5 Minutes
Lithuanian startup Nexos.ai has closed a €30 million Series A round to tackle one of the trickiest problems in corporate tech: how to let employees use powerful AI tools without leaking sensitive data. Backed by Index Ventures and Evantic Capital, the platform positions itself as a neutral control layer between teams and large language models.
Why enterprises are nervous — and why Nexos.ai sees opportunity
Most companies understand AI’s potential but worry about governance, compliance and uncontrolled data exfiltration. Rather than outright bans, Nexos.ai proposes a different path: act as a “Switzerland for LLMs,” a neutral intermediary that lets teams keep productivity gains while keeping corporate data protected.
Founders Tomas Okmanas and Eimantas Sabaliauskas — best known for building Nord Security and other Tesonet ventures — say the current AI landscape risks creating what they call “the biggest corporate data leak” as employees paste sensitive documents into public models. Their approach is to insert a secure control plane that logs, filters and routes AI requests so companies can enforce policies without blocking everyday AI workflows.
Funding, credibility and a high-profile investor network
Just months after a stealth debut that included an $8 million round, Nexos.ai announced a €30 million Series A (roughly $35 million) at a €300 million valuation. Index Ventures and Evantic Capital co-led the round; previous backers Creandum and Dig Ventures also participated. Angel investors include CEOs from Datadog, Klarna, Supercell and Wix, lending both capital and strategic credibility.
Evantic, the new firm founded by former Sequoia partner Matt Miller, brings a unique asset: its "Legends" network of about 140 operators who advise portfolio startups. Okmanas says he’s both a Legend and a beneficiary of other operators’ expertise as Nexos.ai shapes its product roadmap.
Where the money will go
- Accelerate development of private model support for sensitive data
- Scale the AI Gateway and Workspace features
- Expand sales and engineering across Europe and North America
- Grow staff toward a 100-person team within the first year
Product: an AI Workspace for employees and a Gateway for control
Nexos.ai ships two core components today: an AI Workspace that employees use to interact with models, and an AI Gateway that developers and security teams integrate with. The Gateway acts as a control layer for security, cost management, and compliance while also reducing the fragmentation that companies face when dealing with multiple AI vendors.
By routing requests through a single access point, Nexos.ai currently supports roughly 200 AI models and plans to accelerate private-model integrations to keep sensitive data within customer-controlled environments. That architecture aims to give legal, security and IT teams a single place to apply governance while preserving the flexible, model-agnostic workflows that product and support teams want.
Early traction and target customers
Okmanas says the company is doing 50–60 demo calls per week. Nexos.ai is prioritizing two customer groups: tech-forward companies already using AI daily and regulated enterprises that need strict governance and data residency controls. Early disclosed customers include companies inside the founders’ Tesonet portfolio and Bulgarian fintech unicorn Payhawk.
Examples matter. At Hostinger — another portfolio company — an internal AI assistant reduced human support needs and saved roughly €10 million this year, according to Okmanas. That kind of tangible ROI is the playbook Nexos.ai is trying to roll out to larger organizations that still need to convince boards and auditors that AI can be safe and profitable.
What this means for enterprise AI adoption
Imagine an organization where every AI interaction is both productive and auditable. That’s the pitch: allow employees to use modern models without exposing corporate secrets or violating compliance rules. It’s a pragmatic compromise between the productivity gains of AI and the risk-averse instincts of enterprise governance.
Nexos.ai’s founders are betting that many companies won’t accept blunt bans, and that a neutral, model-agnostic control plane can unlock broader adoption. With fresh capital, a strong investor roster and early customer wins, the startup aims to be the intermediary many CIOs and CISOs can trust when they bring AI into core workflows.
Source: techcrunch
Comments
Marius
I've had folks paste invoices into public models, saw a small leak first hand. A neutral control plane could help, but hope it's not just locking teams down or adding big latency. Also who owns the logs? auditors will ask.
mechbyte
Sounds useful, but is it actually watertight? Model inversion, prompt leaks, vendor trust, etc.. curious if audits catch everything..
Leave a Comment