Microsoft-Backed Lace Raises $40 Million to Build Chips 10x Smaller Than Today’s Best — Helium Atom-Beam Lithography Could Reinvent Semiconductors

Lace, a Norway-based semiconductor equipment startup backed by Microsoft, raised $40 million in funding in March 2026 to develop helium atom-beam lithography — a novel chip fabrication technique that the company claims could enable semiconductor designs approximately ten times smaller than those enabled by the current best-in-class extreme ultraviolet lithography used by TSMC, Samsung, and Intel. The company, founded by researchers with backgrounds in physics and precision engineering, is targeting a pilot chip fabrication plant by 2029. TechCrunch reported the funding, noting the broader strategic context: semiconductor progress is increasingly dependent on breakthroughs outside the traditional incumbent players — ASML, TSMC, and the established node progression — and the AI hardware race is creating both the financial incentive and the strategic urgency for new approaches to chip manufacturing at previously unachievable scales. Current state-of-the-art semiconductor fabrication uses extreme ultraviolet light to print circuit patterns at 2-nanometre scale. Helium atom-beam lithography uses streams of individual helium atoms rather than light waves to etch features into semiconductor material, theoretically enabling pattern precision at scales that optical lithography cannot achieve due to the fundamental physics of light diffraction. Microsoft’s backing is strategically notable: the company is simultaneously the world’s largest enterprise cloud provider and one of the primary buyers of advanced AI chips from NVIDIA, AMD, and Intel, giving it a direct commercial interest in future semiconductor manufacturing breakthroughs that could reduce costs or expand performance for AI accelerators. For the broader IT industry, Lace represents the category of deep-tech infrastructure investment that has accelerated significantly since the AI boom: capital is now flowing not just to AI applications and models but to the foundational physics-level research that will determine what AI hardware looks like in 2030 and beyond.