Fractile raises £160m to deliver chips reducing latency and boosting memory bandwidth
This article covers Fractile, an AI startup, raising £160m in a growth funding round to accelerate delivery of its first inference chips and systems to customers. The funding is intended to help the startup address inference latency and memory bandwidth constraints for long-context AI workloads, supporting higher-throughput and real-time applications in areas such as software engineering automation, drug discovery and materials discovery.
Fractile, an AI startup building inference hardware for frontier AI systems, has raised £160 million in a growth funding round to speed the delivery of its first chips and systems to customers. The raise matters because Fractile is targeting a technical bottleneck — inference latency and memory bandwidth — that is becoming a limiting factor as large models and very long-context workloads scale up.
Why it matters
Generative and long-context AI models are moving beyond training constraints into operational limits on inference. Fractile says some workloads now need outputs of up to 100 million tokens, and that on current chips an output of that length can take about a month at roughly 40 tokens per second. Bringing those workloads down to a day would require output speeds of around 1,200 tokens per second. If those figures hold, improvements in inference hardware could materially expand what applications are practical in areas such as software engineering automation, drug discovery and materials discovery.
Inside the product
Fractile develops chips and integrated systems aimed at reducing latency and increasing capacity for long-sequence AI inference. The company combines work across AI research, foundry process innovation and chip micro-architecture to tackle memory bandwidth limits and the complexity of running very large models over long contexts. The firm says its technology is intended to address three core challenges: complexity, latency and capacity when operating frontier models.
Fractile positions faster inference as enabling longer-context real-time interactions and higher-throughput batch workloads that current architectures struggle to handle.
Who are the investors?
The £160 million round was led by Accel, Factorial Funds and Founders Fund, with participation from Conviction, Gigascale, O1A, Felicis, Buckley Ventures, 8VC and existing investors. Fractile said the proceeds will be used to accelerate the path to getting its first chips and systems into customers' hands and to scale engineering and operations.
The mix of venture capital names spans established early-stage investors and firms known for hardware and AI bets, signalling continued investor interest in companies building specialised inference solutions rather than software alone.
If you're researching potential backers in this space:
- Explore our list of AI venture capital firms in the UK
- Browse our directory of AI angel investors in the UK
Founder perspective
Founded in 2022, Fractile has since worked across multiple technical fronts: AI model research, chip micro-architecture and foundry process collaboration. The company is hiring across London, Bristol, San Francisco and Taipei as it moves from development toward customer deployments.
The bigger picture
The round underscores a broader shift in the AI ecosystem where funding is flowing into hardware and systems startups that promise to alleviate inference bottlenecks. As models grow in size and context length, memory bandwidth and latency become priority constraints alongside model architecture and optimisation.
This raise also highlights the UK’s role in advanced semiconductor and AI systems work: Fractile’s hiring in London and Bristol links the company to the UK hardware and research base, even as it operates internationally. The result is another example of how UK and European AI-related startups are attracting large rounds to tackle infrastructure-level problems in the global AI stack.
They've invested in Fractile
Get to know these AI investors
| Investor | Sector | Stage | Activity | Team | Connect |
|---|---|---|---|---|---|
 Accel | 40 investments investments | 8 contacts contacts | |||
 Factorial Funds | 1 investment investment | more info | |||
 Founders Fund | 4 investments investments | more info | |||
 Felicis Ventures | 5 investments investments | 8 contacts contacts | |||
 Buckley Ventures | 2 investments investments | 2 contacts contacts | |||
 8VC | 6 investments investments | 31 contacts contacts |
Click here for a full list of 7,589+ startup investors in the UK