PsiQuantum's $100M CHIPS Award Just Made Photonic Quantum a US Defense Schedule — GlobalFoundries, Lumentum, Coherent Are the Real Trade

PsiQuantum signed a Letter of Intent for $100 million on May 21, 2026, alongside eight other quantum companies; the September 8 signing converted that LOI into a definitive award tied to milestones. Commerce's stated use of funds: scale domestic manufacturing of barium titanate (BTO) optical switches, high-temperature superconducting single-photon detectors (SNSPDs), and advanced packaging — three bottlenecks that PsiQuantum's own technical paper in Nature had identified as the gating items for utility-scale photonic quantum. Unlike the IBM Anderon award, which is a foundry grant to a separate subsidiary, PsiQuantum's check goes straight into a private-company production line that already ships to a real customer pipeline: a $125 million DARPA QBI contract (July 22, 2026), an A$940 million Australian/Queensland government commitment for Brisbane, and a $200 million2025 spend with American suppliers across 38 states.

The headline number undersells the shift. The May LOI raised PsiQuantum's total committed U.S. public + private funding stack to roughly $450 million across federal R&D, DARPA, and Brisbane; the September award is the first to be tied to a manufacturing schedule, not a research milestone. That distinction is what changes PsiQuantum's role in the listed supply chain from \"anchor customer\" to \"anchor customer with co-funded U.S. capacity,\" and it is what re-rates the photonic component ecosystem attached to it.

PsiQuantum has been locked in a 300mm silicon photonics partnership with GlobalFoundries since 2019, fabricating the photonic chipset that integrates SNSPDs, single-photon sources and the BTO-based switches at GlobalFoundries' Malta, New York fab. On September 8, GlobalFoundries closed its own definitive $375 million CHIPS award — five times PsiQuantum's — explicitly to scale a \"secure, domestic quantum foundry for leading architectures and multiple modalities.\" That puts PsiQuantum's photonic chips inside the first U.S. quantum fab explicitly funded to host multiple modalities, which protects single-customer concentration risk for GlobalFoundries and turns PsiQuantum into a captive anchor tenant.

Two implications. First, GlobalFoundries is the only U.S. foundry that can simultaneously serve PsiQuantum's photonic line and the broader AI/datacom silicon photonics market — the same wafer lines run quantum photonics,800G/1.6T transceivers, and co-packaged optics. Second, the $375M CHIPS award is now the U.S. government's direct investment in a foundry that has PsiQuantum as its flagship quantum tenant. That makes GlobalFoundries the cleanest listed exposure to \"photonic quantum becomes a U.S. factory schedule\" — without the customer-concentration risk of betting on PsiQuantum's IPO timing.

PsiQuantum's photonic stack depends on components the broader AI optical build-out already buys by the container load: high-power InP lasers for single-photon sources, ultra-low-loss silicon nitride waveguides, high-efficiency photodetectors, and optical fiber interconnect for the cryostat-to-control-electronics link. Nvidia's venture arm committed $4 billion in March 2026 across Lumentum and Coherent in non-exclusive supply agreements specifically to lock down this layer for Spectrum-X networking — the same lasers and the same packaging feed photonic quantum. Marvell added Polariton in April 2026 to bolster its electro-optics and custom silicon photonic portfolio; Applied Materials runs the deposition and etch tools that pattern the silicon nitride waveguides.

Photonic quantum needs cryogenics only at the detector stage — far less than superconducting, but still a non-trivial infrastructure investment. PsiQuantum has spent two years building that capacity directly: a September 2023 partnership with SLAC to install cryogenic quantum modules inside SLAC's LCLS-II cryoplant, and a March 2026 UK R&D facility at STFC Daresbury's National Cryogenic Facility for \"next-generation high-power cryogenic systems.\" The cold-supply market itself is privately held (Bluefors in Finland, Maybell Quantum in Colorado, Oxford Instruments NanoScience, recently divested by Oxford Instruments and absorbed by Quantum Design International) and Quantum Machines (Israel) and Zurich Instruments (Switzerland) hold the bulk of the quantum-class control electronics.

