TextQL

The bet

Enterprises have data spread across warehouses, BI tools, SaaS products and older systems. TextQL’s Ana agent aims to answer business questions across that landscape without first forcing a costly migration. Its documented stack combines Ana, an enterprise Ontology for business context, and Sandcastles, a sandboxed execution environment. The product emphasis is access, context and governed execution—not simply turning a prompt into SQL.

The deeper thesis is that models will change faster than enterprise definitions do. If TextQL’s ontology becomes the trusted, maintained layer connecting data, business meaning, permissions and whichever models a company uses, it could be unusually sticky. That is the opportunity I recognize from Diageo. It is still a hypothesis: customers must keep that context current in TextQL, and platform vendors are building their own semantic layers.

The founders

CEO Ethan Ding has worked in data roles at Tackle and Bessemer; CTO Mark Hay worked on Meta’s Integrity Infrastructure team and its Sigma and Haxl projects. Ding has said the team rearchitected the product nine times in its first two years. That suggests high iteration speed, alongside the risk of a product still finding its repeatable shape. (Ding profile; Hay profile; Ding on iteration.)

Evidence of use

The best public signal is named customer usage. TextQL reports that Scale AI has sent 28,000+ messages, started 11,500+ threads and runs 42 playbooks in production. Dropbox’s Director of Revenue separately described producing a CFO-ready quarterly update in under an hour, versus most of a day before. These are encouraging references, though the usage figures and case studies are company-published. (Scale; Dropbox finance leader.)

Blackstone is both a customer and the lead investor in TextQL’s announced $17m April 2026 financing. That is a meaningful reference and strategic relationship. It does not establish a committed sales channel into Blackstone’s portfolio. TextQL also describes production deployments with Lumeris and PAR Technology, including embedded analytics, a potentially valuable route to downstream users. (financing; Lumeris; PAR.)

CEO Ethan Ding has publicly claimed 9× year-over-year revenue growth and six-figure contracts over six consecutive weeks. Our Frontier underwriting also includes revenue diligence whose figures cannot be shared in a co-invest memo. The qualitative view is that growth is strong, with some moderation expected. (CEO’s post.)

Commercial shape and competition

TextQL’s AWS Marketplace offer lists a $100,000 annual platform commitment for up to four million agent-compute units, plus metered overage. That is a concrete enterprise price point, not proof of realized ACV. Public self-serve pricing is far lower; dedicated, VPC and on-prem deployments, SSO, SLAs and migration help are enterprise features.

Deployment depth could create durable customer context and workflow adoption. It could also make onboarding labor-heavy. TextQL hires for custom proofs of concept, connector setup, security reviews and post-sale deployment; if each account needs substantial engineering, growth and gross margin will be harder to scale.

Incumbents are moving quickly. Google’s BigQuery Conversational Analytics is generally available across multiple data sources; Microsoft Fabric agents combine several sources; Databricks Genie and Snowflake Cortex Agents add governed natural-language analysis inside their platforms. TextQL’s opening is customers with fragmented, cross-cloud or legacy estates and embedded use cases. Its moat must come from better deployment, context, security and outcomes—not the chat interface alone. (Google; Microsoft; Databricks; Snowflake.)

I underwrite a strategic acquisition path, not a journey to $1bn of ARR. Potential homes exist up and down the stack: cloud and data platforms, BI and governance vendors, and enterprise software companies seeking a governed interface across their products. The AWS Marketplace listing makes TextQL purchasable alongside other data products. One adjacent strategic option there is especially interesting. Marketplace presence supports procurement, but does not establish an AWS channel commitment or buyer interest. Any acquirer would still weigh buying TextQL against building or partnering.

Entry and exit

TextQL announced a $4.1m pre-seed and seed in January 2024, then a $17m strategic round led by Blackstone in April 2026; public databases report $21.1m raised before the financing now under review. The terms shared with me are $30m at $100m pre-money: $130m post-money, with new investors buying about 23.1% before any option-pool changes. The five-month interval makes it important to confirm whether this is a new round or an extension and how much is already committed. (seed announcement; April round.)

The return hurdle remains high even with an acquisition thesis. After another 25% dilution, a 10× gross return requires about $1.73bn in equity value at exit, before preference effects. A smaller takeout could validate the product and still produce a modest venture return.

View: starter-sized venture bet. Production usage, strong revenue growth and strategic relevance make the right tail credible. I expect growth to moderate; the case is that TextQL’s ontology, integrations and customer foothold become valuable to a strategic buyer before it needs to reach billion-dollar ARR. The main risks are incumbent bundling, implementation intensity and whether a buyer values the capability more than building it internally. The $100m pre-money terms still require a clean preference stack and clarity on option-pool dilution.