Oracle bought Cerner, and Cerner customers inherited a cloud strategy

The week

On December 20, 2021, Oracle announced an agreement to acquire Cerner, one of the two largest electronic health record vendors in the United States, for approximately $28.3 billion in cash. Oracle said it would accelerate the modernization of Cerner’s systems and run them on its own cloud infrastructure. Cerner’s customers included many US hospitals and health systems. They also included the Department of Veterans Affairs, which had selected Cerner to replace its legacy record system and was rolling it out site by site, and the Department of Defense. The acquisition closed in June 2022, and the business later operated as Oracle Health.

Acquisitions of enterprise software vendors are common, and customers often experience them as a new logo on the invoice. This one differed in scale and in kind. Oracle was buying a health record vendor and, with it, a large and durable customer base for its database and cloud infrastructure. It said so publicly. For a health system on Cerner, the relationship with its most important software supplier would increasingly be about infrastructure as well as clinical software. Three areas needed fresh thinking, each on its own clock: the product roadmap, the hosting model, and the commercial terms.

The product roadmap

An electronic health record is more than software. It embeds order sets, documentation templates, clinical decision support, revenue cycle rules and hundreds of interfaces built up over years. A health system’s investment in configuration, training and integration usually far exceeds what it paid in licenses. That makes the vendor’s roadmap a strategic dependency, and an ownership change puts that roadmap in question.

After the announcement, customers needed to know whether Oracle would evolve the existing platform, rebuild it on new technology, or both, and on what timeline. A rebuild might bring a modern platform but also a migration, with retraining and revalidation of workflows. Evolution might be smoother but slower. Either path would shift where development effort went, possibly away from features a given customer was waiting for.

Early on, much of this was unknown, and that uncertainty was information in itself. Health systems approaching a major upgrade, a new module or a large integration project had reason to sequence carefully. A practical step was to list the roadmap items the health system depended on over the next three years and ask the vendor to confirm each in writing. Typical items were regulatory updates, interoperability features and revenue cycle functions. Where confirmation was not forthcoming, the health system knew where its plan carried risk.

The hosting model

Many Cerner customers already used vendor-hosted arrangements. Oracle’s stated plan to run the platform on its own cloud meant hosting choices would likely narrow over time. For a health system, that had consequences beyond cost. Security responsibilities, disaster recovery design, data residency and integration with the health system’s other environments would all follow the vendor’s infrastructure decisions.

Health systems with a cloud strategy built around another provider faced a concrete issue. Their record, often their largest and most critical application, might end up on a different cloud from their analytics, identity and productivity platforms. That is manageable, but it affects network design, latency for interfaces, the cost of moving data out for analytics, and the tooling that operations teams use to monitor and recover systems. Planned, it is an architecture decision. Discovered at migration time, it is an unbudgeted project.

The hosting shift also changed the shared-responsibility line for security. On a self-hosted or third-party-hosted record, the health system controls much of the stack. On the vendor’s cloud, it relies on the vendor’s controls, logs and incident response. Contracts and business associate agreements needed to reflect that. They should say what logs the health system can access and how quickly it is told of incidents affecting its environment. They should also say how recovery time and recovery point objectives are tested and reported.

The commercial terms

Acquisitions change the negotiating dynamic. The acquirer has paid a premium and has targets to meet. Customers can expect changes to pricing, bundling and support structures over time. Yet the period right after an acquisition is often when the acquirer most wants to retain customers. That creates a window.

Health systems with renewals approaching could seek protections that would hold through a transition. These included price caps, support for current versions for a defined period, data access and export rights, and clear terms for any hosting migration, including who pays. Those with long remaining terms needed to read change-of-control and assignment clauses to see which protections carried over. Organizations should review these terms with counsel and procurement specialists who know health IT contracts.

Data rights deserved the most attention. A record holds decades of clinical history, and the ability to extract it in complete, documented, usable form is a health system’s ultimate protection against any vendor’s decisions. Contract language on export format, completeness, cost and timeline is worth more than many price concessions. The federal information blocking rules, in effect since 2021, strengthened expectations about access to electronic health information, but they do not replace clear contractual rights.

Some health systems used the moment to evaluate alternatives. Replacing a record is among the most expensive and disruptive projects a health system can undertake, measured in years and significant capital, and an ownership change alone rarely justifies it. A structured review still had value. It asked how well the platform serves clinicians and patients, what it will cost over ten years under plausible scenarios, and what a move would take. The output strengthened the health system’s hand in negotiation whether or not it led to a switch.

With hindsight

Oracle completed the acquisition in June 2022 and in 2024 announced a new, cloud-based health record platform with embedded AI features, while continuing to support the existing system. The VA’s rollout ran into serious problems, including patient safety concerns documented by the VA inspector general, and the department paused most new deployments in April 2023. It restarted in April 2026, when four Michigan medical centers went live, with more sites scheduled through the year. In the commercial market, health record consolidation continued, and a number of health systems moved from Cerner to other platforms, often in connection with mergers. By this point in 2026, Oracle Health remains a major vendor with a large installed base. Its customers’ experience has depended heavily on how carefully they negotiated hosting, support and data terms during the transition.