Rigetti Computing, Inc. (RGTI) stock intrinsic value fair value estimate & owner earnings analysis

Rigetti Computing builds superconducting quantum computers. It designs and fabricates its own chips at its Fab-1 facility, sells 9- to 108-qubit systems under the Novera and Cepheus names, and offers access to its Ankaa-3 and Cepheus-1-36Q machines through its Rigetti Quantum Cloud Services platform. The technology is advancing (the company reports a 99.6% median two-qubit gate fidelity on Cepheus-1-36Q as of January 2026), but the business does not yet earn money. What the business has actually produced: revenue fell every year as a public company, from $13.1 million in 2022 to $12.0 million, $10.8 million and $7.1 million in 2025. There were no Novera sales in 2025, and the company says revenue was hurt by the lapse of the National Quantum Initiative Act in September 2023. About 90% of 2025 revenue came from government entities, and one customer alone was 42%. Revenue picked up in the first half of 2026 to $9.5 million from $3.3 million a year earlier, and the company received orders for two Novera systems worth about $5.7 million in 2025, but contracted revenue not yet recognized was only $2.7 million at June 30, 2026. Against that, operating losses were $84.7 million in 2025 and $54.0 million in the first half of 2026 alone. How we measured cash earnings: we start from operating cash flow, strip out working-capital swings, and charge stock-based compensation and the capital spending needed to keep the business running. The reported net loss of $216.2 million in 2025 is inflated by a $150.6 million non-cash loss on the fair value of warrants, which we ignore because it is an accounting remeasurement, not cash. On our measure the business consumed $121.5 million in 2022, $71.6 million in 2023, $71.7 million in 2024 and $86.9 million in 2025. Excluding interest earned on its cash pile (we count the cash separately), the typical year over 2022-2025 consumed about $82 million, and the burn widened 32% in 2025. The first two quarters of 2026 point to roughly $110 million a year, an 18% widening; we use that as the starting level because two reported quarters confirm it. Nothing we can point to in the filings shows a date when revenue will cover costs: the company itself says it expects to keep incurring losses for the foreseeable future, it has agreed with Quanta Computer (February 2025) that each party will invest at least $250 million in quantum computing over five years, and it says it may significantly increase capital spending, possibly including a new chip fabrication facility. What the company owns and owes: at June 30, 2026 it held $541.3 million of cash and investments and has no borrowings (its Trinity Capital loan was prepaid in December 2024). The $78.4 million warrant liability on the balance sheet is not debt; warrants bring cash in when exercised (they brought in $50.0 million in 2025). We keep about $25 million as working cash and treat the rest, about $516 million, as available. We credit interest on that cash at about 4.1%, the yield it actually earned in 2025. No value is assigned to the technology or to future product sales, because no contract or track record yet shows they will produce profits. The three cases differ only in how fast the cash is spent. Base case: spending continues at the first-half-2026 rate of about $110 million a year until the available cash is used up, roughly six years. The spending is worth about minus $429 million today; with $541 million of cash, $25 million held back and no debt, about $88 million, or $0.26 per share, is left for shareholders. Bear case: the burn keeps widening about 20% a year, as it has recently, driven by the disclosed plans for more capital spending and the heavy dependence on government customers; the cash is gone in about four years. Spending worth about minus $441 million leaves about $76 million, or $0.22 per share. Bull case: the first-half-2026 revenue gains continue enough to hold the burn to its 2025 level of about $94 million a year, stretching the cash to about seven years. Spending worth about minus $415 million leaves about $101 million, or $0.30 per share. In every case the small positive value exists only because the cash is spent over several years rather than at once; once it is gone, further funding has come from selling new shares (about $347 million of at-the-market sales in 2025, and $97.5 million plus a $96.0 million direct offering in 2024), which lifted the share count from 283.5 million at the end of 2024 to 333.7 million by June 2026. Per-share figures use about 340 million shares including unvested stock units. No growth was credited in any case, because none is contracted or backed by a repeatable record: revenue has fallen for three straight years and depends on government budgets. What would make this wrong: sustained commercial orders large enough to cover more than $100 million of annual spending, which would give the technology an earnings value we cannot see today. Until that appears in reported results, the shares are worth roughly the cash that has not yet been spent. Margin-of-safety price: because nearly all of the value is cash that the company plans to spend, and the cases span only $0.22 to $0.30, the estimate requires a margin of safety of about 65% below the base case, a margin-of-safety price of $0.09 per share.