AI use inside big companies keeps rising while oversight over how employees are making use of AI is lagging. Lithuanian-based startup nexos.ai, founded by the creators of Nord Security, pitches a single control layer that unites a developer AI Gateway with a secure AI Workspace for employees. The core idea is to standardize model access, apply policy at the platform edge, export full interaction logs into a company’s security and identity management systems, and keep model spend predictable that finance can track. The goal is to replace scattered AI tools with a unified interface that brings policy, visibility, and flexible access controls to leading models.
Coming from a Deep Bench in Enterprise Security
The solution comes in multiple parts. Developers route traffic through an AI Gateway that centralizes model access, routing, policy, and spend controls. Employees use the Workspace for governed chat and reusable agents. Both surfaces feed auditable logs back to the customer’s security tooling. In practice, a product manager could build an agent that calls two model families with PII redaction, while security reviews each interaction in one dashboard. The founders say model neutrality matters so teams can swap providers as performance or cost shifts without rebuilding the control plane.
Chief Commercial Officer Justas Morkunas stresses the importance of observability in the AI system. “Whatever happens within the system, we are logging everything, we are sending to your security, so companies can control the stuff that happens within the AI.”
The company was founded by the creators of Nord Security. Same co-founders, but new startup with its own funding, team, and ownership. That separation enables a fast route to market and accelerated capital plan that is optimized for the hyper pace of the AI markets.
Morkunas describes the origin of the company as a response to internal sprawl that is increasingly happening with ad-hoc and piecemeal use of AI in the enterprise. Teams are using a hodge podge of APIs with the result that costs spiked, and no one could describe who sent what where.
The products started with the Gateway. Market feedback pushed the company to add a Workspace within weeks so buyers could roll out governed, ready-to-use experiences rather than assemble toolkits. “We started building in January 2025, and just a few months later we had evolved to add a workspace,” shared Morkunas.
First customer wins arrived fast. “By June, we closed the first customer from FinTech.” Seed funding also came early to put the company “on the map,” with a $30 million Series A and rapid hiring with a heavy bias toward R&D. The company already has around 100 employees.
The company is also building out partnerships that serve the security narrative the team wants customers to see. “A week ago, we announced a partnership with CrowdStrike.” CrowdStrike, for its part, has been expanding an AI security ecosystem that reaches into purpose-built AI clouds and NVIDIA’s stack, a signal that buyers can expect deeper integrations across AI infrastructure and security controls.
Why the Timing is Right for an Enterprise Focus
The ideal customer profile is focused on industries that require controls over their AI usage. These organizations face a wave of tooling, inconsistent policies across model providers, and cost spikes from uncontrolled usage. Morkunas points to a board-level shift from curiosity to return. “We were thinking that for at least the upcoming two years, we’ll have a gap that nobody’s expecting ROI from AI. But it changed very quickly,” explains Morkunas, “Everybody switched very quickly to ‘green money’ that needs to show real return. We’re selling straight to the top in the organization.”
At that level of enterprise engagement, the sales approach runs through analyst stages and board materials as much as developer communities. Buyers want speed with control, and they want evidence in the form of logs, policies, and cost discipline, not pilots that drift.
The enterprise AI market is facing an increasingly crowded field, mapped to two buyer conversations. First are bottom-up developer- and user-led approaches that start with individual groups experimenting and implementing point solutions. This can lead to rapid iteration but also chaotic and uncontrollable systems.
On the other hand, driving from a top-down approach can accomplish whole-organization coordination, cost management, and more strategic alignment to business imperatives. At this level, companies like Salesforce, Microsoft, SAP, Oracle, and others are aiming to get executive-level attention.
Morkunas places nexos.ai across both layers and argues consolidation reduces governance debt. “Our aim is to tell the companies that we can reach the organization horizontally to meet 60-70% of the organization’s AI needs.”
Many vendors can ship a demo. Few can repeat wins across conservative buyers at scale. Morkunas keeps returning to go-to-market mechanics. “It’s all about distribution… You want to win the hearts of the enterprises.” He expects the channel to matter once direct playbooks harden. “If you just go and say, ‘Please push my product to the market,’ it doesn’t work.”
Headquartered in Vilnius, Lithuania, and backed by Evantic Capital, Index Ventures, Creandum, Dig Ventures and angels including Olivier Pomel of Datadog, Sebastian Siemiatkowski of Klarna through Flat Capital, Ilkka Paananen of Supercell, and Avishai Abrahami of Wix.com , nexos.ai is showing that the market is wide open for competition from focused teams who know the problems they need to solve.
