HWM — Howmet Aerospace Inc.

The market's vote, not the sell-side's words, is bearish as of 2026-09-13: the stock fell 11.44% in one week and 18.47% in one month to $229.61, 26% below the 52-week high of $310, and it now sits 12.08% below its 20-day SMA and 15.68% below its 50-day SMA with RSI at 30.19. The proximate trigger was an early-September 2026 SpaceX announcement that it would bring gas-turbine blade and vane casting in-house, which introduced a customer-insourcing narrative to a stock priced for perfection; Bernstein and Citi downplayed it as non-systemic. That conflicts sharply with the fundamental news flow - a Q2 beat on 2026-08-06, a raised guide, $300M of Q2 buybacks, and a 17% dividend increase - and with sell-side positioning (consensus rating Buy, 21 buys vs. 3 holds and 1 sell, price targets of $329 consensus / $337 recent average). This is a valuation and narrative de-rating against an intact operating story, not a demand collapse.

The trend is unambiguous acceleration, not a blip. Quarterly revenue went $1.658B (Q3 2023) to $1.731B (Q4 2023) to $1.891B (Q4 2024) to $2.168B (Q4 2025) to $2.547B (Q2 2026). Year-over-year growth ran ~9-11% through 2024, then +13.8% (Q3 2025), +14.6% (Q4 2025), +19.1% (Q1 2026) and +24.1% (Q2 2026) - four consecutive quarters of accelerating growth, with the August 2026 release citing +21% organic. Margins tell the same story: operating margin went from 18.6% in Q3 2023 to 22.9% in Q3 2024 to 25.4% in Q2 2025 to 27.9% in Q2 2026, and gross margin from 24.4% to 37.3% over the same span. One caveat worth flagging: Q1 2026's 32.6% operating margin (operating income of $753M on $2.313B revenue) looks flattered by a non-recurring item - gross profit of $854M implies only ~$101M of operating expense versus ~$240M in Q2 - so the 27.9% Q2 GAAP operating margin is the better run-rate, and management's own adjusted EBITDA of $817M (+39% YoY) in Q2 confirms underlying profitability actually improved even as the GAAP margin dipped sequentially. Cash generation is real: quarterly free cash flow of $479M in Q2 2026, $359M in Q1 2026 and $530M in Q4 2025, with full-year 2026 FCF projected near $1.9B. Earnings quality is clean - the accruals ratio is -0.37% as of 2026-06-30, meaning cash flow exceeds reported net income.

Expensive, and modestly rich against its own history as of 2026-09-13. At $229.61 the trailing P/E is 49.2x versus a five-year average of 42.2x, P/S is 10.1x versus a five-year average of 5.0x, P/B is 16.0x, and EV/EBITDA is 35.0x. PEG is 1.375. Forward P/E on the next-FY consensus EPS of $5.335 is 43.0x. The one genuine mitigant is that the selloff has already compressed the multiple - at the 52-week high of $310 the trailing P/E was ~66x (310 / $4.678 TTM EPS), so roughly a third of the premium has been worked off. Everything hinges on durability: a 43x forward multiple is only defensible if the 20%+ top-line growth and margin expansion persist, and the consensus FY2027 EPS figure of $5.335 implies just ~1% growth over management's FY2026 guidance of ~$5.27, which does not support that multiple on its face. FCF yield is roughly 2.1% on 2026 projected FCF of ~$1.9B - fine for a compounder, not a cushion.

There is a real gap between the two earnings anchors, and how you resolve it determines the trade. The backward-looking anchor says the market is paying ~43x forward for what consensus models as ~flat 2027 EPS growth over 2026 - a poor trade if you believe those estimates. The momentum anchor says those 2027 estimates are stale: Q2 2026 alone printed $1.33 diluted EPS with revenue still accelerating into a Q3 target of ~$2.75B, and the CAM fastener deal closed 2026-04-06 with synergies landing into 2027, so $5.335 looks too low. In other words, the price is discounting roughly $5.3 of 2027 earnings while the business is currently running at a pace closer to $5.8-6.0. The September 2026 SpaceX headline is what forced that discount to widen. Practically: the stock is not pricing in growth that hasn't shown up - it is pricing a growth stall that has not shown up either. The Q3 print in late October 2026 is the arbiter; if the $2.75B revenue run-rate lands with margins intact, the stale-consensus argument wins and the multiple re-rates; if Q3 misses, 43x on ~1% growth becomes the correct read and there is more downside.

Unambiguous short-term downtrend, now approaching oversold. As of 2026-09-13 the RSI(14) was 30.19 - at the threshold, not yet below it - and the price sat 12.08% below the 20-day SMA (~$261), 15.68% below the 50-day SMA (~$272) and 6.72% below the 200-day SMA (~$246). The stack is fully bearish: price < 200-day < 20-day < 50-day, meaning the short-term average has already crossed below the 50-day, and the loss of the 200-day (which had been the trend support) is the most important technical damage. Trailing returns confirm the momentum break: -11.44% one week, -18.47% one month, -13.25% three months, -3.02% six months. Timing implication: RSI near 30 and a 26% drawdown from the high make a reflex bounce plausible, so chasing shorts here is poor risk/reward; but buying a falling knife into a broken 200-day is also poor practice. The disciplined entry is a close back above ~$246 (the 200-day) with RSI reclaiming the high-30s, or a stabilization base above $227-230. Losing the prior close area of $227.91 opens the 52-week low at $182.78.

