Percona report warns of database lock-in & AI woes

Percona has released its 2026 State of Open Source Database Management Report, which found that 54% of respondents said vendor lock-in makes it harder to adapt to changing regulatory requirements.

The report points to a broad reassessment of database infrastructure as organisations contend with rising cloud spending, licensing bills, operational pressures and the demands of artificial intelligence workloads.

Cost pressures were clear in the findings. Some 31% of those surveyed said rising cloud costs were the biggest barrier to reducing the total cost of ownership of their database estates, while 23% cited rising licensing costs as their main problem.

Performance issues also featured heavily. Some 42% of respondents identified inefficiency or slow throughput as a challenge, 41% said database problems were wasting engineering time, 40% cited downtime, and 37% pointed to difficulty scaling systems.

Compliance pressure

Regulation emerged as another area where database choices are shaping operational risk. More than half of respondents said vendor lock-in made it harder to respond to changing compliance demands, while more than half also cited fragmented tools or policies as a challenge.

The survey found that 46% of respondents identified both software supply-chain traceability and a reactive security posture as obstacles in their database environments. The findings come as companies face tighter scrutiny over how they manage software vulnerabilities and maintain oversight of their technology stacks.

A separate result highlighted the age of many production systems. Almost all respondents, 97%, said their organisation was running at least some database versions beyond official end-of-life, while 44% said many of their databases were in that state.

That level of outdated software suggests many businesses are balancing risk, cost and disruption when deciding whether to modernise. Legacy environments can be difficult to replace, particularly when applications, compliance processes and in-house expertise are closely tied to long-established platforms.

AI barriers

The report suggests AI readiness is constrained less by a lack of tools than by underlying organisational and architectural problems. Fragmented data architecture and skills shortages were tied as the top barriers to AI adoption, each cited by 23% of respondents.

That indicates many organisations still face basic integration and staffing problems before they can make broader use of AI systems. In practice, database teams are being asked to support new workloads while also managing reliability issues, cost controls and regulatory demands.

The research was based on a survey of 300 US-based database administrators, site reliability engineers and platform engineers. All respondents said they used at least one of the following database systems: PostgreSQL, MySQL, MongoDB, Redis or Valkey.

Peter Farkas, Chief Executive Officer at Percona, said the pressures on database teams are widening beyond traditional concerns such as speed and uptime.

"Database infrastructure is being asked to do more than ever before," Farkas said.

"Cloud economics are changing, AI is introducing new workload patterns, regulations are becoming more complex, and organizations are increasingly conscious of where their data and technology are controlled. The organizations that will be best prepared are not necessarily those that can predict what comes next. They are the ones that preserve the freedom to adapt when it arrives," he said.

The combined findings suggest cost, performance, regulation and AI demands are becoming more closely linked in infrastructure planning. As a result, database decisions are increasingly affecting not only day-to-day operations but also how easily companies can respond to shifts in policy, budgets and technical requirements.

Farkas said this is changing how leaders should think about database strategy.

"Database leaders need to build for change, not for a single version of the future," he said.

"That means choosing infrastructure they can adapt as regulations shift, workloads evolve, and costs rise, without having to rebuild their entire architecture. Control and flexibility need to be requirements, not afterthoughts."