Tag: Atomic Semi

  • Jim Keller’s startup is building a factory to mass-produce small semiconductor fabs —Atomic Semi rebrands as ‘Fab2’ underlining intended role as a ‘fab fab’

    Featured image Jim Kellers startup is building a factory to massproduce small semiconductor fabs Atomic Semi rebrands as Fab2 underlining intended rol

    The future of silicon manufacturing isn’t about gargantuan factories; it’s about distributed innovation, and a startup named Fab2 is leading the charge by reimagining how microchips are made. Founded by chip architect Jim Keller and DIY fabrication pioneer Sam Zeloof, this company has successfully rebranded from Atomic Semi to capitalize on its revolutionary concept: the “fab fab“—a system designed to mass-produce small semiconductor fabrication facilities and all the complex tools contained within them.

    Fab2 isn’t just a design firm; they are building the entire ecosystem. The company designs and engineers every component necessary for a chip factory in-house, from precision pumps and valves to advanced lithography systems and vacuum chambers. They assemble these components into sophisticated machines, which then form complete fabs, aiming for mass production of these specialized manufacturing units.

    This approach shatters the traditional model of semiconductor production, which typically involves moving 300mm wafers through massive production lines. Instead, Fab2 targets small, software-defined fabs capable of prototyping chips far smaller than a wafer in mere hours. This shift is rooted in Sam Zeloof’s early pioneering work, where he demonstrated the viability of fabricating lithographic chips down to roughly 300nm features in his parents’ garage long before co-founding the company with Keller.

    The core challenge in this high-speed, small-scale manufacturing model lies in throughput. While the method utilizes electron-beam lithography—which writes patterns directly rather than projecting them through a mask—it inherently makes the process slower than some massive industrial techniques like EUV scanning. This trade-off means Fab2 excels in prototyping and low-volume runs, positioning them perfectly for innovation rather than high-volume commercial foundry production.

    To realize this vision, Fab2 has established a dynamic operational footprint spanning three sites: an expansive 120,000 square foot facility in Austin serves as the hub for research and production; a 30,000 square foot site in Lockhart houses the actual “fab fab“; and the original 25,000 square foot “garage fab” remains in San Francisco.

    Backed by significant capital and visionary investors—including Naval Ravikant, Nat Friedman, and Fred Ehrsam—the startup secured a reported $15 million seed round in 2023, led by the OpenAI Startup Fund. The company has attracted talent, currently employing around 84 people as of May 2026, demonstrating the appetite for this novel approach to chipmaking.

    Fab2’s model offers a compelling alternative to industry behemoths like Tesla and SpaceX, which are pursuing massive, centralized megafabs such as Terafab. While one aims for a single Austin facility targeting a terawatt of compute, Fab2 champions a decentralized solution. They represent two competing blueprints for expanding U.S. chipmaking capacity: consolidate production into giant campuses, or distribute it across many small, replicable, and agile fabrication units.