Deliverance AI has emerged from stealth with £6 million in annual recurring revenue. The London-based company says it has signed six enterprise customers and grown to more than 30 employees within three months of incorporation.
The UK-founded business is positioning itself as an operating system for enterprise AI for customers that want to run AI systems in private, sovereign or air-gapped environments. It is targeting government bodies, regulated industries and large companies that need tighter control over data, models and decision-making than public cloud services may allow.
The launch comes as companies reassess how to move AI projects beyond trials and into day-to-day operations. Many large organisations have spent heavily on computing infrastructure and pilot schemes, but still face questions over governance, auditability, cost control and where sensitive data is processed.
Its software is designed to address those issues by providing a runtime for AI agents, model routing, knowledge layers, audit trails and cost attribution. It also offers engineering support to help customers deploy the system in production.
The platform is already being used for professional services workflows, sales and operations tasks, and finance and business process work, according to the company. In one customer deployment, it said the software reduced costs by nearly 75% while also cutting the time needed to start and complete tasks.
Partner strategy
Alongside the launch, Deliverance AI outlined partnerships with HPE and NVIDIA. HPE provides the private cloud base for customer-controlled AI deployments, while NVIDIA supplies computing systems and software, including DGX Spark and NemoClaw.
The partnerships are intended to help regulated organisations deploy AI in their own environments rather than rely on infrastructure controlled elsewhere. That pitch is likely to resonate with customers concerned about data residency, infrastructure jurisdiction and the risk of extra-territorial access to information.
The company argues that the issue has become more pressing as businesses look to use AI on sensitive internal data. For sectors such as government, telecoms, financial services and other heavily regulated industries, the question is no longer just whether an AI model performs well, but whether its use can be governed and audited under local rules.
Mick McNeil, Chief Executive Officer and Founder of Deliverance AI, said the missing element in many corporate AI projects is not hardware but the operating framework around it.
"Enterprise AI will not scale on trust-me promises. The organisations with the most valuable data need AI that can operate inside their own environment, under their own controls, with clear governance over which models are used, where data goes and how decisions are made. Companies have spent heavily on AI infrastructure, but infrastructure alone does not give you an AI outcome. The missing layer is an operating system for agentic AI: somewhere to run agents, govern them, give them context, measure them and make them accountable. That is what Deliverance AI has been built to provide," McNeil said.
Sovereign focus
The company's emphasis on sovereign deployment reflects a broader shift in the European AI market. As generative AI tools spread across businesses, concerns about relying on US-based cloud platforms have become a growing factor in procurement discussions, particularly among public bodies and operators of critical infrastructure.
Deliverance AI said its software can run across hyperscale cloud, private cloud, on-premises systems and air-gapped environments. It is headquartered in London and serves customers across the UK, Europe, the Middle East and North America.
McNeil previously held senior leadership roles in cloud, computing and AI businesses at Microsoft, Northern Data Group and Logicalis. His new company is entering a crowded market for enterprise AI software, where established cloud providers, specialist infrastructure groups and start-ups are all trying to define how companies should manage AI agents in production.
One area where Deliverance AI is trying to stand apart is model routing. The company said its system directs tasks to different AI models based on performance, cost, risk and governance requirements. It argues that this approach can help customers avoid dependence on a single model provider, cloud platform or infrastructure vendor while giving them more room to adapt as the market changes.
That may appeal to buyers wary of lock-in after a first wave of AI spending. Large enterprises have often built early AI projects around a single vendor, only to find that switching later is expensive or operationally difficult.
McNeil said the company's work with HPE and NVIDIA is central to that effort.
"HPE and NVIDIA are central to how regulated enterprises will deploy AI at scale. HPE gives customers the private cloud architecture they need. NVIDIA provides the accelerated computing platform and emerging software stack for secure agentic AI, including DGX Spark and NemoClaw. Deliverance AI is the operating system, helping customers turn that foundation into governed, measurable and accountable AI workflows," McNeil said.
Anthony Hills, NVIDIA Regional Director, UK&I, Enterprise & Public Sector, added: "As AI becomes a foundational productivity layer for every enterprise, organisations are increasingly focused on turning accelerated computing infrastructure into controlled, measurable business outcomes. The integration of Deliverance AI's Agentic Operating System with the NVIDIA-accelerated computing platform and software stack provides a managed infrastructure designed for running governed agentic workflows within private and regulated environments."
James Brooks, UKIMEA Hybrid Solutions Leader, HPE, said: "Deliverance AI's emergence shows the growing demand for sovereign AI among enterprises. HPE Private Cloud AI meets this need by enabling agentic workflows to run within private, sovereign and air-gapped environments, keeping customers firmly in control of their data, models and decisioning."
As enterprises move from AI experimentation to large-scale deployment, demand for platforms that combine governance, flexibility and infrastructure independence is likely to grow. Deliverance AI is betting that sovereign, accountable and vendor-agnostic AI operations will become a core requirement for organisations seeking to deploy agentic AI in production.