Fortune 500 Companies Adopt Open-Source AI for Data Control

Major enterprises are rapidly embracing open-source artificial intelligence models, driven by the need to maintain control over proprietary data while meeting strict regulatory requirements, according to Ollama CEO Jeffrey Morgan.

Speaking with Yahoo Finance, Morgan disclosed that his company's platform is adding nearly one million new installations each week. More significantly, token usage—a measure of actual AI model activity—has doubled every month for the past six months, signaling what he described as "insatiable" demand for open models.

Enterprise adoption across regulated industries

Fortune 500 companies are already deploying open-source AI at scale, Morgan revealed, though this adoption has remained largely under the radar. The client base spans retail, healthcare, and industrial sectors—including some of the most heavily regulated industries where data governance is paramount.

The appeal centers on proximity and control. Enterprise clients want AI capabilities that work directly with their data without transmitting sensitive information to third-party model providers or servers in foreign jurisdictions. For companies facing stringent compliance requirements, this localized deployment model eliminates risks that come with conventional cloud-based AI services.

Why it matters

The shift toward open-source AI represents a fundamental challenge to the dominance of proprietary models from major tech vendors. When Fortune 500 companies can deploy powerful AI locally while maintaining full data sovereignty, the traditional cloud AI business model faces pressure. This trend also suggests that regulatory concerns and data security—not just model performance—are driving enterprise AI architecture decisions.

Business model and expansion plans

Ollama operates on a freemium subscription model. Individual developers can download and use the platform at no cost, then upgrade to paid subscriptions when they need access to larger cloud-hosted open models running on US and European servers.

The company is introducing a team plan designed specifically for business and enterprise adoption. This offering will use subscription pricing with predictable token-based scaling, allowing organizations to expand their AI compute resources as needed across different business functions.

Competition and model performance

When asked about competitive threats from open-source models developed in China, Morgan emphasized that enterprise customers care more about where and how a model runs than its country of origin. Security and operational control take precedence over provenance.

Morgan highlighted strong customer interest in Nvidia's new Nemotron 3 Ultra models, which he noted deliver top performance for complex, long-horizon agent tasks—the kind of sustained AI operations enterprises need for substantive workflow automation.

These details were first reported by Yahoo Finance in an interview with Morgan.