Source Review: Air Liquide Baytown $850M, 9,000 t/d ASU (June 2024)

Verdict: Consistent Confidence: high

Primary-source corporate announcement. The 9,000 t/d O₂ capacity (= 3.29 Mt/yr if continuous) at $850M capex gives $258/(t/yr) capex coefficient — within ~30% of Thunder Said Energy's $200/(t/yr) rule and the SDC anchor's $137/(t/yr) (Linde Starbase). A single Baytown-class facility supports ~750 Block-3 Starship launches/year (at 4,422 t LOX/launch). For q1's "100 Mt/yr LEO requires ~1,350 ASUs and ~$1.1T capex" line, Baytown is the empirical capex anchor. The investment is conditional on regulatory timeline (ExxonMobil low-carbon hydrogen project alignment), so it's a near-term real-world data point, not just an industry-average estimate.

Abstract

Air Liquide's 24 June 2024 announcement of its Baytown, Texas, low-carbon oxygen production facility provides a concrete tier-C anchor for the capex of large-scale ASU capacity. The facility will be built at ExxonMobil's Gulf Coast site (Baytown TX), comprising four large modular air-separation units with combined capacity of 9,000 metric tons of oxygen per day — a 50% increase in Air Liquide's Texas capacity. Capital expenditure is up to $850 million USD. The investment is conditional on policy and regulatory milestones tied to ExxonMobil's adjacent low-carbon hydrogen project.

Key claims

• baytown-asu-capacity: "9,000 metric tons per day of oxygen production" (= 3.285 Mt/yr if continuous).
• baytown-asu-capex: "up to 850 million dollars" total capex.
• texas-capacity-step-up: Adds "50% increase in Air Liquide's Texas capacity" — implies prior Texas Air Liquide capacity was ~18 kt/d O2, total post-build ~27 kt/d.
• modular-asu-design: "four Large Modular Air separation units" — modular scaling pattern.

Reviewer notes

Tier C/D — corporate press release. Primary data on real-world large-ASU capex and capacity. The implied $/(t-O2/d) capex is $850M / 9,000 t/d = $94k per (t/d) of capacity, or about $258/(t/yr) — broadly consistent with Thunder Said Energy's $200/(t/yr) rule of thumb. One Baytown facility's 3.29 Mt/yr capacity is roughly enough to fuel ~750 Block-1 Starship launches per year (at ~4.4 kt LOX per launch). Cross-reference with Linde Starbase ASU (2,000 t/d, $100M capex) from the SDC anchor doc.