Quantum: the most empirically real AI-adjacent bubble

When we measure how tightly twelve editorial AI bubbles actually trade as one, the highest residualized intra-cluster correlation in our entire taxonomy belongs to Quantum. Specifically $QBTS, $RGTI, $IONQ, $ARQQ, and $QUBT — five quantum-computing pure-plays — with a 252-day residualized correlation of 0.76.

That's higher than Semi Equipment (0.82 raw, but still the cleanest semi bloc), higher than Memory / HBM, higher than Compute / GPUs. By the empirical test, Quantum is the most coherent thematic bloc in our coverage of the AI supercycle.

It's also one of the more counterintuitive results, because three of the five names were essentially unprofitable revenue-pre at the time of publication. Most "real bubble" intuition leans toward fundamentals-driven names; Quantum is mostly the opposite. The reason it clusters tightly is because of that, not in spite of it.

What "0.76 residualized correlation" actually means

If you're new to the residualization methodology, the 90-second version: every stock's daily return decomposes into "what the market did" + "what was specific to this stock." When you correlate raw returns, you mostly measure the first part — the shared market exposure that every U.S. stock has. When you correlate the residuals (what's left after subtracting market beta against $SPY), you measure only the idiosyncratic component — the part that says "these stocks are responding to something specific that the broad market isn't."

Full walk-through with worked examples in What is residualized correlation?.

For Quantum, that residualized correlation is 0.76. The raw is 0.78. Almost no drop under residualization. That's the signature of a real thematic bloc — these names move together for reasons that have nothing to do with general market gravity.

Why pre-revenue stocks correlate so tightly

The members:

• $QBTS — D-Wave Quantum, annealing-based quantum
• $RGTI — Rigetti Computing, superconducting qubits
• $IONQ — IonQ, trapped-ion qubits
• $ARQQ — Arqit Quantum, quantum-encryption infrastructure
• $QUBT — Quantum Computing Inc, photonic approaches

What they have in common: minimal current revenue, no near-term path to GAAP profitability, and an existential dependence on a single thesis — that someone, somewhere, builds a useful fault-tolerant quantum computer before the buildout patience of public markets runs out.

They have nothing else going on. There's no ancillary business line for any of them that would diverge from the central narrative. That's why their residualized returns cluster so tightly. The shared exposure is total: when sentiment about the path-to-quantum-utility shifts (in either direction), every name in the bubble responds in the same direction at roughly the same magnitude.

This is the inverse of the failure mode in our Hyperscalers bubble, where each name has so many ancillary revenue streams that the AI thesis is diluted into noise. Quantum has the opposite problem — pure thesis exposure with nothing to dilute it.

The macro thesis: post-GPU compute optionality

Why does the quantum narrative move so coherently? Because the underlying thesis is binary at the bubble level.

If a fault-tolerant qubit architecture commercializes by ~2030 — meaning useful compute on problems classical computers can't tackle (factoring, simulation, optimization) — the entire bloc gets revalued upward in tandem. None of the five companies are currently differentiated enough on technology that "the right one" gets all the upside. They all rip if the technology breaks through.

Conversely, if 2030 comes and goes without commercial fault-tolerance, public markets lose patience. Capital exits. The whole bloc deflates together.

This is the structural reason for the high residualized correlation: a binary outcome at the technology level forces a synchronized response at the asset level. Whoever wins the eventual race gets repriced first, but on the way up everything moves together because the market can't yet credibly distinguish which architecture leads.

Pure binary outcome → why pricing is sentiment-driven and synchronized.

Catalysts that move the bloc

The bubble responds to a tight set of inputs:

• Major hyperscaler/research benchmark announcements. When IBM, Google, or Microsoft posts a new quantum-volume record, the bubble moves on the read-through to the bloc's commercial timeline.
• Government quantum funding announcements. US, UK, EU, China — public-sector funding telegraphs the willingness to subsidize the buildout to commercialization.
• Single-company breakthroughs. Any of the five posting a benchmark surprise (or, asymmetrically, a setback) drags the rest with it ±15-25% in a single day.
• Capital-markets activity. Large secondaries, dilutive financings, or M&A change the bubble's effective float and shift sentiment about runway.

Watch the IBM Quantum and Google Quantum AI research disclosures most closely. They're independent of the public-equity bubble names but the read-through to commercialization timelines moves the bubble more reliably than anything any of the five publicly-traded names report themselves.

How to size the bet

The 0.76 residualized correlation has direct implications for position sizing:

Treat the bubble as a single position. If you own all five names equal-weight, you do not have five independent bets — you have one bet sized 5×. The bubble-level idiosyncratic risk dominates the within-bubble dispersion.

Right-tailed payoff distribution. If quantum commercializes, the bloc's expected return is high. If it doesn't, the bloc's expected return is materially negative. The distribution is fat-right-tailed, which means: own a small position in the basket, never concentrated. The right-tail outcome compensates for the modal disappointment outcome.

Etf shorthand. $QTUM is the only meaningful pure-play ETF for the bloc. It also includes broader quantum-adjacent names ($NVDA, IBM, etc.), so the residualized correlation of QTUM to the pure-play bubble subset is lower than the pure-play subset's internal correlation. If you want clean bubble exposure, an equal-weight basket of the five names tracks tighter than QTUM does.

Position-sizing math, briefly

If you target X% of your portfolio in "AI bubble" exposure, and you've decided Quantum is your single concentrated thematic bet within that, the right framing is: this is one position with binary return characteristics, sized small enough that a 70-90% drawdown in the basket doesn't blow up the portfolio. Most of the public quant guidance for binary-outcome assets converges on 1–3% of NAV at maximum, with the position rebalanced monthly to maintain target weight as the bubble marks higher.

This is conservative on purpose. Quantum is the strongest empirically-validated bubble in the taxonomy — but "strongest" by correlation tightness is not "highest expected return." It's "most coherent." Coherent blocs can still go to zero coherently.

---

The deeper point: Quantum is the bubble in our taxonomy with the cleanest read-through from "the technology bet works" to "every constituent rips together." Most editorial bubbles don't have that level of coherence. When you find one that does, the trade isn't to pick the winner — it's to size the bubble correctly and accept that the constituent risk is fungible.

Browse all 12 bubbles, ranked by empirical realness for context on where Quantum sits in the broader AI supercycle taxonomy.