every part of its systems: the chips, the lasers and the network that connects them”