With Sift Stack, two former SpaceX engineers are bringing back the software that helped launch rockets on the factory floor.

Shouts of “atoms, not bits!” – a statement about Silicon Valley’s obsession with digital innovation – reached its peak last week with word that Jeff Bezos was forming a partnership. $100 billion expenses for the purchase and maintenance of factories.

But the manufacturing industry is not just a hardware problem. It relies heavily on advanced software and AI tools, and this shift is reshaping the companies that are building the foundations of the world of manufacturing.

Karthik Gollapudi, CEO Sift StackThe El Segundo, California-based company, whose facilities help design and manufacture complex systems such as aerospace vehicles and automobiles, is seeing a radical change. He says that this change has changed the mindset of his company in the last six months.

Gollapudi and his co-founder, CTO Austin Spiegel, founded the company in 2022 after working on software tools at SpaceX that managed massive amounts of telemetry data — real-time information from sensors on physical objects — during testing, manufacturing, and launch.

Many companies that build advanced systems use their own archive tools or cook their own Python scripts, but Sift saw an opportunity to provide companies with a more advanced tool. Customers range from United Launch Alliance, the largest US rocket builder, and other defense contractors, to robotics and power grid management startups.

However, Gollapudi says the advent of AI tools for data analysis forced a change in his business. The types of customized movements that were once known as the company’s signature are now on the table in the world of AI and deep learning models. But the company’s ability to improve data performance suddenly became more important.

“Our long-term view of how we saw this in five years is playing out this year,” Gollapudi told TechCrunch.

This means managing the volume of data from today’s systems using software. Some of the vehicles the company works with have more than 1.5 million sensors moving simultaneously, across multiple modes and time scales.

Organizing and storing data for AI applications is the company’s goal — “the most valuable information is revealing that it can be read by machines,” Gollapudi said. If AI agents are making decisions about design or analyzing test data to identify potential problems, Sift’s goal is to make that data available to them.

Jeff Dexter, VP of software at Astranis, a satellite company that uses Sift to manage testing, production, and operations, said good data tools are needed for companies like his that can test 10 million applications a day.

“Inevitably, it ends up costing us millions of dollars a month to store data,” Dexter said. “It’s like, did this cost a million dollars? With technology like Sift, I don’t care how much data there is.”

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