Between December 2024 and May 2025, four customer data platforms with real customers and real revenue stopped being independent companies. ActionIQ went to a conversational-AI firm. mParticle went to an ad-tech business. Lytics went to a content management vendor. Census went to a data-pipeline company. Not one of those acquirers was another CDP, and not one of them wanted the same thing.
That is the part worth sitting with. A run of acquisitions in a category is easy to wave away as ordinary market churn. This run is not ordinary. It is a verdict on a business model: the standalone, pure-play CDP, sold on its own to enterprises as a product they log into, is getting harder to sustain. Reading the deals carefully tells you why, and it tells you what to check before you sign or renew a contract.
The pattern is older than the headlines
CDPs being absorbed is not new. It is the default ending for the category, and the early examples set the template.
Salesforce bought the data management platform Krux in 2016. TechCrunch reported the deal at roughly $340 million in cash plus stock, with a total value that could reach about $700 million. Krux became Audience Studio inside the Salesforce suite. Arm acquired Treasure Data in 2018, a deal widely reported at around $600 million, before it was spun back out under SoftBank ownership as the CDP market accelerated. Acquia bought AgilOne in 2019 and folded it into a digital experience platform.
Then came the largest of them all. In October 2020, Twilio agreed to buy Segment, the vendor that effectively defined the modern CDP, for about $3.2 billion in an all-stock deal. At the time it read as validation: the category had produced a multi-billion-dollar outcome. What happened next reads differently. By early 2024 Twilio was under activist-investor pressure, and as CB Insights documented, there was open speculation that Segment would be sold. Twilio ran an operational review and announced in March 2024 that it would keep Segment rather than sell, with the stated goal of pushing it to break-even non-GAAP operating income by the second quarter of 2025. A flagship CDP, inside a public company, managed toward break-even is its own kind of signal.
So consolidation did not begin in December 2024. What changed is the speed, and the type of buyer.
Four deals in six months
Take them one at a time, because the differences are the point.
ActionIQ to Uniphore. Announced in early December 2024. ActionIQ was an enterprise composable CDP that had raised on the order of $144 million across its funding history. Uniphore is a conversational-AI company built around contact-center automation, voice AI and agent assist. It bought ActionIQ, alongside the data-engineering firm Infoworks, to feed what it calls a "Zero Data AI Cloud". Uniphore's own framing was blunt about the reason: AI initiatives stall in proof of concept because the underlying data is not prepared in a way models and agents can consume. The CDP, in other words, was bought as fuel for agents, not as a marketing product in its own right.
Lytics to Contentstack. Announced in January 2025, with the transaction reported as closing in December 2024. Contentstack is a composable content management and digital experience platform. Lytics gave it real-time customer profiles and audience data for both known and unknown visitors. The logic here is a content company wanting to personalize what it serves. The CDP became the targeting engine behind a content stack.
mParticle to Rokt. Announced in mid-January 2025 at $300 million. This is the deal that should give a CDP buyer the longest pause, and the numbers are why. mParticle had raised about $272 million in total funding and, per CDP.com's profile, generated roughly $76 million in revenue in 2024. A $300 million price against $272 million of capital raised is barely above the money put in. For a venture-backed company that once expected a far larger exit, that is a sober outcome. Rokt is an ecommerce technology business that monetizes the transaction moment, the few seconds during checkout when a relevant offer can be placed. It wanted mParticle's live feed of customer data to make those moments smarter. AdExchanger reported that more than 90 percent of the two companies' clients overlapped as enterprise retailers and brands. The CDP here was bought by ad-tech to sharpen commerce targeting.
Census to Fivetran. Announced on May 1, 2025. Census was a reverse-ETL and composable CDP vendor; TechCrunch noted it had raised more than $80 million and was last valued at around $630 million in 2022. Fivetran moves data into warehouses. Census moves data out of warehouses and into business applications. Buying Census let Fivetran offer the round trip, ingestion and activation under one roof, and the product was folded in as Fivetran Activations. The CDP here was absorbed into a data-platform company as the missing half of a pipeline.
Terms were not disclosed for every deal. ActionIQ's price has been reported at roughly $145 million but was not officially confirmed; Lytics and Census terms were not disclosed. What is solid is the shape of the wave, and the shape is the lesson.
Four buyers, four different appetites
Line the acquirers up and a structure appears. These were not four versions of the same purchase. They were four kinds of company, each treating the CDP as a component of something larger.
A conversational-AI firm wanted the data layer that makes its agents work. A content and experience platform wanted a personalization engine. An ad-tech and commerce company wanted real-time targeting signal. A data-pipeline vendor wanted to complete a pipeline. CX and service suites, content suites, ad-tech, and data infrastructure: that is the full set of acquirer types, and the common thread is that none of them sells a CDP as the headline product. Each one wanted the CDP as plumbing inside a different machine.
This matters because it explains why the pure-play model is hard. A standalone CDP has to win a budget line on its own, against buyers who increasingly get customer-data features bundled into tools they already own. Its costs are real: identity resolution, connectors to dozens of destinations, real-time infrastructure, and a sales motion long enough to land six-figure enterprise contracts. Meanwhile the 2026 Gartner Magic Quadrant for Customer Data Platforms dropped four vendors outright, ActionIQ, Redpoint Global, mParticle and Zeta Global, for failing updated inclusion criteria around CDP software customers and contract value. The bar for counting as a serious independent CDP went up at the same moment several independents stopped being independent. The mParticle math says the financial reward for staying the course was thin. When the standalone path is expensive to run and modest to exit, the rational move for many vendors is to become a feature of a bigger platform.
