UK chip startup Olix raises $312M at $3.3BN valuation
Tech.eu John Reynolds ● Covered by 2 sources
UK chip startup Olix just raised $312m, tripling its valuation to $3.3bn in months. It's betting specialised inference chips can dethrone Nvidia's one-size-fits-all approach.
Based on reporting by Tech.eu, John Reynolds — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Two years is not a long time to build a chip company, let alone one investors now value at $3.3bn. But that is where Olix, founded by 25-year-old James Dacombe, finds itself after closing a $312m Series B. The number alone is striking: it triples the valuation Olix carried just months ago, when it raised $220m in February at just over $1bn.
The backers say plenty about how seriously the market is taking this. Fundomo, a New York VC, led the round, with Arm and Hudson River Trading joining in, plus angel money from Netflix co-founder Reed Hastings. Existing investors, including the UK government's Sovereign AI fund, all increased their stakes rather than sitting this one out. Olix also brought in Matt Briers, the former Wise CFO who took that company public, and added Nick McKeown, a Stanford professor and former Intel executive, to its board.
What Olix is actually building is an optical digital processor with what it calls a novel memory and interconnect architecture, designed to run current AI models and handle inference workloads. Dacombe, who is simultaneously running CoMind, a brain monitoring and treatment startup, framed the pitch to the FT in blunt terms: Nvidia's era of handling every AI chip workload is ending, and the industry is heading into what he called a world of specialists.
Olix's own explanation leans on a factory metaphor. Producing a single AI token, the company argues, requires hundreds of operations, each with different hardware demands, yet the industry keeps running every stage through the same general-purpose chip. Each new generation just makes that generalist chip a bit better rather than building something purpose-made for any single stage. Olix's bet is that splitting those stages across specialised chips will produce a meaningful jump in performance and cost efficiency, rather than incremental gains.
The company says it is in the final stages of designing and testing its chips, with first customer deliveries expected next year. Rather than selling bare silicon, Olix plans to package its chips into a server rack alongside its own software and networking gear. The fresh funding, it says, will go toward getting those chips to market, building out the broader silicon platform, and locking down manufacturing and supply chain commitments — the unglamorous but essential work of actually shipping hardware at scale.
My take — AI-written commentary, not fact-checked reporting
A two-year-old company hitting a $3.3bn valuation before it has shipped a single chip says more about investor appetite for anything that might dent Nvidia than about Olix's chips themselves. The specialist-versus-generalist argument is genuinely interesting, and the roster of backers and hires is unusually credible for a startup this young. Still, the real test isn't the funding round, it's whether those chips land with customers next year as promised — plenty of ambitious chip architectures have looked brilliant on paper and stalled at manufacturing.
Read more about this at: Tech.eu