Aeluma is a development-stage semiconductor company in Goleta, California, trying to do something the incumbents say is hard: manufacture high-performance photonics on large-diameter, mismatched substrates (up to 300 millimeters) instead of the expensive, supply-constrained indium-phosphide wafers the industry defaults to. The pitch is that this makes the components at the heart of AI data-center interconnects - high-speed photodiodes, shortwave-infrared imaging arrays, quantum-dot lasers - cheap enough to scale to mass markets. That is the whole investment case, and it is a case built on milestones rather than the profit-and-loss statement. The March quarter delivered a familiar development-stage pattern: revenue essentially flat at about $1.2 million, a gross margin that compressed as material costs tied to government-funded deliverables came through, and a GAAP net loss that widened to $(1.8) million. There is nothing in those numbers that justifies a roughly $330 million market value on its own.
What justifies the market value is the optionality. As of March 31, Aeluma held $37.8 million in cash against no debt and a $40 million stockholders' equity, with a fully untapped $50 million at-the-market facility and, two months after the quarter closed, a letter of intent with the U.S. Department of Commerce for up to $30 million of CHIPS research-and-development funding for its non-InP photonics platform. The quarter itself was a financial step backward; the platform funding path is the story. Investors at roughly $18 are not paying for what the company earns. They are paying for the chance that a scaled, AI-datacom photonics business exists here at all. The stock, which ran from a low near $10.70 in early April to a 52-week high near $31.50 in mid-May and then pulled back, has rebounded into the high-teens on the CHIPS news. That volatility is the second half of the thesis: almost nothing about this name is priced for a gentle outcome.