That gap is the article's most important caveat. Investors looking for a clean public-market line on the cryogenic and control-electronics layer behind the CHIPS quantum awards will not find a U.S.-listed name; the listed exposure sits one step upstream, in photonic components and silicon photonic wafer fabrication. If the U.S. government decides to onshore this layer next — which the IBM-HRL acquisition (closed August 26, 2026) and Honeywell's 58% stake in Quantinuum implicitly argue for — that is the trade that gets created next, not the trade that exists today.

TSMC's CHIPS lane is built around $6.6 billion in direct CHIPS funding plus a $265 billion cumulative Arizona investment program, of which $100 billion was incremental and announced in July 2026. That lane is fab capacity — millions of wafers a year of advanced logic, packaging, and HBM-adjacent silicon. PsiQuantum's $100 million does not buy wafer volume; it buys photonic quantum process development, three component categories, and an early-stage U.S. supply base for a modality that does not yet ship in volume. The two lanes share the legal chassis (the CHIPS and Science Act) and the supply-chain security framing, but they target different revenue lines: TSMC's $265B serves AI compute, PsiQuantum's $100M serves a future quantum utility-scale market that the Quantum Insider sizes at roughly $1 trillion in economic value by 2035, but which generates essentially zero hardware revenue today.

PsiQuantum is not alone in the package — but it is the only photonic winner. Quantinuum ($100M, trapped-ion, the only trapped-ion recipient; Honeywell majority-owned, ~48% of voting power post-IPO) uses a wholly different supply chain dominated by ion-trap fabricators and UV lasers. D-Wave ($100M, annealing + 10,000-qubit gate-model) and Rigetti ($100M, superconducting) both need dilution refrigerators and superconducting wafer fab — the same cryogenic gap noted above. Atom Computing ($100M, neutral atom) and Infleqtion ($100M, neutral atom) need vacuum and laser systems. Diraq ($38M) targets silicon spin qubits, a fifth modality. And then there is IBM's $1B for the Anderon foundry: not a modality bet but a foundry bet on superconducting quantum, paired with the August 26 close of the HRL Laboratories acquisition to fold in-house cryogenic and materials expertise.

The packaging tells a story. Of the $2.013B announced in May and finalized in stages through September,68% goes to foundry capacity (IBM + GlobalFoundries) and 32% to discrete modality bottlenecks — and within the modality bucket, photonic quantum wins the cleanest check: a single architecture with an existing commercial foundry relationship, a DARPA QBI Stage C validation, and a multi-billion-dollar private capital base. The defensible read: Commerce is funding quantum the way it funded semiconductors — anchor foundries first, specialty fabs second, and modalities that lean on existing commercial tooling get a faster check. Photonic is exactly that modality.

Days to quarters: the immediate catalyst is PsiQuantum's September 8 award driving incremental domestic pull-through at GlobalFoundries' Malta fab — SiPho revenue ($277M Q2 2026, +62% YoY) is the line that re-rates. Lumentum and Coherent continue to ship into AI datacom first; quantum is upside optionality, not yet a driver. Marvell's optical interconnect book is the same: AI-driven with photonic quantum as a secondary tailwind. The first-quarter-and-a-half trade is GlobalFoundries; the second-quarter trade is the photonic component layer as quantum specification awards trickle out through2027.

One to three years: the structural bet is that photonic quantum achieves utility-scale — DARPA's QBI program targets 2033 for \"industrially-useful\" quantum. If PsiQuantum's Brisbane site, Chicago IQMP, and the Malta SiPho line ramp as planned, photonic becomes the modality with the only domestic foundry (GlobalFoundries) hosting a real multi-modality quantum fab. The listed winners are the photonic component ecosystem that doubles for AI. The risks are commercial (fault-tolerance milestones slip), geopolitical (export controls fragment the photonics supply chain), and competitive (a third modality — neutral atom, ion trap, or a Chinese photonic entrant — gets a faster federal check).