OLIX, a custom silicon startup building hardware for AI inference, has raised £240m in a Series B round at a £2.4bn valuation, two years after it was founded. The round was led by Fundomo with participation from Arm, Hudson River Trading and angel investors including Reed Hastings, the co-founder of Netflix. Existing investors Hummingbird Ventures, Crane, Plural, Creandum, Phoenix Court and Transition all increased their commitments.
The industry's approach to building AI inference hardware is reaching its efficiency limits, according to OLIX. A datacentre is a factory whose product is the token, and producing a single token takes hundreds of operations, each placing different demands on hardware. Any other factory would give each stage a machine built for it, yet the token factory runs every stage on the same general purpose chip, with each new generation delivering an even better generalist but never a specialist. OLIX's systems instead use specialised chips for each stage of the token production process. Within the X-1 platform, models are fully unrolled across a large number of chips, creating a production line that allows each chip to focus on a single part of the model, while every chip retains a flexible compute fabric rather than baking in a particular model's architecture.
Holding the system together is a novel 'slow and wide' optical interconnect that moves data directly between chips using light instead of copper, at ultra-low latency and energy cost, made possible by rack-scale codesign across every part of the link. Workloads are scheduled across racks by a fully deterministic compiler. The first chip OLIX is building within X-1 is DX-1, a decode accelerator designed for the stage at which a model reasons and generates its output. For 100B parameter size models, DX-1 can achieve the pareto optimal inference performance, delivering simultaneously over 10,000 tokens per second per user at higher output token throughput per watt than general purpose chips running large batch sizes, and the architecture scales performantly to models of 10 trillion parameters and above thanks to a multi-rack scale-up domain architecture. By keeping a model in fast on-chip memory, SRAM, DX-1 uses no advanced packaging and no high-bandwidth memory, the components the industry is in shortest supply of.
Alongside the round, Professor Nick McKeown joins the board of directors. One of the defining figures of modern networking, McKeown co-invented software-defined networking, OpenFlow and P4, is Professor Emeritus of Computer Science and Electrical Engineering at Stanford University and winner of the 2025 Marconi Prize, and co-founded several companies including Nicira, acquired by VMware, and Barefoot Networks, acquired by Intel, where he went on to lead Intel's networking business. Matt Briers has also been appointed Chief Financial Officer after nine years as CFO of Wise, which he joined in 2015 from Google when it had around 500,000 customers and was loss-making. Briers built the finance function that carried Wise from private company to the public markets while it remained profitable throughout, taking it through a 2021 direct listing on the London Stock Exchange, the first by a technology company in London and the exchange's largest technology listing, valuing Wise at £8.75 billion, approximately $12 billion.
The Series B funds the path to deliver DX-1 to OLIX's first customers by H2 2027 and the build-out of the wider custom silicon platform behind it, together with the manufacturing and supply chain commitments that scaling frontier inference hardware requires. Hiring is underway across silicon, photonics, compiler and systems engineering in London, Bristol, Austin, Toronto and San Francisco.
The future of AI will be built on chips that power models. Countries that build chips will build leverage. Olix is exactly the kind of ambitious company we want to back through Sovereign AI. In just two years, it has established itself as one of the UK’s most exciting AI startups, developing breakthrough chip technology with the potential to help shape the future of AI.
