The Real Cost of Data Lock-In — and How TextQL Plans to Break It

TextQL challenges the expensive data platform status quo with natural language querying that works across systems at enterprise scale.

Enterprise data platforms have a comfortable business model. They make it expensive to leave. TextQL wants to change that.

The three-year-old startup built a system that lets business users query data across multiple platforms using natural language. But they're not just another company promising to make SQL easier. They're attacking the high-margin business models of platforms like Snowflake, Databricks, and SAP. Here's what I learned during the 64th IT Press Tour.

The Lock-In Problem

Moving enterprise data between systems costs millions of dollars. Not because the technology is particularly complex, but because the platforms make it that way.

"If you're a CFO and you want to move business logic from SAP to NetSuite, you're looking at a contract with Accenture for $50 million over five years," said Ethan Ding, TextQL's CEO. "And that only moves about 10% of your actual business logic."

Each platform has its own query language and table format. Moving data requires changing formats, rewriting queries, and finding containers large enough to hold everything during the transfer. The platforms charge a markup of 10 to 20 times on their underlying infrastructure costs because they know customers can't easily leave.

This creates a strange economic situation. Today, data platform costs represent about 10% of a company's IT budget. Most of the cost is salaries. But as AI coding agents make data scientists more productive, unit costs for data work will drop. That drives up demand. More demand means companies will spend more on data platforms — potentially reaching 50% of IT budgets.

"When data platform costs become 50% of your spending, those 20x markups become a lot less cool," Ding said.

What TextQL Built

TextQL created what they call a "Rosetta Stone system" for enterprise data. It translates between different query languages and table formats through an intermediary layer. This means you can query data across multiple systems without migrating everything to one place.

The system handles scale. Hundreds of thousands of tables. Trillions of rows. Petabytes of data. It reconciles when the same customer appears differently across systems — "Fred" in one database, "Fred Frankel" in another, "frank.f@gmail.com" in a third.

During a demo at their New York office, Ding showed how a business user could ask complex questions about flight data without writing any code. The AI agent wrote queries, caught errors, tried different approaches, and produced visualizations. All in natural language.

"You should be able to ask questions and have the AI figure it out," Ding explained. "Instead of waiting weeks for someone to build a dashboard, you get answers in minutes."

The Business Case

The cost difference is significant. Traditional data platforms charge based on hours of CPU running. TextQL charges based on compute usage — only for the minutes when data is actually in transit.

"We can spin up a really big computer on the fly," Ding said. "You don't have to pay the database provider for hours of runtime. You pay us for minutes in transit."

The bigger advantage is speed. A data scientist might analyze the cheapest houses to buy each month by state. With TextQL, you can run that analysis per city per week or per zip code per day. You can get alerted the moment something changes in your data.

This is how high-frequency trading firms operate. They make thousands of tiny, data-driven decisions every second. TextQL wants to bring that same capability to every business decision.

Target Customers

TextQL focuses on CFO offices, financial services, and healthcare organizations. These sectors have massive amounts of structured data spread across multiple systems. They also have strict compliance requirements around HIPAA and FINRA.

"Security and governance are huge in these areas," said one of the attending journalists. Ding confirmed that TextQL runs in customer environments and works with whatever security measures they already have in place.

The company can deploy on-premises or in the cloud. Their most mature deployments run on AWS, but they're building out support for Azure and other platforms.

The Journey

TextQL didn't get here easily. They started in December 2022 and rewrote their entire codebase seven times. The first version failed. So did the second and third.

"We lost every customer until January of this year," Ding admitted. "But since January, we haven't lost a single customer or pilot."

The company raised $5 million and built a 15-person team of senior engineers. They started serious customer conversations in June. Enterprise sales cycles in this space typically take 12 months. TextQL shortened theirs to six.

They project $7 to 7.5 million in revenue and recently hired experienced enterprise sales executives who've worked with Fortune 500 companies for decades.

Why This Matters

Natural language querying isn't new. Companies have been promising it for 35 years. SAP's Business Objects claimed to offer it in the early days. It wasn't actually natural language — it was memorizing keywords and typing them perfectly.

Every few years, another wave of startups promises to solve this problem. Most focus on small datasets or documents that fit on your computer. TextQL focused on the hard problem: enterprise-scale data across multiple systems.

"If you connect me to a database of infinite complexity, I can answer questions your best data scientist can't answer in 15 minutes," Ding said. "Put us in a room with competing software. Do a one-to-one comparison."

The company's confidence comes from solving the technical challenges. They can handle the scale. They can work across systems. They can make it affordable.

Whether they can break the lock-in that major platforms have spent decades building — that's the real question. But if they can deliver on their claims, they're offering something enterprises badly need: a way to use their data without being held hostage by their data platforms.