Databricks to buy A16z-backed machine learning startup Tecton

Databricks will acquire Tecton, a machine learning startup, to boost its AI agent offerings, CEO Ali Ghodsi told Reuters.

Tecton provides software for large-scale, low-latency data analysis and deployment and was last valued at US$900 million in 2022.

Financial terms were not disclosed, but the deal will be paid in Databricks’ private shares.

Founded in 2020 by former Uber engineers, Tecton has raised US$160 million from investors including Sequoia Capital, Kleiner Perkins, Andreessen Horowitz, and Bain Capital Ventures.

Ghodsi said Tecton’s technology could help improve response times in Databricks’ Agent Bricks platform, used to build and automate AI workflows for enterprises.

🔗 Source: Reuters

🧠 Food for thought

1️⃣ High-valued private companies increasingly use equity as acquisition currency

Databricks’ acquisition of Tecton demonstrates how companies with soaring private valuations can use their own shares as bargaining chips rather than depleting cash reserves.

The deal structure uses Databricks’ private shares to acquire Tecton, which was valued at $900 million in 20221. This approach becomes particularly attractive when a company’s valuation jumps dramatically. Databricks saw its value increase significantly to more than $100 billion in just eight months1.

This equity-based strategy allows rapidly growing private companies to preserve cash while still competing for acquisitions against well-funded rivals.

The approach also helps sellers by giving them exposure to potential future upside if the acquirer continues growing, rather than a fixed cash payout that might look modest compared to the acquirer’s trajectory.

2️⃣ Real-time data processing becomes critical bottleneck for AI applications

Databricks’ focus on Tecton’s low-latency capabilities highlights how speed has become the new competitive battleground for enterprise AI tools.

CEO Ali Ghodsi specifically cited response time as a “top priority for customers building interactive services,” noting that “humans hate to wait” when using AI applications, particularly for voice interactions1.

This emphasis on speed reflects a fundamental shift from traditional batch processing approaches to real-time AI that can respond instantly to user queries.