AI infrastructure specialist OLIX has secured $312 million in Series B funding, strengthened its leadership team, and advanced plans to commercialise its next-generation AI inference platform.
London-based AI infrastructure company OLIX has raised $312 million in a Series B funding round, valuing the business at $3.3 billion just two years after its founding.
The funding marks another milestone for Founder James Dacome as OLIX prepares to bring its first AI inference platform to market and expand its global engineering footprint.
The financing attracted new investors including Fundomo, Arm, Hudson River Trading, and angel investor Reed Hastings, co-Founder of Netflix. Existing backers Hummingbird Ventures, Crane, Plural, Creandum, Phoenix Court, and Transition also increased their commitments, signalling continued confidence in the company’s long-term vision.
Alongside the investment, OLIX has strengthened its leadership with two high-profile appointments. Professor Nick McKeown has joined the board of directors, whilst former Wise CFO Matt Briers has been named Chief Financial Officer.
McKeown is recognised as one of the pioneers of modern networking, having co-invented software-defined networking, OpenFlow, and P4. A Professor Emeritus at Stanford University and recipient of the 2025 Marconi Prize, he has also founded several networking companies, including Nicira and Barefoot Networks.
OLIX believes the current approach to AI inference hardware is approaching its efficiency limits.
Rather than relying on increasingly powerful general-purpose processors, the company has developed an architecture that assigns specialised chips to individual stages of the AI inference process. By fully distributing AI models across multiple chips, its X-1 platform creates a production line where each processor performs a dedicated task whilst retaining the flexibility to support evolving model architectures.
At the heart of the system is a proprietary optical interconnect that transfers data between chips using light rather than copper, reducing both latency and energy consumption. Combined with a deterministic compiler that schedules workloads across multiple racks, OLIX believes the approach will improve AI inference performance whilst reducing operational costs.
The first processor in the platform, DX-1, has been designed specifically for the decoding stage of AI inference, where models generate responses. The company says the accelerator is capable of delivering more than 10,000 tokens per second per user for 100-billion-parameter models, whilst maintaining higher output token throughput per watt than conventional general-purpose chips.
DX-1 has also been engineered with scalability and supply chain resilience in mind.
The processor stores AI models in fast on-chip SRAM memory to improve energy efficiency and reduce latency. It also avoids the use of advanced packaging and high-bandwidth memory, two components that remain in particularly short supply across the semiconductor industry.
According to OLIX, the architecture has been designed to scale efficiently to AI models exceeding 10 trillion parameters, supporting increasingly demanding inference workloads as frontier AI continues to evolve.
The new funding will support the delivery of DX-1 to OLIX’s first customers during the second half of 2027, whilst accelerating development of the wider custom silicon platform underpinning its AI infrastructure.
Investment will also fund manufacturing capacity and supply chain commitments needed to scale production, alongside continued recruitment across silicon, photonics, compiler, and systems engineering teams in London, Bristol, Austin, Toronto, and San Francisco.
Founded in London in 2024, OLIX designs every element of its AI inference systems, from chips and optical networking technology to complete rack-scale infrastructure. With fresh capital, expanded leadership, and commercial deployment on the horizon, the company is positioning itself to support the next generation of frontier AI workloads.
