Company
Forlais at the company level. The institutional overview, careers, and a public statement of who Forlais is and what Forlais does.
Forlais is a British frontier AI research and development company building new models, systems and infrastructure for artificial intelligence.
Our work spans model architecture, memory, autonomous systems, compression, reliability, software platforms and specialised computing. Through this research, we are developing technologies designed to make AI more capable, efficient, reliable and accountable.
What we do: Research, developed into working technology.
We research and develop new ways for artificial intelligence to learn, represent information, retain knowledge and operate across different environments.
This work has produced a broad portfolio of architectures, models and systems, including floatless AI, persistent memory, autonomous research systems, lossless compression and specialised computing designed around the requirements of advanced AI.
Our research is developed into working technology, with each programme contributing to the wider direction of Forlais and the systems we are bringing to market.
Our research: Original research across several areas of artificial intelligence.
Forlais conducts original research across several areas of artificial intelligence. Our programmes explore:
Model architecture and representation
Integer-native and floatless computation
Persistent and self-organising memory
Autonomous systems
Lossless model compression
Reliability and auditability
Structural learning
AI infrastructure and specialised hardware
Together, this work is creating a new technical foundation for the models, products and platforms being developed by Forlais.
CNA: One of Forlais' most important advances.
Our floatless AI architecture, designed natively around a new form of computation and developed without conventional floating-point operation.
CNA is one of Forlais' most important advances in artificial intelligence.
Foundation: Provides the foundation for a new class of models.
Central to: Model efficiency, reliability, auditability and operating cost.
Strengthened by: Research programmes across Forlais contribute knowledge, methods and systems to CNA and the technology built around it.
Carried into: An important part of the technology behind EvaEsi, bringing a new class of AI systems into the frontier market.
EvaEsi: Where research becomes a platform.
EvaEsi is the primary route through which Forlais will bring its research to market.
It provides a way for us to develop the work contained across our research portfolio into scalable models, products and services that are more precise, reliable and accountable.
EvaEsi is not separate from the research conducted at Forlais. It is where that research becomes a complete, accessible and commercially deployable AI platform.
AI products for consumers, developers and organisations across: Web, Desktop, Mobile, Enterprise, API.
Reliability: Frontier AI capability with stronger reliability.
Memory: Persistent memory carried across the platform.
Accountability: Clearer accountability in how systems answer.
Operating cost: Lower operating requirements by design.
A growing research portfolio: Across intelligence, learning, memory, autonomy and computing.
Forlais is developing a growing portfolio of research programmes across intelligence, learning, memory, autonomy and computing.
These include CNA, Athena, Hermes, Orion, Sagas, Zi, Genesis, Eros, Eos, Adriane, Nammu × Argus, Ouah, FCN, Odysseus, the Icarus Files and Talos.
Across these programmes, we are developing models, memory systems, autonomous architectures, compression technologies and new forms of AI infrastructure.
Each programme investigates a different part of the wider problem. Together, they create the technical layers from which Forlais can build complete AI systems.
Our work goes beyond model development alone.
Talos: Specialised computing for Forlais technology.
Forlais is building the software, platforms, infrastructure and hardware needed to support the technologies emerging from our research.
Talos extends this work into specialised computing.
It is a new processing chip being developed around the requirements of Forlais technology and future AI systems, with the aim of making advanced AI practical in local and offline environments.
Talos is being designed as part of a wider computing architecture rather than as an isolated component. Its purpose is to support the models, systems and infrastructure emerging from our research with greater efficiency, control and independence.
Model architecture, Memory, Platforms, Infrastructure, Specialised hardware, Talos.
By working across every layer, from model architecture and memory to platforms, infrastructure and specialised hardware, Forlais can develop AI as a complete system rather than treating each part as a separate problem.
Where we are going: The intelligence, the platform and the infrastructure, as one connected system.
Forlais is preparing to bring its technology to market through EvaEsi while continuing to expand its research portfolio.
Our experimental work has been running since 2025, producing the findings, systems and technical direction that now underpin the company.
The next stage is to bring these parts together.
Our direction is to make Forlais increasingly self-reliant across the full AI technology stack, from research and architecture through to models, products, infrastructure and specialised computing.
This gives us greater control over how our technology is designed, deployed and improved. It also allows us to pursue substantial operating efficiencies and enter the market with a fundamentally different economic model.
Forlais is building towards a future in which the intelligence, the platform and the infrastructure are developed as one connected system.
That future is now within reach.
No idea too strange.
We chase ideas wherever they lead. The thoughts most people dismiss out of hand are often the ones worth chasing hardest.
The hard part is building it.
The limit is rarely the idea. It is turning the whole of a blueprint into code that can carry it, without losing what made it powerful.
Find the penicillin.
We watch for the anomaly that does not fit. The detail most would tidy away is often the door to something greater.
Truth above all.
We do not overclaim or hide. A result that looks miraculous is a reason to test harder, not to celebrate.