Ajelix launches enterprise AI platform backed by its own GPU infrastructure

Ajelix on Sept. 8, 2026, said it has acquired its own GPU compute infrastructure and made its enterprise AI platform generally available. The move is aimed at giving organizations more control, governance and data sovereignty as they deploy AI at scale.

Why it matters: - Enterprise AI adoption often breaks down when tools are hard to govern, hard to audit or dependent on infrastructure a company does not control. - Ajelix is positioning its own compute stack and enterprise platform as a way to make AI deployments more reliable for regulated and operations-heavy teams. - The platform is designed to support legal, finance, compliance and other teams that need source-backed answers and tighter controls.

What happened: - Ajelix announced the acquisition of its own GPU compute infrastructure and the general availability of the Ajelix Enterprise Platform on Sept. 8, 2026. - The company said the platform is a complete operational layer for organizations deploying AI at scale. - Ajelix said the platform is now available for demo bookings at Ajelix enterprise.

The details: - Ajelix said it now owns and operates its own servers and hosts frontier open-source models such as GLM-5.3. - The company said data sovereignty is built into the infrastructure rather than handled only through contracts. - The enterprise platform includes custom AI agents that can read company documents, use company tools and escalate high-stakes decisions to humans. - Each answer is backed by a source citation, according to Ajelix. - A no-code workflow builder lets users connect agents, tools and people, then run workflows on a schedule, by event or through a webhook. - A knowledge layer turns documents, policies and datasets into material the AI reasons over and links answers to exact source passages. - Built-in governance includes role-based access, audit trails, guardrails, policy enforcement on every message and spend controls by team and project. - Companies can run any LLM provider through one governance layer. - Ajelix said the platform is available in three configurations: Ajelix Cloud, Self-hosted and Cloud Providers on AWS, Azure or Google Cloud. - Ajelix said the platform was developed alongside enterprise clients in financial services, legal, insurance, public sector, healthcare, logistics, manufacturing, energy, marketing and real estate. - The company said it has spent four years building toward the launch. - Ajelix said it began as an Excel formula tool, grew to 350,000 users without external funding and launched agentic AI chat in February 2026. - Ajelix said it was founded in 2022 and now has more than 350,000 users across 150+ countries. - Ajelix said it is a member of the NVIDIA Inception Program, Latvian Investment and Development Agency Incubator Program, OVHcloud Startup Program, AWS Activate and Google for Startups.

Between the lines: - Ajelix is making a broader argument that enterprise AI needs infrastructure ownership, not just software features, to meet security and governance requirements. - The launch also shows the company moving from individual productivity use cases toward infrastructure for organizational deployment. - Co-founder Agnese Jaunosane said the company kept hearing a common enterprise question about where data goes, and that owning the compute was the way to address it. - Jaunosane said companies that adopt AI successfully treat it like a critical system with governance, accountability and visibility into usage and cost.

What's next: - Organizations can book demos now and evaluate the platform in cloud, self-hosted or major public-cloud deployments. - Ajelix will likely use the new infrastructure and enterprise stack to pursue more regulated and large-scale customers. - The company is betting that control over compute, auditability and governance will be a key differentiator as AI moves deeper into enterprise workflows.

The bottom line: - Ajelix is trying to sell not just AI software, but the operational layer enterprises need to trust AI in production.