Ex-SpaceX engineers bring rocket-launch software to factory floor with Sift Stack

Last week, the rallying cry “atoms, not bits!” — a phrase that captures Silicon Valley’s increasing focus on physical manufacturing rather than digital products — reached a peak when news broke that Jeff Bezos is assembling a $100 billion fund to consolidate and automate factories.

However, factory automation is not solely a hardware challenge. It increasingly relies on advanced software and AI tools, and this shift is transforming the companies building the infrastructure for physical manufacturing.

Karthik Gollapudi, CEO of Sift Stack — an El Segundo, California-based company whose tools aid in designing and manufacturing complex machines such as spacecraft and cars — senses the ground shifting beneath his feet. He says these changes have reoriented his company’s priorities over the past six months.

Gollapudi and his co-founder and CTO, Austin Spiegel, launched the company in 2022 after working on software tools at SpaceX that handled enormous volumes of telemetry data — real-time performance data streamed from sensors on physical components — during testing, manufacturing, and launch.

Most companies that build advanced machines rely on off-the-shelf database tools or custom Python scripts, but Sift identified an opportunity to offer a best-in-class solution. Its customers include United Launch Alliance, a major U.S. rocket builder, other defense contractors, as well as robotics and power grid management startups.

However, Gollapudi says the emergence of AI tools for data analysis compelled a shift in his business. The customized workflows that once defined the company’s signature offering have become table stakes in an era of AI and deep learning. Meanwhile, the company’s ability to manage data infrastructure suddenly became far more valuable.

“Our long-term vision of how this would unfold over five years is actually happening this year,” Gollapudi told TechCrunch.

This means handling the massive data flow from today’s software-intensive machines. Some vehicles the company works with have over 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 — “a lot of the value is in making it machine-readable,” Gollapudi said. If AI agents are to make decisions about manufacturing or analyze test data to flag potential issues, Sift aims to make that data accessible to them.

Jeff Dexter, VP of software at Astranis — a satellite company that uses Sift to manage testing, manufacturing, and operations — said that robust data infrastructure is critical for companies like his, which may run 10 million automated software tests in a single day.

“Inevitably, we reach a point where it’s costing us millions of dollars each month just to store data,” Dexter said. “It really becomes a question of whether that million dollars is well spent. With technology like Sift, I don’t worry about how much data is there.”

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