AgileRL secures £5.5m to cut reinforcement learning costs using distributed training
This article covers AgileRL, a London-founded SaaS startup that has closed a growth funding round of £5.5m to develop and commercialise tools to make reinforcement learning cheaper and faster for commercial teams. The funding will support a San Francisco office and hiring and aims to lower barriers to enterprise adoption of reinforcement learning, targeting commercial teams and enterprises across sectors such as finance and robotics.
AgileRL, a London-founded SaaS startup, has closed a growth funding round of £5.5 million to develop and commercialise tools that make reinforcement learning cheaper and faster for commercial teams. The funding will support a new San Francisco office and hiring across engineering and go-to-market as the company pushes its open-source platform and managed product, Arena, towards wider enterprise use.
Why it matters
Reinforcement learning can unlock decision-making and automation use cases that supervised models struggle with, but building production-grade RL is costly and complex. Companies often need dedicated research teams, specialised simulators, large compute and bespoke pipelines, which keeps the technology inside a handful of large tech firms.
AgileRL’s pitch is to lower those barriers with a plug-and-play stack that companies can adopt without rebuilding the whole research infrastructure. The company says its tooling cuts training time and cost by around 10x, which, if realised in production settings, could broaden RL adoption beyond deep-pocketed incumbents.
Inside the product
AgileRL offers a free open-source platform alongside Arena, a managed, full-stack RLOps product. The platform supports on-policy, off-policy, offline, multi-agent training, contextual multi-armed bandits and integration with large language models. Key technical claims include evolutionary hyperparameter optimisation, distributed multi-GPU training, environment validation and one-click deployment to production.
The project has seen more than 300,000 downloads and lists engineers at organisations such as JPMorgan, Wayve, IBM, Huawei and Forster & Partners among users. Those names indicate interest from financial institutions, autonomous vehicle labs, large corporate R&D teams and architecture consultancies, suggesting a range of potential enterprise use cases from trading and controls to simulation-backed design workflows.
In the announcement, Andrew Nestor, Machine Learning Engineer at Decision Lab, said:
Arena has significantly streamlined our RL development workflow, making training and deploying agents a breeze. The platform’s hyperparameter tuning capabilities have dramatically accelerated our experimentation and improved the performance of our models.
Who are the investors?
The round brings AgileRL’s total funding to £5.5 million. It was led by Fusion Fund with participation from Flying Fish, Octopus Ventures, Entrepreneur First and Counterview Capital.
Investors backing the round are placing a bet on tooling that reduces the engineering overhead of reinforcement learning and packages research primitives into deployable workflows. The mix of backers includes venture firms and a notable early-stage operator (Entrepreneur First), reflecting interest in both commercial potential and technology-led teams.
If you're researching potential backers in this space:
- Explore our list of saas venture capital firms in the UK
- Browse our directory of saas angel investors in the UK
Founder perspective
In the announcement, Param Kumar, co-founder of AgileRL, said:
Having built a reinforcement learning system from scratch at my last company, I saw firsthand how costly and complex it is. No company in 2026 should need a full AI research lab just to use RL. Our goal is to make reinforcement learning a standard tool in every company’s tech stack.
AgileRL says the fresh capital will fund a San Francisco office and more than a dozen hires across engineering and go-to-market functions, signalling a move to serve US customers while continuing to develop its open-source community and managed product.
The bigger picture
The funding reflects growing interest from SaaS investors in infrastructure that lowers the barrier to advanced AI techniques such as reinforcement learning. If AgileRL’s claims about speed and automation hold up in enterprise pilots, the company could help expand RL beyond research labs into practical applications across finance, robotics, simulation and other sectors.
For the UK ecosystem, the deal is another example of London-founded AI toolmakers securing transatlantic ambitions and investor support for developer-focused products that aim to commercialise academic advances.
They've invested in AgileRL
Get to know these saas investors
| Investor | Sector | Stage | Activity | Team | Connect |
|---|---|---|---|---|---|
 Fusion Fund | 3 investments investments | more info | |||
 Flying Fish | 2 investments investments | more info | |||
 Octopus Ventures | 45 investments investments | 14 contacts contacts | |||
 Entrepreneur First | 14 investments investments | 10 contacts contacts | |||
 Counterview Capital | 1 investment investment | more info |
Click here for a full list of 7,526+ startup investors in the UK