Data Fabric Platforms

A data platform lead writing their November 2025 shortlist memo would have written "fabric Wave still planned." The memo would have been wrong twice over. The Forrester Wave: Data Fabric Platforms, Q4 2025, went live on 30 October 2025, authored by VP, principal analyst Noel Yuhanna, and the name on the cover is the category's third in seven years: Big Data Fabric in 2018, Enterprise Data Fabric through 2024, Data Fabric Platforms now.

Three names in seven years, and each rename was a product decision. That is the history of this market in one sentence, and the Q4 2025 scorecard is where the third decision lands.

The third name: The Forrester Wave: Data Fabric Platforms, Q4 2025

The evaluation scored fourteen providers against twenty six criteria across current offering, strategy, and customer feedback. The publicly confirmed placements run two Leaders, and they bracket the market's deployment spectrum.

Microsoft received the highest scores in both the current offering and strategy categories, with the highest possible scores in nineteen of the twenty six criteria, including end-to-end data fabric, agentic AI, real-time integration, and unified data governance and security. The hyperscaler's entry into this Wave at the top of both axes is the category's gravity shift made explicit: Fabric sits inside Azure, the data sits inside OneLake, and the integration argument that used to be the fabric vendors' entire pitch is now bundled with the cloud bill.

Cloudera became a Leader for the first time, with the highest possible scores in seven criteria: end-to-end integrated fabric, unified data catalog, real-time performance and scalability, vision, roadmap, metadata management, and agentic AI. Forrester's fit verdict names the other pole: "an ideal choice for organizations that want robust data processing, scalable storage, and persistent data management." The report credits Cloudera's private cloud and on-premises deployments with a stronghold in industries carrying data sovereignty or legacy system requirements.

Microsoft leads where the data can live in the cloud. Cloudera leads where it cannot. The two confirmed Leaders are the two answers to the same question, and the remaining twelve placements sit behind Forrester's paywall.

Three names, three products

The lineage tells you what each rename was for.

Big Data Fabric, Q2 2018, was the era of volume: fifteen vendors including Denodo, IBM, Oracle, SAP, Talend, TIBCO, and Cloudera, scored on how well they moved and processed the data estates that Hadoop had just made enormous. The word "Big" did the work.

Enterprise Data Fabric, the Q2 2022 and Q1 2024 editions, was the era of governance. Yuhanna's Q2 2022 framing named the stakes: delayed insights can have a devastating effect on a firm's ability to win, serve, and retain customers. The Q1 2024 edition, fifteen vendors against twenty six criteria, named Denodo a Leader for a "world-class data fabric solution" built on data virtualization, and it recorded the shift that matters most in this lineage: the market was moving from read-only analytical fabric toward a transactional fabric, with bidirectional read-write capabilities for modern applications like microservices and real-time customer experience.

Data Fabric Platforms, Q4 2025, is the era of operation. The "Big" is gone because every estate is big. The "Enterprise" is gone because the buyer is no longer only the enterprise architect. What remains is the platform claim: not a connector layer, not a governance layer, but the layer applications read and write through.

The fabric became the nervous system

The transactional shift from the Q1 2024 edition and the agentic AI criterion in Q4 2025 are the same sentence, written eighteen months apart.

A fabric that only feeds dashboards is a pipe. A pipe moves data. A fabric that supports bidirectional read-write for microservices and real-time experience is something else: it is the layer decisions run through. Add the new criterion and the sentence completes itself. AI agents read data, write decisions, and act. If the agents are to be trusted, the fabric is the nervous system they read and write through, and the governance of that traffic is the new product.

Look at where the 2025 maximum scores concentrate in the two confirmed Leaders: agentic AI, unified data governance and security, metadata management, real-time integration. None of those criteria existed in the 2018 edition's world. The scorecard is no longer asking how well the vendor moves data. It is asking how well the vendor governs what the machines do with it.

What it means when the analyst wrote the definition

Here is the honest limitation, and it is specific to this category.

Forrester defined the data fabric category. Yuhanna has been its primary author across every edition, from Big Data Fabric through today. That is a strength and a caution at once. The strength is continuity: no other category in this series can claim a single author across seven years and three names. The caution is that the Wave scores conformance to a definition its own author wrote. Buyers are not measuring vendors against an independent market reality. They are measuring vendors against Forrester's conception of the market, updated every two years, and the conception has changed names three times because it keeps changing substance.

The practical consequence is that the scores travel less well than they appear to. A Leader score in the 2024 edition meant the vendor matched the definition of a governed, read-mostly fabric. A Leader score in 2025 means the vendor matched the definition of a transactional, agent-ready platform. Same word, different product, and the vendors that led one edition are not guaranteed to lead the next. The definition is the product, and the product keeps changing.

There is also the question the report leaves open by design: where the fabric ends and the data cloud begins. The transactional shift walks the fabric into the path of the platforms that hold the data. The Q4 2025 Wave does not settle that fight. It scores the fabric side of it, and the buyer's first architectural decision remains which side their data lives on.

Three questions for the fourth name

Which pole is yours? Microsoft's win is the cloud-native answer, Cloudera's is the sovereignty-bound one. If your data cannot leave your private cloud or your jurisdiction, half of this Wave's leadership picture is not available to you, and the shortlist should be built from your pole inward, not from the scores outward.

Are you buying a pipe or a nervous system? If your use case is analytics and reporting, most of the 2025 criteria are buying capacity you will not use. If your use case is operational read-write with microservices and real-time experience, the transactional shift is the criterion that matters, and it is the one the 2024 edition was already scoring on promises.

Who governs the agents, and can you see the audit trail? Agentic AI is a scored criterion now, and for European buyers the follow-up is not optional: which agents can write, what can they write to, and does the platform log every machine decision about data at a level a data protection authority could read. The nervous system's traffic is the product. Its records are your liability.

The category has renamed itself every time its product changed, and it has never stopped changing. The Q4 2025 Wave is the scorecard of the operational turn, published under the third name, with the definition's author still at the wheel. Read it as the market's current self-description, not as a verdict, and expect the fourth name to be a different product again.

The "where is your data allowed to live" question this category answers architecturally gets a jurisdictional answer in Sovereign Cloud Services, where the first scored evaluation names the hyperscalers themselves as four of its five confirmed Leaders.

Forrester does not endorse any vendor named here, and tier placement should not be read as a recommendation to buy.

The neighboring coverage here is Data Integration Tools. AI agents don't wait for a nightly batch job, which is why Gartner's newest quadrant rewards vendors who serve data where it lives over vendors who still move it into a warehouse first.