Four Quantum Architectures Now Public: Which Bet Wins?

I've been tracking quantum stocks for nearly three years, and what changed in the last six months is bigger than any single name. The number of quantum pure-plays jumped from four to seven, with three more in the IPO pipeline. But the architectural story is what really matters — for the first time in history, all four fundamentally different ways to build a quantum computer are publicly investable.

That removes the worst constraint quantum investors have lived with: being forced to bet on one approach. Now I can hold positions across architectures and let the technology decide.

Superconducting Path Splits Into Two Bets

D-Wave: The Only Public Annealer

D-Wave uses superconducting hardware, but with an annealing approach instead of gate model. Annealing is purpose-built for optimization — supply chain routing, drug discovery scheduling, financial risk modeling. Customers are paying for it today.

The numbers speak for themselves. 2025 full-year revenue grew 179% year-over-year. The detail that caught my eye: in January and February 2026 alone, D-Wave booked more new contracts than its entire 2025 annual revenue. And here's what most coverage missed — in January 2026, D-Wave completed an acquisition that adds gate model capability alongside annealing. Eight qubits operational now, 17 by mid-2026, 49 by 2027.

Rigetti: The Concentrated Gate-Model Bet

Rigetti uses the same superconducting substrate but in gate model form, the path IBM and Google follow. The concentration risk here is real. Two contracts represent 63% of Rigetti's 2026 revenue estimate — India's C-DAC organization for a 108-qubit system and Novara QPU hardware. Both delivery timelines have already slipped from H1 to H2 2026.

Quantum hardware delays come with the territory. But when more than half your revenue rides on two slipping contracts, quarter-to-quarter results stay volatile.

Trapped Ion: The Largest Pure Play

IonQ replaces superconducting circuits with individual atoms held in electromagnetic fields. Trapped ion is regarded as the most precise approach.

When I last covered IonQ in October, backlog was $77 million. It's nearly 5x that today. Revenue tripled year-over-year (up 202%), and the company announced a $1.8 billion acquisition of SkyWater Technology, a trusted U.S. semiconductor foundry. If that deal closes, IonQ would own end-to-end manufacturing on U.S. soil — a major advantage for Pentagon contracts. IonQ is now cleared to compete for U.S. defense quantum work.

Neutral Atom and Photonic: The Two New Architectures

Inflection: Nvidia Picked This One Twice

Inflection went public in February 2026 and brings the third architecture: neutral atom. The scaling argument is the strongest of any architecture — neutral atoms can reach thousands of qubits without the engineering headaches that plague other approaches.

What separates Inflection isn't the press release. Nvidia selected the company for two separate quantum programs. Out of every quantum company on the planet, Nvidia chose this one twice. And Inflection isn't a pure tech bet — it has two revenue streams. The sensing side has deployed hardware on the International Space Station for NASA. Revenue topped $32 million last year with 23% growth guided this year. That's commercial traction for a company that's been public a couple of months.

Xanadu: The Photonic Hardware-Plus-Software Play

Xanadu brings the fourth architecture: photonic. Photons don't need the cryogenic cooling superconducting qubits demand, so the Aurora system runs at room temperature. That's a real scaling advantage.

But the part most investors will miss is the software. Xanadu built PennyLane, an open-source quantum programming framework used by half of all quantum developers worldwide — 160,000 monthly downloads, usage up 161% last year. That adoption holds regardless of which hardware architecture wins, because PennyLane works across all of them. With Xanadu you get the photonic hardware bet and the de facto software standard in one ticker.

Architecture-by-Architecture Cheat Sheet

Architecture	Public Stock	Edge	Stage
Superconducting (anneal)	D-Wave	Solves real optimization today	Revenue +179%, leading commercialization
Superconducting (gate)	Rigetti	General-purpose quantum	Two-contract concentration
Trapped Ion	IonQ	Highest precision, defense access	Revenue +202%, largest market cap
Neutral Atom	Inflection	Scales to thousands of qubits	$32M revenue + Nvidia backing
Photonic	Xanadu	Room-temp + software standard	Early, PennyLane held by 50% of devs

The Highest-Risk Name on the List

I have to call this out plainly. Quantum Computing Inc. (QUBT) had only $682,000 in 2025 revenue. The 10-K carries a going-concern warning. The company is facing securities fraud lawsuits — short sellers allege QUBT overstated its NASA relationship and that some of that revenue was never real.

Those are very serious allegations. With $682K in revenue, an active going-concern flag, and pending litigation, this is the riskiest name in the universe. Until the legal picture clarifies, I'd keep it on the watch list, not the buy list.

FAQ

Q: Why does owning multiple architectures matter? A: Quantum has no winning approach yet. Owning one architecture means betting the others fail — owning a basket means letting the technology pick.

Q: Which name has the most defensive moat? A: Xanadu, in my view. PennyLane's developer footprint is hardware-neutral, so it benefits whichever physical approach wins.

Q: Is QUBT investable at all? A: Not for me. Until the lawsuits resolve and revenue scales meaningfully, the risk-reward is asymmetric in the wrong direction.

What I'm Doing

I'm spreading exposure across architectures rather than picking a single winner. IonQ and D-Wave are core because their commercial traction is the clearest. Inflection and Xanadu are option value — small positions on potential breakouts. Rigetti I size carefully given the contract concentration. And QUBT stays out of the portfolio until the going-concern and litigation overhangs resolve.