xmemory pre-seed £3m to improve AI reliability using a write-time memory layer
This article covers xmemory, a London-based AI startup, which has raised £3m ($4m) in a pre-seed funding round to develop a "precision-first" memory layer to improve the reliability, governance and accuracy of AI systems. The funding will be used to move the startup out of closed beta, run pilot programmes with design partners and refine its platform for enterprise use, supporting organisations deploying AI that require traceable, high-precision outputs from internal tools, agents and decision systems.
xmemory, a London-based AI startup, has raised £3m ($4m) in a pre-seed funding round to develop a “precision-first” memory layer aimed at improving the reliability, governance and accuracy of AI systems. The funding will be used to move the company out of closed beta, run pilot programmes with design partners and refine its platform for enterprise use.
Why it matters
Organisations deploying AI for internal tools, agents and decision systems increasingly complain about inconsistent or untrustworthy outputs. Much of that stems from how data is stored and retrieved. xmemory’s raise is notable because it targets the infrastructure problem upstream of model queries: improving the quality of data at write time rather than relying on retrieval-augmented generation at query time. If successful, that change could make AI systems easier to audit and govern and reduce risky reliance on semantic search for routine operational questions.
Inside the product
xmemory is building a memory layer that structures, cleans and organises information as it is written, so later reads return higher-quality, more consistent answers. The company says this approach could allow organisations to answer up to 95% of critical operational and business questions with high precision, reserving broader semantic search for only the most complex queries. The platform is currently in closed beta; the new funding is earmarked for product development and pilots with early design partners to validate the workflow in real-world settings.
Who are the investors?
The round was led by Fly Ventures, Begin Capital, AAL VC, 33East and Inovia Capital. Angel investors and industry mentors also participated, including Vadim Barshtak, former CTO of Miro, alongside Mark Shmulevich, Alexey Dosovitskiy and Konstantin Vinogradov.
Investors are backing xmemory’s focus on data reliability and governance for enterprise AI, betting that a disciplined memory layer can reduce downstream risk and integration costs for organisations building production AI systems.
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 by Alex Petrov, xmemory positions its work as infrastructure for organisations that need traceability and precision from their AI systems. The company is prioritising pilot engagements with customers whose use cases require strict accuracy and audit trails, such as internal automation, knowledge management and decision-support tools.
The bigger picture
The raise underlines continued interest in tooling and infrastructure that support enterprise AI adoption across the UK and Europe. As organisations move from experimentation to production, there is growing demand from AI investors for technologies that reduce model risk and make outputs auditable. xmemory’s pre-seed round joins a broader wave of startups focused on data-layer solutions rather than model-only improvements.
The funding also highlights how early-stage capital in the UK is being deployed to address practical barriers to adopting AI in regulated and operational contexts, a trend likely to continue as companies and public-sector organisations look for safer, more governable deployments.
They've invested in xmemory
Get to know these AI investors
Click here for a full list of 7,526+ startup investors in the UK