Two former SpaceX engineers are bringing software to help launch rockets from the factory floor with Sift Stack.

The cry of “atoms, not bits!” — a phrase that captures Silicon Valley’s growing obsession with physical manufacturing for digital products — reached a fever pitch last week with news that Jeff Bezos was creating a $100 billion fund to get factories up and running and automated.

But factory automation is not purely a hardware problem. We increasingly rely on sophisticated software and AI tools, and these changes are reshaping the companies that build the infrastructure of the physical manufacturing world.

Karthik Gollapudi, CEO of Sift Stack, an El Segundo, California-based company that provides tools to support the design and manufacture of complex machines such as spacecraft and automobiles, sees a change. He said the changes have shifted the company’s focus over the past six months.

Gollapudi and his co-founder, CTO Austin Spiegel, started the company in 2022 after working on SpaceX’s software tools to manage massive amounts of telemetry data – real-time performance information streaming from sensors on physical components during testing, manufacturing and launch.

Most companies building advanced machines use off-the-shelf database tools or write their own Python scripts, but Sift saw an opportunity to provide companies with best-in-class tools. Customers range from major U.S. rocket manufacturer United Launch Alliance and other defense contractors to robotics and power grid management startups.

But Gollapudi says the emergence of AI tools for data analysis has changed his business. The kind of custom workflows that once stood out as the company’s signature products have become important in the world of AI and deep learning models. But suddenly a company’s ability to manage its data infrastructure has become more important.

“Our long-term vision of how this would play out over five years is actually being put into action this year,” Gollapudi told TechCrunch.

This means managing the intensive data flows that occur in today’s software-intensive systems. Some of the vehicles the company works on have more than 1.5 million sensors streaming data simultaneously across multiple formats and time scales.

The company’s goal is to organize and store that data for AI applications. “There is a lot of value in exposing this data to be machine-readable,” Gollapudi said. Whether you want an AI agent to make decisions about manufacturing or analyze test data to flag potential problems, Sift’s goal is to make that data available.

Jeff Dexter, vice president of software at Astranis, a satellite company that uses Sift to manage testing, manufacturing and operations, said a good data infrastructure is critical for a company that performs 10 million automated software tests a day.

“Inevitably, we got to the point where it was costing us millions of dollars each month just to store our data,” Dexter said. “Really, it’s like whether a million dollars is well spent. With technology like Sift, you don’t have to worry about how much data you have.”

Gollapudi told TechCrunch that Sift has raised a $42 million Series B in 2025 at a post-money valuation of $274 million, led by StepStone with participation from GV (Google’s venture arm), Riot Ventures, Fika Ventures, and CIV.