Howmet Aerospace: How fast should capacity commitments move?
Download case PDF (6 pages)Request teacher’s pack
Strategy case for MBA students and working professionals | Hypothetical decision date: October 2, 2026, 4:00 p.m. ET
Growth is arriving in different forms
Howmet Aerospace entered autumn 2026 with strong demand signals in commercial aerospace and gas turbines. In the first half, reported commercial-aerospace revenue was $2.574 billion, up from $2.076 billion a year earlier; gas-turbine revenue was $605 million, compared with $437 million. These figures describe sales, not bookings, line-level utilization, or the incremental capacity required to serve future demand. [P2, Note D, six-month end-market table, PDF pp. 13–14]
Management's March Technology & Markets Day presentation showed gas-turbine revenue of about $1 billion in 2025 and forecast more than $2 billion over the following three to five years, based on internal assumptions. It also forecast large commercial-aircraft deliveries rising from 1,611 in 2025 to 2,149 in 2030, citing Jefferies/Cirium. Neither forecast was a Howmet order book or a project-level capacity plan. [P4, slide 11, PDF p. 11]
Then, on August 6, CEO John Plant said capital-expenditure requirements were increasing and would need to rise further in 2027 to support future organic-growth expectations in aerospace and gas turbines. He also said gas-turbine customers were revisiting and adding to their demand outlooks. Howmet did not disclose a 2027 capex amount, a list of proposed projects, or a public decision date. [P3, CEO remarks]
For this exercise, assume a hypothetical October 2026 capital review is deciding how to pace discretionary 2027 capacity commitments. The choice is how much to commit broadly in advance and how much to stage behind customer, technical, and operating evidence. No actual review, budget, project list, or project economics is implied.
What the public record says about capacity and cash
Howmet spent $453 million of cash capital expenditure in 2025. The company said spending principally supported Engine Products capacity expansion, automation, and sustaining and return-seeking projects across its segments. Its 2025 Form 10-K expected 2026 capex of about 5% of sales, including aerospace and gas-turbine capacity expansions. These are historical spending and a 2026 outlook, not a 2027 budget. [P1, Item 7, Investing Activities and capex outlook, PDF pp. 31–32]
First-half 2026 cash capex was $198 million, versus $221 million in the prior-year period. Cash from operations was $1.036 billion and free cash flow was $838 million. The same half-year cash-flow statement reported $1.929 billion of acquisitions, net of cash acquired. CAM closed on April 6 for about $1.8 billion; Brunner was acquired in February. Acquisition spending is not capex. Segment capex is also reported on an accrual basis, which can differ from cash capex. [P2, Statements of Cash Flows, PDF pp. 5–6; capex reconciliation, PDF p. 13; Note C, acquisition detail]
The August earnings release reported $800 million of common-stock repurchases year to date through July. In March, before CAM closed, management had described 2026 priorities including $450–490 million of capex, more buybacks than in 2025, disciplined M&A, and leverage near 1.0x. The March priorities and August update have different dates and scope. They show other uses of management attention and capital, but do not establish a binding 2027 trade-off or financial constraint. [P3, CEO remarks and summary; P4, slide 60, PDF p. 60]
Customer funding offers a possible way to share some growth risk. The 2025 Form 10-K says Howmet sometimes receives customer payments—primarily in gas turbines and commercial aerospace—for future deliveries and to fund future volume growth. Deferred revenue was $253 million at December 31, 2025; the associated delivery periods are generally one to five years. The filing does not identify a particular project, customer, refund term, or amount available to fund a proposed 2027 investment. These payments remain obligations to deliver product, not free cash or proof of project-level demand. [P1, Item 7, Contractual Obligations and Deferred Revenue, PDF pp. 32–33]
Customer concentration can make schedule changes consequential. GE Aerospace and RTX represented approximately 14% and 10% of first-half 2026 third-party sales, respectively, primarily in Engine Products. The figures do not disclose customer-specific order growth or commitments for a proposed project. [P2, Note D, PDF p. 14]
A decision under competing risks
The cost of waiting is not abstract. Howmet's 2025 Form 10-K warns that a rapid, significant demand increase could exceed its ability to ramp production quickly, risking lost growth opportunities or competitive position. But the disclosure does not identify a current bottleneck or the lead time for any contemplated project. Strong end-market sales and forecasts may indicate durable opportunity; they may also fail to map directly to Howmet's own orders or incremental capacity need. [P1, Item 1, risk factors]
There are also operating alternatives to immediate fixed investment. Howmet's March strategy deck lists driving labor productivity and automation in manufacturing as strategic initiatives (P4, slide 59). In the first half of 2026, Engine Products absorbed approximately 720 net headcount in support of expected revenue increases (P2, MD&A, Segment Information—Engine Products, PDF p. 29, HTML lines 1144–1145). These disclosures do not quantify project-level slack, utilization, qualification readiness, or output that could be recovered without new equipment.
