Howmet at Jefferies Global Industrials Conference 2026: demand outruns capacity

Howmet Aerospace reaffirmed its path to exceed $10 billion of revenue in 2026, after reaching $9.1 billion in trailing-12-month sales, up 18% year over year. Management indicated it will likely raise its prior outlook for industrial gas-turbine sales to double to about $2 billion within three to five years from 2025, driven by AI data-center power demand, aerospace utilization and long-term turbine OEM contracts. The company is operating at capacity and is adding roughly 2.5 plants' worth of incremental capability across the U.S., Europe and Japan, while reporting more than 30% year-to-date yield improvement. Leverage remains low at 1.2x despite approximately $2 billion of acquisitions, alongside continued organic investment, buybacks and dividend increases.

Analysis

HWM’s incremental earnings power is likely more sensitive to throughput and mix than to end-market order growth: each point of yield, uptime, or qualification-driven price realization converts directly into constrained-capacity revenue. The key non-obvious upside is that advanced IGT content can shift toward higher-value single-crystal and cooling architectures, raising dollar content per turbine even if turbine unit growth normalizes. This makes HWM a more direct beneficiary of AI-power buildout than broad data-center hardware names, with a 6-18 month earnings revision catalyst once capacity ramps are reflected in formal guidance.

The near-term risk is that the market capitalizes the demand narrative before the physical expansion produces shipments. New-furnace commissioning, skilled-labor ramp, scrap rates, and customer qualification schedules can defer revenue while depreciation and start-up costs arrive immediately; fixed-price contract terms may also limit spot-like pricing upside. GE Vernova’s vertical-integration push via CPP is not an immediate substitution threat, but it increases the probability that future captive capacity targets lower-complexity casting work, leaving HWM with a narrower but more technically demanding mix.

Consensus may underappreciate the duration embedded in the installed base rather than the next turbine order cycle. Higher utilization creates a compounding aftermarket annuity, but the stock’s premium multiple will be vulnerable if AI-related power projects are delayed by grid interconnection, gas availability, or financing conditions. For the next 1-3 months, February guidance is the clean catalyst; falsification is a guidance increase that fails to exceed current elevated buy-side expectations, evidence of ramp-related margin pressure, or a material deterioration in OEM delivery schedules.

Key Decisions for Investors

• Accumulate HWM on broad industrial/rates-driven weakness ahead of the February earnings update; target a 9-12 month holding period. Underwrite only if consensus 2027 EPS still leaves room for a mid-single-digit or better uplift from IGT capacity and mix; exit or reduce on a guide-up below that threshold or a material increase in launch-cost commentary.
• Express relative exposure through long HWM / short XLI over 6-12 months rather than an outright industrial-beta position. HWM should retain superior pricing and aftermarket conversion in a slowing macro backdrop, while XLI remains exposed to broader capex and freight cyclicality; reassess if oil and rates drive a generalized aerospace/industrial multiple de-rating.
• Maintain a watch on long HWM versus GEV only after the latter discloses CPP integration plans and supply-allocation policy. A confirmed captive-capacity strategy focused on products presently sourced externally would impair the HWM thesis; absent that evidence, both companies can benefit from the turbine bottleneck and the pair lacks clean isolation.
• Do not chase NVDA or AI-infrastructure proxies on this read-through. The more actionable confirmation signal is turbine OEM backlog, lead-time, and margin commentary from GEV, SIEGY, BKR, and CAT over the next two reporting cycles; weakening power-equipment orders would precede a revision risk for HWM.