Wise OLIX Secures $312M Series B at $3.3 Billion Valuation

Just two years after its founding, OLIX has secured $312 million in Series B funding at a $3.3 billion valuation, signaling rapid investor confidence in its approach to artificial intelligence inference hardware. The company believes current AI chip design is hitting efficiency limits and proposes specialized chips for each stage of processing; its X-1 platform fully unrolls models across numerous chips, creating a focused production line. Alongside the funding, OLIX appointed Professor Nick McKeown, co-inventor of software-defined networking, to its board of directors, and Matt Briers, former CFO of Wise, to the role of Chief Financial Officer.

OLIX Secures $312m Series B with Nick McKeown Appointment

OLIX’s recent $312 million Series B funding round values the company at $3.3 billion. The financing included participation from Fundomo, Arm, and Hudson River Trading, along with increased commitments from existing investors like Hummingbird Ventures and Crane, demonstrating broad support for OLIX’s vision. Reed Hastings, the co-founder of Netflix, contributed as an angel investor, extending interest beyond traditional venture capital within the technology hardware space.

This capital will directly fund the delivery of OLIX’s DX-1 decode accelerator to initial customers by the second half of 2027, alongside the expansion of its custom silicon platform and the necessary manufacturing and supply chain infrastructure. The company explains that producing a single token takes hundreds of operations, each placing different demands on hardware, advocating for a production line approach where each chip focuses on a single part of the AI model.

This strategy aims to unlock a step change in AI performance and cost, particularly as model architectures rapidly evolve. McKeown, a defining figure in modern networking and co-inventor of software-defined networking, OpenFlow, and P4, brings substantial expertise to the company. He is Professor Emeritus of Computer Science and Electrical Engineering at Stanford University and a 2025 Marconi Prize winner, having previously co-founded Nicira, acquired by VMware, and Barefoot Networks, acquired by Intel.

His leadership at Barefoot Networks saw him lead the company’s networking business, suggesting a strong ability to translate innovation into commercially viable products. OLIX is also bolstering its operational leadership with the appointment of Matt Briers as Chief Financial Officer.

Briers joins from Wise, where he served as CFO for nine years, successfully guiding the company from a loss-making startup with around 500,000 customers to a 2021 direct listing on the London Stock Exchange, valuing Wise at approximately $12 billion. OLIX highlights that he built the finance function that carried Wise from a private company to the public markets while it remained profitable throughout, reflecting OLIX’s ambition for a similar trajectory towards profitability and public markets.

The company’s X-1 platform utilizes an optical interconnect, moving data between chips with light instead of copper to reduce latency and energy consumption, facilitated by rack-scale codesign and a deterministic compiler. OLIX believes this will make existing frontier AI more affordable and abundant and, more importantly, unlock the deployment of far more powerful models in the future.

DX-1 Decode Accelerator and Rack-Scale Optical Interconnects

OLIX is challenging conventional approaches to artificial intelligence inference hardware with a focus on specialized silicon and a novel interconnect architecture. Rather than continually scaling general-purpose chips, the company proposes a system built around dedicated processing stages for each part of the AI, mirroring the assembly line approach of other manufacturing industries. For 100 billion parameter models, the company reports DX-1 can achieve over 10,000 tokens per second per user, exceeding the throughput per watt of general-purpose chips operating with large batch sizes.

This performance is enabled by retaining the model in fast on-chip SRAM memory, a design choice deliberately avoiding reliance on high-bandwidth memory or advanced packaging, components currently facing significant supply chain constraints. The architecture is designed to scale to models exceeding 10 trillion parameters, leveraging OLIX’s multi-rack scale-up domain architecture.

OLIX is currently hiring engineers in London, Bristol, Austin, Toronto, and San Francisco, focusing on silicon, photonics, compiler, and systems engineering.