Atomera Inc (ATOM) stock intrinsic value fair value estimate & owner earnings analysis

Atomera (Tempe, Arizona) develops and licenses Mears Silicon Technology (MST), a thin film of re-engineered silicon that chipmakers deposit on wafers to make transistors faster, more reliable and more power-efficient. It does not make chips. Its model is to charge a chipmaker an R&D licence fee when MST is installed on a tool in the customer's fab, a larger high-volume manufacturing (HVM) licence fee once the customer qualifies MST in its process, and then royalties on every MST-enabled product sold. Atomera runs its own development work on two leased epitaxial deposition reactors in Tempe: an Applied Materials Centura and an ASM Epsilon. Where the business stands. Ten years after its 2016 IPO, no customer has reached the HVM or royalty stage. Revenue was $65,000 in 2025 and $135,000 in 2024, and has never been above about $550,000 in a year. The only commercial licence, signed with ST Microelectronics in April 2023, stalled in October 2025 when ST said it would not finish qualifying MST because it was moving the target process to 300mm wafers. Atomera also has two joint development agreements. The first, with a leading semiconductor provider, has MST installed in the customer's fab under a paid manufacturing licence and has met its development milestones. The second, with a major foundry, is still running test wafers. Integration licences with a fabless RF chip company, a foundry and Asahi Kasei Microdevices are in evaluation. None of these agreements commits a customer to production. In April 2025 Atomera signed a marketing agreement with a large chip-equipment maker aimed at leading-edge logic and DRAM customers. The 10-K itself says the company will keep burning cash until customers sign HVM licences and ship royalty-bearing products in volume. Earnings power. We measure owner earnings as operating cash flow, less working-capital swings, stock compensation and the cost of keeping the business running. For Atomera that cost includes the payments on its reactor finance leases, which are really equipment purchases paid over time. On that basis the business has lost money every year: between $9.8 million and $20.4 million a year since 2016, and $20.4 million ($0.66 a share) in 2025. Over 2021-2025, the years with today's leased-tool cost base, the median was about $0.69 a share. The losses are getting bigger. The 2025 loss was about 10% larger than 2024's (per share it looked slightly better only because more shares were issued), and the first half of 2026 ran at an annualised loss of about $24 million. Management says it plans to spend more on R&D. We therefore assume no improvement in the first year. The losses have been paid for by selling new shares: weighted diluted shares grew from 12.7 million in 2018 to 30.8 million in 2025. There were 39.0 million shares outstanding at June 30, 2026, after a $25 million sale of 5 million shares at $5.00 in February 2026 and continued at-the-market sales. The company says it has enough capital for at least 12 months from its February 2026 annual report. What we can and cannot credit. At June 30, 2026 Atomera held about $38.4 million of cash and short-term investments against about $3.1 million of total liabilities. It has no borrowings beyond equipment leases. Cash spending was $16.1 million in 2025 (operating outflow of $14.9 million plus $1.2 million of lease payments) and about $8.9 million in the first half of 2026. Licence fees and royalties would change everything, but no HVM licence, royalty rate or fee schedule has been signed or disclosed. After ten years without one, we treat future licensing as unproven, so none of the three cases puts a value on royalty income. Bear case, $0.00 a share. Spending continues at the first-half 2026 pace, and the planned R&D increase lifts it further. No licence is signed. The cash runs out around 2028. Continuing after that depends on selling more shares, which leaves today's shareholders with essentially nothing. Liabilities are small, so the company does not appear to owe more than it is worth. The equity is simply used up. Base case, $0.03 a share. The development programme continues for about two more years at a burn rate between the 2025 and 2026 levels. Licence revenue stays immaterial, as it has been. What is left after paying liabilities, discounted to today, is a few cents a share. Bull case, $0.90 a share. The company's own pipeline delivers. The first development partner, which already makes MST wafers in its own fab, converts to a distribution and royalty agreement, and licence fees plus early royalties roughly cover spending. Shareholders then keep today's net cash of about $35 million. Any royalty profit above break-even is not counted, because no terms exist to value it. What would make this wrong. A signed HVM licence with disclosed fees and royalty terms, from either development partner, ST or a customer brought in through the equipment-maker partnership, would give royalties a basis for valuation and could raise the value well above these figures. On the other side, each year without one uses up roughly $17-18 million of cash and adds more shares. Customer concentration is high: one customer provided 77% of 2025 revenue and another 23%. The film recipe is kept as confidential know-how, which the company notes does not expire the way patents do. That protection depends on licensees keeping their confidentiality obligations. Figures for the first half of 2026 come from the company's tagged financial data. The full text of that quarterly report was not available to us.