Quantum’s Cost Wall
Last quarter, IonQ booked $65 million in revenue.
It lost $271 million.
That’s an operating loss roughly four times its revenue. And IonQ is the strong one — it’s sitting on $3.1 billion in cash, the longest runway in the business.
Rigetti, the same quarter: $4.4 million in revenue, a $26 million operating loss. Six dollars lost for every dollar earned.
These are public companies, with multi-billion-dollar valuations, selling a product that doesn’t have unit economics yet. Let me show you the wall they’re walking into.
Two different walls
Quantum computers come in flavors, and the cost structure depends on the flavor.
The superconducting machines — IBM, Google, Rigetti — have to be cooled to within a hair of absolute zero. That takes a dilution refrigerator.
These are not mini-fridges. IBM’s large-scale unit reportedly costs over $800,000, with electricity bills north of $100,000 a year. The refrigeration for a 500-qubit-class system can top $2 million per unit.
And the cooling runs on helium-3 — an isotope so rare it sells for anywhere from $1,000 to $20,000 a liter, produced in meaningful quantity by exactly two governments.
The trapped-ion machines, like IonQ’s, skip the dilution fridge. Different physics, no helium-3 wall.
But notice: IonQ avoided the cooling wall and still lost $271 million on $65 million. So dodging one wall doesn’t save you. The economics are brutal across the board.
Why “more qubits” makes it worse, not better
Here’s the part that turns a hard problem into a wall.
The qubits these machines have today are error-prone. To get one reliable qubit — a logical qubit, the kind you could actually compute with — you need many physical qubits correcting each other. Often hundreds or thousands.
So when a company says “we’re scaling to more qubits,” read it as: we’re multiplying the most expensive parts of the machine.
More physical qubits means more cooling. More cooling means more dilution refrigerators, more power, more helium-3.
The cost doesn’t fall as you scale. It compounds. You’re stacking the scarcest isotope on top of the most expensive fridge, to build error correction that eats most of the qubits you add.
That’s the cost wall. It’s not that quantum can’t scale. It’s that scaling it, with today’s approach, makes the bill grow faster than the capability.
What the “revenue” actually is
The last honest question: what are these companies even selling?
Mostly, not commercial value. The revenue is government contracts, research grants, and cloud-access fees — people renting time on machines that can’t yet beat a regular computer at anything that matters commercially.
That’s not a product business. It’s a funded science program with a ticker symbol.
Which can be a perfectly good thing to fund. The science is real, and it might genuinely change the world in a decade. I’m not mocking the physics.
I’m pricing the equity. And the equity is valued like an industry, while the income statement reads like a lab.
Why this matters before the paywall
When quantum stocks dropped 7–9% in a single session this May, the headlines called it “profit-taking.” On companies with no profit. That phrase alone tells you how detached the prices have drifted from the economics.
Behind the paywall:
The burn-to-revenue ratios across the listed quantum names, lined up — and how many quarters of runway each actually has at current loss rates.
The full cost stack of a logical qubit: fridge, power, helium-3, and the error-correction multiplier — and why that stack, not the algorithm, is the real bottleneck.
The one input that could break the whole superconducting roadmap if it stays scarce — and the trapped-ion and photonic approaches that route around it.
And the honest answer to the only question that matters for an investor: not “will quantum work?” but “will any of these companies still be the ones that own it when it does?”
If you’ve read this far, you can see the wall. The paid section measures how far away it is…