Olix Raises $312M for AI Inference Chips at $3.3B Valuation

London-based deep tech innovator Olix has successfully closed a substantial $312 million Series B funding round, propelling its valuation to $3.3 billion. This significant capital infusion underscores the intense global competition to develop specialized hardware for artificial intelligence inference, the process of running AI models in real-world applications.

The latest funding round saw participation from a high-caliber group of investors, including chip design giant Arm, quantitative trading firm Hudson River Trading, and a notable contribution from Netflix co-founder Reed Hastings. The round was reportedly led by New York-based deep tech investment firm Fundomo, which also participated in Olix's previous $220 million Series A round earlier this year. Existing backers such as Hummingbird Ventures, Crane, Plural, Creandum, Phoenix Court, and Transition all increased their stakes.

Adding strategic weight to the investment, the UK government's state venture fund, Sovereign AI, also joined the round. This marks Olix as the fifth company to receive an equity investment from the fund, highlighting the UK's commitment to fostering domestic AI capabilities. AI Minister Kanishka Narayan emphasized the strategic importance of domestic chip development, while Sovereign AI Head of Ventures Joséphine Kant drew a connection to Britain's historical innovation legacy through the company's holding name, Bletchley Industries.

Olix is addressing perceived inefficiencies in current AI inference hardware by developing its proprietary X-1 platform. This system moves away from general-purpose chips, instead distributing AI models across numerous specialized chips. Each chip handles a specific segment of the model, connected via a novel optical interconnect that utilizes light for data transfer. The company's initial offering, the DX-1 decode accelerator, is designed to significantly boost token generation speeds for large language models, promising superior throughput per watt compared to conventional hardware, particularly for models up to 100 billion parameters and scalable to trillions.

A key differentiator for Olix's DX-1 chip is its ability to hold model parameters in fast on-chip SRAM, crucially avoiding the need for advanced packaging and high-bandwidth memory (HBM), components currently facing industry-wide supply constraints. This strategic design choice could provide a significant advantage in the current hardware market dynamics.

The company is also bolstering its leadership team with the appointment of two prominent figures. Nick McKeown, a renowned expert in networking and co-founder of successful ventures like Nicira and Barefoot Networks, joins the board of directors. Additionally, Matt Briers, former CFO of fintech firm Wise, where he guided the company through its landmark 2021 direct listing, has been appointed as the new Chief Financial Officer.

This funding round arrives amidst a surge of investment in AI chip startups, which collectively raised approximately $8.3 billion globally in 2026 alone, according to Dealroom data. The focus has shifted from model training to inference, driven by the escalating computational and energy demands of AI deployment in everyday applications. Competitors like Nvidia, Cerebras, and d-Matrix have also seen substantial capital injections, underscoring the intense race for market leadership in this critical technology sector.