IonQ, Inc. IONQ

Key data

The business

IonQ builds quantum computers by trapping individual charged atoms and manipulating them with lasers, an approach that gives it the field's cleanest error rate: it hit 99.99% two-qubit gate fidelity in 2025, a world record. It sells two ways: cloud access to its machines through Amazon, Microsoft and Google, and full systems like the new Tempo. Partners named in its own release include AWS, AstraZeneca and Nvidia, working on drug discovery, materials and defense. It has also been buying companies, folding in a SkyWater chip-fabrication business and a market-intelligence unit, so the revenue line now mixes high-concept quantum access with an ordinary foundry whose orders are cancellable until production and which carries little backlog. The moat is a genuine hardware lead in fidelity plus a patent thicket. The hole in it is that quantum computing still has no proven killer application anyone pays real money to run.

The numbers

Read the operating line, not the bottom line, because the bottom line is noise. Revenue is climbing hard and steadily; the loss from actually running the business is widening just as fast.

Revenue rose 287% year over year in Q2. But look across the net income column: a $753 million profit, an $805 million profit, then an $1.87 billion loss, on a business doing $80 million a quarter. That is not operations, it is the mark-to-market on warrant and derivative liabilities swinging with the share price, a non-cash number that reverses every quarter. The two consensus "beats" this year, reported EPS of $2.07 and $1.93 against a forecast near negative $0.46, were entirely these accounting gains. The clean line, operating loss, has gone from $161 million to $337 million a quarter in a year and points one direction.

Half of 2026 already out-sold all of 2025. Now the price. At $14.8 billion the stock costs 60 times its last twelve months of sales. To grow into that at even 10 times sales, generous for an unprofitable hardware name, IonQ needs about $1.5 billion of revenue; the Street's own model shows $932 million in 2029, which at 10 times is $9.3 billion, below today's value. So the math only works if the multiple stays extreme for years or revenue blows past what analysts see, and gross margin is just 36% because the foundry mix drags it down. What this memo believes that the tape does not: the reported profits are a mirage, and the print that settles it is the operating loss, which will keep widening while revenue climbs.

Management

Insiders are one-way sellers: about $9.4 million sold across 15 sales in twelve months against $224,000 bought in two, led by Rima Alameddine's $4.7 million in November 2025, plan status not disclosed on the Form 4s. The deeper tell is stock-based compensation running at 169% of revenue, meaning the company pays its people with far more stock than it collects in sales, and funds the cash losses with repeated equity raises visible in the cash balance jumping from $54 million to $1.24 billion inside eighteen months. Management is a capable engineering team and a serial acquirer paying in its own shares. Their guidance credibility is thin: the only "beats" were warrant gains, and on the line that matters they miss and widen.

How it fails or surprises you

Dilution eats the holder. The business burns cash and pays for it by issuing stock; share count and the 169%-of-revenue comp bill grow every quarter. Even if revenue triples, your slice shrinks. The print that shows it first: diluted shares outstanding and the SBC line, quarter over quarter, rising faster than anyone models.

The growth is bought, not grown. A chunk of the 287% jump is the acquired SkyWater foundry, a low-margin business with cancellable orders, folded into a "quantum" story. Strip acquisitions and organic quantum revenue may be a fraction of the headline. The reveal: an organic-versus-acquired disclosure, or gross margin sliding further below 36%.

Quantum advantage lands (right tail). The 99.99% fidelity lead converts the $485 million of remaining performance obligations into a named, large-scale enterprise or government contract with real margin, and the category re-rates on proof rather than hope. The first sign: RPO stepping up sharply alongside a gross-margin inflection, not another partnership press release.

Closing thoughts

Nothing in the next four quarters resolves this. No single print tells you whether quantum computing becomes a real industry, and the reported earnings are worthless as a signal because warrant accounting scrambles them. What matters is survival and the pace of dilution: IonQ has $1.24 billion of cash and almost no debt, so it will not fail soon, but it is spending that hoard and its own shares to buy time for a technology whose payday has no date. The left tail is not bankruptcy, it is a decade of dilution while the killer app never arrives and the 60-times-sales premium bleeds out. The right tail is real and large if the fidelity lead ever cashes in, but you are paying today for an outcome no customer has yet paid IonQ enough to prove. On the evidence, the downside is fatter, because the burn is certain and the payday is not.

The bet is still that quantum computers start doing paid work and that IonQ's trapped-ion machines are the ones customers keep buying, funded until then by printing stock. What breaks it is the gap between the two numbers that count: revenue climbing while the operating loss climbs faster, and the share count rising underneath both. Watch those three lines together. If revenue growth ever stalls before the operating loss turns, the story is over regardless of what the fidelity record says.