Ultra-high vacuum systems

Every trapped-ion and neutral-atom quantum computer is, before it is anything else, a vacuum vessel. Ion lifetimes and atom-array fidelities collapse when a stray background molecule hits a qubit, which is why these machines specify ultra-high to extreme-high vacuum — 10^-11 torr and below — and why recent trapped-ion engineering work treats the vacuum system as a first-order qubit-lifetime budget item, not plumbing.

Who owns the choke

Atlas Copco quietly assembled most of it. Edwards (acquired 2014), Leybold (2016), ion-pump specialist Gamma Vacuum and cryostat maker Montana Instruments all sit inside its Vacuum Technique business. Agilent inherited Varian's VacIon franchise — Sandia's QSCOUT trapped-ion testbed runs on an Agilent ion pump and UHV gauge. VAT owns the valve layer, Busch-controlled Pfeiffer the turbopump layer, SAES Getters the NEG-pump layer, Kurt J. Lesker the custom-chamber layer. The moat is not patents; it is fifty years of tacit knowledge in outgassing rates, bakeout recipes, getter chemistry and all-metal sealing, qualified across CERN, LIGO and the national labs — references no entrant can shortcut — wrapped in installed-base service economics. The market is also too small to invade rationally: an oligopoly preserved by apathy.

What would break it

A frontal competitor needs a decade of materials science and a reference base; nobody is trying. The live threats are subtler. Vertical integration: Infleqtion manufactures its own UHV glass cells, keeping the choke in-house. But the bigger make-vs-buy datapoint went the other way: Universal Quantum, which needs million-ion machines, signed an MoU with Atlas Copco's Edwards, Leybold, Gamma and Montana Instruments brands in December 2025 to industrialize quantum vacuum rather than build it. Second, cryopumping: at 4 kelvin every cold surface pumps for free, so cryogenic trap architectures could cut active pump content per machine. Third, modality risk: superconducting and photonic machines do not need 10^-11 torr; if they win, this choke matters less.

Be honest about the size

Quantum UHV is, generously, tens of millions of dollars a year inside a vacuum industry priced off semicap cycles. Vacuum Technique is roughly a quarter of Atlas Copco's revenue, and quantum is a sliver of that sliver. Nobody prices ATLKY, MKSI or VACNY off quantum today — correctly. The chokepoint is real while the market is small; the trade is owning the toll road before the traffic shows up.

What forces a repricing

DARPA's Quantum Benchmarking Initiative advanced 11 companies into Stage B in November 2025 — IonQ, Quantinuum, QuEra and Atom Computing among them, all UHV-dependent modalities. Stage C decisions, expected within roughly a year of Stage B's start, fund utility-scale prototype builds: that converts single-chamber lab procurement into industrial vacuum orders with multi-year service tails. Behind it, the National Quantum Initiative Reauthorization Act cleared House committee in April 2026 carrying a quantum supply-chain resilience mandate — rare named federal attention for this layer of the stack. Near term, Atlas Copco's July 16 Q2 print is the third since the Universal Quantum MoU — the April Q1 report already flagged vacuum-driven order growth — and any quantum vacuum commentary from the one conglomerate that owns three UHV pump franchises is the tell.