• Commercial aerospace OEM build-rate ramp - Boeing and Airbus backlogs remain the primary demand driver for Howmet's engine and fastener content, and any OEM rate cut would hit the highest-margin lines first.
• AI data-center power demand driving industrial gas turbine (IGT) blade and vane orders - a genuine incremental tailwind for Howmet's leading share position, but also the exact segment where SpaceX's September 2026 insourcing announcement landed.
• Interest-rate easing and active debt management - Howmet cut annualized interest expense by ~$12M via a cross-currency swap and paid down a JPY term loan, but long-term debt still rose to $4.05B at 2026-06-30 from $2.86B at 2025-12-31 to fund the CAM acquisition, making absolute rates matter more now than a year ago.
• Defense spending and geopolitical tension supporting the defense portion of the aerospace mix, plus tariff and raw-material (nickel superalloy, titanium, aluminum) cost pressure across cross-border supply chains.
• Sector rotation and multiple compression in premium industrials - after a multi-year re-rating, high-P/S, high-EV/EBITDA names like HWM are the first to be sold when rates or risk appetite shift, which is precisely what the 26% drawdown from the February-to-September 2026 highs reflects.

• Q3 2026 earnings in late October 2026 (tentatively October 28-29) - the single most important test of whether management's ~$2.75B quarterly revenue target and ~28% operating margin hold up, and whether the stale FY2027 consensus EPS of $5.335 gets revised upward.
• CAM fastener acquisition integration - closed 2026-04-06 for ~$1.8B, with material synergies guided into 2027; successful delivery would validate the leverage taken on and lift forward estimates.
• 2027 capital-return update and the path to management's 1.0x net debt-to-EBITDA target by year-end 2026 - deleveraging enables a step-up in buybacks beyond the $300M deployed in Q2 2026 and further dividend increases after the 17% Q3 hike to $0.14.
• Resolution (or escalation) of the SpaceX vertical-integration question - if other OEMs or power-turbine customers follow, it becomes a structural share-loss narrative; if it stays a one-off, it becomes a buying opportunity.
• Continued Boeing/Airbus rate decisions and IGT order momentum tied to AI data-center power buildouts, both of which flow directly into the Engine Products segment that drove Q2's +32% revenue surge.

• Valuation risk is the dominant risk. At 49.2x trailing, 43.0x forward and 35.0x EV/EBITDA versus a 5-year average trailing P/E of 42.2x, there is very little room for error, and the 26% drop from $310 to $229.61 in roughly seven months is a live demonstration of how fast that multiple can compress.
• Customer insourcing / vertical integration - the early-September 2026 SpaceX announcement that it will cast natural-gas turbine blades and vanes in-house is the specific overhang behind the recent selloff, and it raises the question of whether GE, Safran, Boeing or Airbus expand in-house content over time.
• Cyclical aerospace exposure - a Boeing or Airbus production hiccup, an engine program delay, or softening air travel would hit OE and high-margin aftermarket volumes simultaneously, and the 24% Q2 2026 growth rate will be an exceptionally tough comparable entering 2027.
• Leverage and integration execution - long-term debt rose to $4.05B at 2026-06-30 from $2.86B at 2025-12-31, cash fell to $563M from $2.435B at 2026-03-31, and the quick ratio is a thin 0.872; CAM synergies are guided, not yet delivered.
• Segment concentration in Engine Products, which powered the Q2 beat with +32% revenue growth - a single program (LEAP, GTF) or single-customer disruption would be disproportionately material to consolidated margins.

Howmet is a genuinely excellent operating business going through a painful de-rating. The fundamentals are accelerating, not deteriorating: four straight quarters of accelerating revenue growth culminating in +24% YoY in Q2 2026, operating margins up from 18.6% to 27.9% over three years, 18.2% ROIC, 28.2% ROE, negative accruals (clean earnings), and ~$1.9B of projected 2026 free cash flow. The problem is price. At $229.61 as of 2026-09-13 the stock still trades at 49x trailing and 43x forward earnings, 10x sales versus a 5-year average of 5.0x, against a consensus FY2027 EPS that implies roughly no growth. The 18.5% one-month drawdown and RSI of 30.19 have taken real froth out and make this a materially better entry than a month ago, but I would want the late-October Q3 print and a reclaim of the 200-day (~$246) before sizing up. Own it for the quality of the franchise; do not own it because it is cheap, because it is not.

Forward P/E on a self-built FY2027 estimate: I modeled FY2027 revenue at ~$11.1B (roughly 10% growth off the ~$10.05B FY2026 guide), a ~21.5% net margin (up from 20.5% TTM on mix, CAM synergies and lower interest after the $12M annualized swap benefit), and ~396M diluted shares after continued buybacks, producing FY2027E EPS of ~$6.00 - above the arguably stale consensus $5.335. I applied a 40x forward multiple, just below the stock's 5-year average trailing P/E of 42.2x, to reflect decelerating growth (~24% now toward low-double-digits) offset by best-in-class margins and FCF; 40 × $6.00 = $240. Cross-check: ~26x FY2027E EBITDA of ~$3.7B less ~$3.2B net debt on ~396M shares yields ~$235. The gap versus the street's $329 consensus is deliberate - $329 implies roughly 55x my 2027 EPS, which the data on forward growth rates does not support.