Every large company has a version of the same problem. Customer data lives in ten different places: the CRM, the e-commerce platform, the loyalty system, the call center software, the email tool. Marketing wants to send a personalized offer to customers who bought product X but not product Y, and then measure whether it worked. In practice, that request sits in a queue for weeks, gets pulled by a data analyst, exported into a spreadsheet, and loaded into a campaign tool that does not share data back with the original system.
Databricks, a company now running at over $5.4 billion in annual revenue and valued at $134 billion, has launched a product called CustomerLake to address exactly this. It is a customer data platform, or CDP, built directly inside Databricks' existing infrastructure. A CDP is software that pulls customer information from all your different systems, creates a single unified profile for each customer, and makes that profile usable for marketing. What is different here is that CustomerLake does this inside the same platform where the data already lives, rather than requiring companies to copy everything into a new, separate system.
The CDP market itself is a crowded, complicated space. The market sits at around $3.3 billion today and is expected to reach over $17 billion by 2034. But the growth masks real problems. Only 64% of deployed CDPs deliver significant value, a number that has been falling. The top complaints from buyers are not about missing features; they are about complex setup, rising costs, and poor support. Adobe's platform can cost $50,000 to $100,000 per year just to start. Salesforce has renamed its equivalent product six times since 2020. These are not signs of a market that is working well for its customers.
Databricks is not trying to be another CDP vendor. It is arguing that the CDP category itself should cease to exist as a separate product. The logic: if your data is already in Databricks, why send it somewhere else to run marketing? CustomerLake's AI agents can look at the full picture of a customer, including purchase history, product usage, support tickets, and financial transactions, and then decide what to offer, through which channel, and when, all without a data team needing to manually pull and transfer records.
The partner list at launch includes Adobe, Meta, Braze, LiveRamp, The Trade Desk, Twilio, and Snapchat, among others. Databricks is careful to position this not as a war against those companies but as a way to send better data to tools companies already use. Early customers include HP, Circle K, and Getnet by Santander, suggesting the product has been tested at genuine enterprise scale.
The honest caveat: CustomerLake is in private preview, meaning it is not yet available to everyone and the real-world results are still early. The CDP market has a long history of products that promised to solve fragmentation but added more of it instead. That said, Databricks has a structural advantage that pure CDP vendors do not: the data is already there. Companies using Databricks for finance, operations, and supply chain analytics already have their customer data inside the platform. CustomerLake is asking those companies to stop exporting it.
For any business operator who currently has a marketing team waiting on data team tickets, or who is paying for multiple overlapping tools to answer basic questions about their customers, this product is worth watching closely. The private preview is open now; contact Databricks directly if your company already runs on their platform.