Investment research offers a framework, not a company answer. Pindyck's review of irreversible investment explains how waiting can preserve the option to learn when investment is sunk, while long construction times and competitive costs of delay can support earlier action. Bontempi, Golinelli and Parigi find in a panel of Italian manufacturers that uncertainty and investment planning interact, with results dependent on capital reversibility and operating flexibility. Their estimates are not Howmet coefficients, project hurdle rates, or capacity forecasts. [A1, PDF pp. 1–4, 19–30, 35–39; A2, Tables 1–2 and pp. 16–22]
The pacing choice
Three policies are available for discussion. They are case alternatives, not announced Howmet policies.
| Policy | Operating approach | Potential advantage | Exposure |
|---|---|---|---|
| Commit broadly in advance | Release most discretionary long-lead capacity commitments across aerospace and gas turbines using current forecasts, customer outlooks, and expected growth. | Can secure lead-time capacity and reduce the risk of arriving late to customer demand. | Commits capital before project-level schedules, utilization, execution readiness, and returns are public or established. |
| Stage commitments behind evidence gates | Continue sustaining, productivity, or low-regret work; release additional discretionary capacity in tranches as customer schedules, qualification, bottlenecks, and ramp readiness clear defined gates. | Preserves some ability to learn and redirect later commitments. | Gates can delay capacity, and the relevant evidence may arrive after the time needed to ramp. |
| Require stronger customer backing | Prioritize projects with clearer customer schedules, advances, reservation terms, or other documented support before releasing major commitments. | Connects investment release to customer willingness to secure future supply. | Public disclosures do not reveal project-level terms; advances create delivery obligations and can still leave utilization or execution risk. |
These approaches can be combined. For example, management could begin long-lead work broadly while reserving the largest irreversible expenditures for projects that pass later gates. Any thresholds would be proposed policy choices; Howmet has not disclosed them.
Your task is to recommend a commitment posture and a compact set of evidence triggers for accelerating, continuing, staging, or reviewing the next tranche. Specify what would count as customer-backed demand, operating readiness, and capacity need; identify when the evidence should be reviewed and who should own each trigger. Explain what would change your position. Do not treat reported sales as bookings or assume that customer advances eliminate delivery risk.
Exhibit 1. Demand evidence and forecast horizon
| Evidence | Reported or projected | What it does—and does not—show |
|---|---|---|
| H1 2026 commercial-aerospace revenue | $2,574m, versus $2,076m in H1 2025 | Reported end-market sales, not order growth or a capacity gap. |
| H1 2026 gas-turbine revenue | $605m, versus $437m in H1 2025 | Reported end-market sales; segment results do not establish line-level utilization. |
| Gas-turbine revenue outlook | About $1bn in 2025; $2bn+ over three to five years | March management forecast based on internal assumptions, not contracted Howmet sales. |
| Large commercial-aircraft deliveries | 1,611 in 2025; 2,149 in 2030 | March forecast citing Jefferies/Cirium, not a Howmet order schedule. |
| Management's 2027 capex statement | Requirements expected to rise further | August CEO outlook; no amount, projects, or return data disclosed. |
Sources: P2, Note D, PDF pp. 13–14; P3, CEO remarks; P4, slide 11. Forecasts are labeled as forecasts; reported sales are not bookings.
Exhibit 2. Capital deployment context
| Measure | Amount / statement | Period and interpretation |
|---|---|---|
| Cash capex | $453m | FY2025; principally Engine Products expansion, automation, sustaining and return-seeking projects. |
| Cash capex | $198m vs. $221m | H1 2026 vs. H1 2025. Cash-flow measure. |
| Cash from operations / free cash flow | $1,036m / $838m | H1 2026. Consolidated cash generation. |
| Acquisitions, net of cash acquired | $1,929m | H1 2026 cash flow; includes CAM and Brunner acquisition activity, not capex. |
| Common-stock repurchases | $800m | Year to date through July 2026, per August release. |
| Segment capex | $440m | FY2025 accrual basis; not the same measure as $453m cash capex. |
Sources: P1, Item 7, Investing Activities and segment capex table, PDF pp. 31, 53–54; P2, Statements of Cash Flows and capex reconciliation, PDF pp. 5–6, 13; P3, summary and CEO remarks. Periods and accounting bases differ.
Exhibit 3. Customer funding and execution risk
| Evidence | Public disclosure | Guardrail |
|---|---|---|
| Customer payments for future growth | Sometimes received, primarily in gas turbines and commercial aerospace; $253m deferred revenue at Dec. 31, 2025; delivery generally over one to five years | Not all deferred revenue is project funding; terms and project allocations are undisclosed; payments carry delivery obligations. |
| Customer concentration | GE Aerospace ~14%; RTX ~10% of H1 2026 third-party sales, primarily Engine Products | Sales concentration is not a project-specific order or commitment. |
| Rapid-ramp risk | 10-K says a significant rapid demand increase could exceed ramp capability | Disclosed risk, not evidence that a present bottleneck exists or how long a specific project takes. |
Sources: P1, Item 7, Contractual Obligations and Deferred Revenue, PDF pp. 32–33; P1, Item 1, risk factors; P2, Note D, PDF p. 14.
Decision limits
The public record does not identify proposed sites or equipment, project costs, spend phasing, lead times, installed capacity, utilization, bottlenecks, qualification timelines, customer cancellation or advance terms, project margins, expected returns, labor plans, or a 2027 capex ceiling. It also does not establish what information would arrive before a project must be committed. Students should state assumptions rather than invent these inputs or calculate project-level returns from company-wide figures.
Howmet announced on October 1 that it would release third-quarter 2026 results on October 29, after this case's information cutoff. No results after October 2, 2026, 4:00 p.m. ET are part of the decision record. [P5]
Sources
Company sources and academic reading limits are documented in the accompanying research packet: site/research/howmet-aerospace-inc/source-register.md. Source IDs in this case refer to that register. The case is hypothetical and contains no actual management recommendation or later outcome.