Be precise about what this does not mean. It does not mean the CDP job is going away. Companies still need unified, person-level customer data. The function is healthy. It is the standalone company selling that function as its own product that is under pressure.
What consolidation costs the buyer
If your CDP gets acquired, or you are buying one that might be, the risk is not abstract. It shows up in three concrete ways.
Roadmap risk comes first. When a CDP becomes a component of a parent's strategy, its roadmap serves that strategy. A reasonable worry with the Rokt deal, raised in CDP.com's analysis, is whether general-purpose CDP features will keep getting the same investment as ecommerce-optimization features, given that Rokt's core business is the checkout moment. The same question applies across the wave. Under Uniphore, does the CDP develop as a broad marketing platform or as agent infrastructure for contact centers. Under Contentstack, does it serve every channel or mainly the ones that personalize content. Acquisition does not automatically degrade a product, but it does redirect it.
Forced migration is the second cost. Acquirers rebrand and re-platform. ActionIQ has been folded into Uniphore's platform. Census became Fivetran Activations. Salesforce's Krux became Audience Studio, and Salesforce later retired its entire DMP suite. Rebranding is cosmetic; re-platforming is not. If the acquirer eventually merges your CDP into a different architecture, or sunsets a tier, you inherit a migration you did not plan and did not budget. Twilio's planned sunset of its Engage Premier tier, noted in the Gartner coverage, is the kind of event that turns into a project for the customers on it.
Support and attention is the third. Integration consumes engineering and support capacity. Teams get reorganized, founders move into new roles, and the people who knew your account move on. Service can dip during the transition even when the acquirer means well. A slipped support ticket is the smallest of the three, though: weigh each risk by how expensive it is to absorb.
The honest counterpoint: not every acquisition is bad news for customers, and the parent's health is the variable that decides. A CDP bought by a struggling parent is at real risk of being starved. A CDP bought by a healthy, growing one may get more investment than it could fund alone. Rokt reported around $600 million in 2024 revenue while growing more than 40 percent year over year, and said it would increase investment in mParticle. That is a materially better backdrop than a distressed buyer. So the question is not only "who bought it" but "can the buyer afford to keep it good, and does its strategy still include my use case."
Which independents still look durable
If standalone CDPs are under pressure, are any independent positions still defensible? A few patterns hold up better than others.
Scale and a suite around the data layer is one. Salesforce, Adobe and Oracle do not sell a CDP as a lonely product; the CDP is the data foundation under a broad application suite, and Salesforce reinforced exactly that logic by agreeing in May 2025 to acquire the data-management company Informatica for about $8 billion to strengthen its data foundation for agentic AI. These vendors are consolidators, not targets.
Warehouse-native independence is a second. Vendors whose model keeps customer data in the company's own cloud warehouse, and activates it from there, are selling architecture as much as software. Hightouch, built on reverse ETL and now positioned as an agentic marketing platform, was placed as a Leader in the 2026 Gartner quadrant, an unusual result for a composable vendor. Because the data stays in infrastructure the customer already owns, the switching dynamics differ from a packaged CDP that holds a copy of everything.
A clear, hard-to-copy specialty is a third. Tealium, with deep roots in tag management and real-time event routing, and Amperity, known for AI-driven identity resolution, have held independent positions by being genuinely strong at a specific job. Independence built on a real technical moat is more durable than independence built on being a general-purpose CDP. Even so, the Gartner write-up flags viability questions for some niche players, so durable is a relative word, not a guarantee.
The pattern that does not hold is the mid-size, general-purpose, packaged standalone CDP with no suite, no warehouse-native architecture, and no defensible specialty. That is the profile most exposed to being acquired or out-competed.
What to take from the wave
The acquisitions of ActionIQ, mParticle, Lytics and Census, set against the longer history of Krux, Treasure Data, AgilOne and Segment, point one direction. The customer-data function is essential and not in doubt. The standalone company selling it as a product is the part being absorbed, into CX and service platforms, content platforms, ad-tech and data infrastructure, because each buyer wants the data layer as a component and running a pure-play CDP is expensive while the exits are unremarkable.
For a buyer, that turns the purchase decision into a diligence exercise. Ask who owns the vendor and how financially healthy that owner is. Ask whether the owner's strategy still has room for your specific use case, or whether the roadmap is bending toward the parent's core business. Get a concrete roadmap covering the next 18 to 24 months for the capabilities you depend on. Read your contract for what happens on acquisition, sunset or migration. And weigh architecture: a CDP that holds a copy of all your data is harder to leave than one that activates data from a warehouse you already control. Consolidation is not a reason to avoid CDPs. It is a reason to choose one with your eyes open about who might own it next.
Council summary
This post argues that the 2024 to 2025 run of CDP acquisitions is a verdict on a business model, not routine churn: each acquirer, a contact-center AI firm, a content platform, an ad-tech business and a data-pipeline vendor, wanted the data layer as a component rather than a headline product. Every named deal was checked against primary sources, including Rokt's $300 million for mParticle against $272 million raised and the 2026 Gartner Magic Quadrant that dropped four vendors; the financial claims, dates and acquirers stand as written, with wordy passages tightened. The takeaway is concrete: the customer-data function is healthy, but buying a CDP is now a diligence exercise about who owns the vendor, whether the owner's strategy still covers your use case, and how hard the architecture is to leave.
Comments