The three-week-old deferred item from the June quantum scan is now closed: Quantinuum priced its IPO at $60 per share, sold 28 million shares on Nasdaq under the ticker QNT, and raised $1.68 billion in gross proceeds [https://www.quantinuum.com/press-releases/quantinuum-announces-closing-of-upsized-initial-public-offering], converting the pure-play quantum tier from a private-valuation inference exercise into a measurable comparable set. That matters now because it lands alongside IonQ's already-public $260–270 million FY26 revenue guide, creating for the first time a two-name hardware-leading tier where relative valuation is legible — and therefore testable.
That legibility cuts both ways. IonQ (IONQ) trades at roughly 37× forward revenue on a guide backed by Q1 2026 actuals of $64.7 million (+755% YoY) and a $3.1 billion cash pile. QNT's revenue record is substantially thinner; the implied post-IPO market capitalisation, inferring from the "upsized" characterisation and the pre-IPO $8–15 billion anchor, likely sits in the $10–20 billion range — but that inference rests on a fully-diluted share count I have not yet verified (S-1/424B5 from SEC EDGAR is the next-scan queue item, alongside Honeywell's residual stake percentage). Until then, Honeywell (HON) remains the only liquid public proxy for that residual economic interest. If QNT's first post-IPO earnings print discloses trailing revenue materially below IonQ's guide, the gap in implied revenue multiple becomes the operative test of whether the tier consolidates around a shared valuation logic or fragments.
The second development is commercial rather than financial. The Quantinuum × HPE strategic collaboration on quantum-HPC integration for enterprise, announced 21 June [https://www.quantinuum.com/news], opens a distinct sub-tier above the hyperscaler cloud-API channel. HPE is not a quantum-native cloud, but it is a systems-integration channel embedded into pharma, energy, and industrial R&D customers — the same base Quantinuum already serves directly through BMW, BP, Synopsys, and Mitsubishi Electric. The commercial logic is that HPC-hybrid workloads (classical HPC plus quantum accelerator) monetise earlier than pure-quantum-API calls because the classical HPC side carries existing customer budget; only the quantum acceleration is incremental spend. It is the same near-term commercial bridge pattern IonQ's Clavis XG Multiplex QKD product represents at the security layer — distinct revenue path, distinct sales cycle, both operating well before architectural-race resolution.
D-Wave (QBTS) provides the most analytically interesting data point of the 19-day window. The IDC MarketScape Quantum Computing 2026 Vendor Assessment named D-Wave one of two market leaders on 7 July [https://www.dwavequantum.com/company/newsroom/] — five weeks after the dual-rail gate-model roadmap publication and before D-Wave has delivered its first 17-physical-qubit gate-model milestone (roadmap date: 2026, followed by 49 and 181 physical qubits in 2027 and 2028, targeting 10 and 100 logical qubits in 2030 and 2032 respectively). The speed of that designation is notable: analyst categorisation rarely converges this fast on an architecture that has not yet delivered its inaugural milestone. My read is that IDC is partly anchoring on the annealing installed base plus roadmap ambition — a position the analyst is taking ahead of the milestone rather than in response to it. The practical consequence is real regardless: "Leader" designation is a required-and-checked procurement filter at Fortune 500 buyers and federal accounts, which gives D-Wave near-term commercial cover on the gate-model pivot before the hardware is in hand. The $1.5 million NSF grant of 30 June [https://www.dwavequantum.com/company/newsroom/] is small in dollars but coherent with the federal supply-side backstop running at cadence alongside the dual CHIPS letter of intent.
On the regulatory front, the NIST PQC PIV working drafts published 12 June [https://csrc.nist.gov/news] are the first concrete sub-milestone I have caught this year on the migration calendar. PQC-capable Personal Identity Verification credentials advancing from concept to working draft pulls the federal PIV rekey cycle — tens of millions of credentials over a 3–7-year re-issuance window — into observable procurement lead-times. This is structurally constructive for the PQC-migration downstream-beneficiary names: Cloudflare (NET), Thales, and Entrust sit on a mandated federal-credential procurement cycle that is now visibly advancing. The NIST IR 8547 2035 RSA/ECC deprecation horizon is unchanged; the Gidney May 2025 arithmetic (under one million noisy qubits, under one week for RSA-2048 — arXiv:2505.15917) remains the load-bearing cryptographic-relevance anchor at 8–15 years out.
One update trims a prior inference. The Anthropic Fable 5 directive, imposed and lifted within 18 days (resolved 30 June, per the adjacent ai-capex-ceiling/2026-07-09 scan), resolves as a transactional voluntary safeguard exchange rather than a Section 1758 ECRA categorical action. The June quantum scan implied the Anthropic precedent extended export-control political economy durably into quantum hardware. The correct forward read is that any analogous quantum-hardware directive would likely resolve on similar transactional-negotiation terms, provided the vendor can assemble a comparable four-commitment safeguard package. The Trusted-Foundry rent thesis — rooted in DoD certification cycles of 18–36 months and mature-node supply-side scarcity — stands on its own fundamentals; the export-control amplifier is simply thinner than I thought.
The picks-and-shovels hierarchy remains unchanged in direction but is reinforced in conviction. Cryogenics — Oxford Instruments (OXIG.L) and Coherent (COHR) — and Trusted-Foundry mature-node control electronics sit above the pure-plays in durability because their economics are agnostic to which of the six live architectures eventually leads. GlobalFoundries (GFS) is the name that sharpens most on multi-theme grounds: the chips-compute/2026-07-06 scan promoted the compound-semi feedstock chokepoint (Ga/Ge/Sb at the coupled MOFCOM/BIS regulatory intersection), and GFS sits at the intersection of that axis and the quantum Trusted-Foundry layer simultaneously — the overweight rationale is no longer a single-thesis bet. SkyWater's close into IonQ, announced at roughly $1.8 billion and pending Q2–Q3 2026, will be the most informative near-term data point on IonQ's FY27 cost-base intention and on how Trusted-Foundry capacity consolidates.
What would change my mind
Three things would materially break this view. IonQ printing Q3 2026 revenue below $50 million sequential would signal the $260–270 million FY26 guide is unrealistic and re-rate the pure-play multiple across the tier. QNT disclosing trailing revenue below $150 million on its first post-IPO earnings print while trading at a premium multiple would compress the tier and raise a serious question about whether the revenue base can support the valuation structure. And a US-government export-control action against a named quantum-hardware vendor that does not resolve on transactional-negotiation terms within 30 days would establish the durable domestic-export-regime category the Fable 5 resolution ruled out, repricing the geographic revenue mix across all named names.
What I'm watching next
S-1/424B5 on SEC EDGAR for QNT's fully-diluted share count and Honeywell's residual stake percentage. IonQ Q2 2026 revenue print (expected early-to-mid August) — the first sequential test of the FY26 guide. SkyWater close confirmation and the associated balance-sheet impact on IonQ. D-Wave's 17-physical-qubit gate-model milestone delivery against the 2026 roadmap date — whether it validates or exposes the IDC designation. NIST PQC PIV working-draft progress toward finalisation and CISA's PIV inventory-mandate compliance reporting schedule. IBM Quantum blog (currently returning 403; Starling, Kookaburra, and Cockatoo processor cadence remain inferred against the 2029 roadmap until access resolves).