OpenSearch survey flags AI infrastructure vendor worries

The OpenSearch Software Foundation and Linux Foundation Research have published a report on AI-related data infrastructure trends. It found that 83% of organisations already run AI workloads or plan to do so.

The findings highlight growing concern among technology leaders about vendor dependence, platform costs, and the ability to manage data across different systems as AI use expands. The report found that 71% of organisations see deploying infrastructure independently of a specific vendor or cloud provider as a strategic priority.

Cost pressure also emerged as a major factor in infrastructure choices. The research found that 80% of respondents would reduce the number of platforms they use if doing so cut costs, while 80% cited total cost of ownership as a main requirement for a data infrastructure platform.

Security and compliance ranked almost as highly, with 79% identifying them as key needs. Interoperability was another leading issue, with 78% saying cross-vendor and cross-tool compatibility is critical when assessing infrastructure options.

AI uptake

The report presents AI deployment as widespread across businesses, with adoption strongest among larger enterprises. The share of organisations running or planning AI workloads rises above 90% for large companies.

Generative AI and large language model applications were the main use case for data infrastructure, identified by 82% of respondents. Hybrid search, which combines keyword and semantic search, was the preferred retrieval method for 68%.

These findings suggest that the systems used to search, index, retrieve, and monitor information are becoming central to how companies build and manage AI services. The report argues that this has increased demand for data platforms that can support governance, operational oversight, and spending control.

OpenSearch featured prominently in the study, which tracked awareness and use of the open source project. According to the findings, 89% of respondents were aware of OpenSearch in 2026, compared with 68% two years earlier.

The research also found that 77% of organisations view OpenSearch as either a core or supporting part of their AI infrastructure, while 62% already use it in AI workloads. Search and retrieval remained the most common use case, cited by 91%, but the study also pointed to broader use in log analytics, observability, and real-time analytics.

Vendor concerns

The results come as companies weigh how much control they want over the data systems that underpin AI applications. Choosing a platform for these workloads often involves long investment cycles, migration costs, and operational risks, making vendor independence a more significant consideration for buyers.

Bianca Lewis, Executive director of the OpenSearch Software Foundation, said the findings reflect what users have been discussing with the organisation. "This research confirms what we're hearing directly from the community," Lewis said.

She added: "As organizations expand AI efforts, they are worried about the risk of proprietary lock-in, and they need a neutral, open platform they can control. That's why enterprises are shifting to an open model, enabling cost predictability, data sovereignty, and long-term stability."

The report presents open source and vendor-neutral approaches as increasingly attractive to organisations that want to avoid being tied to a single cloud provider or software supplier. That preference appears to overlap with broader efforts to rationalise technology estates as AI spending rises.

User examples

The study also included comments from organisations using OpenSearch in production settings. Those examples focused on regulated and sensitive environments, where control over data handling and customisation were cited as important.

"Our use of data infrastructure technology dates back more than ten years to early versions of our primary data platform. When licensing changes disrupted that ecosystem, we transitioned first to an intermediate distribution and eventually moved directly to OpenSearch. Today, we use the technology heavily across multiple major areas, including global anti-money laundering, public security, and anti-fraud solutions," said Terry Quigley, Principal software developer at SAS.

Art Abal, Managing director of Vana Foundation, described privacy requirements as a deciding factor in infrastructure design. "Handling highly sensitive data under a consent framework requires extreme customizability to maintain strict privacy promises. While commercial solutions often violate those privacy bounds, OpenSearch provides a customizable framework that allows organizations to maintain aggregate data layers while upholding user trust," Abal said.

The report depicts a market in which AI adoption is no longer a niche activity but a mainstream operational concern, with infrastructure decisions increasingly shaped by governance, cost, and control. Among respondents, 44% said they preferred one core platform supported by specialised tools.