Type: thesis-shift (subtype: corrects — structural framing error)

event_id: 2026-05-10-ati-10k-hwm-competitor-not-customer-three-layer-stack-was-wrong

Source: ATI Inc FY2025 Form 10-K filed 2026-02-20 (SEC EDGAR accession 0001628280-26-010140, ati-20251228.htm)

Shift: Following up the same-day Q1 2026 ATI divergence finding by pulling the 10-K customer-concentration disclosure. Two findings, both load-bearing:

1. ATI customer-concentration disclosure (FY2025 10-K, Item 7 / financial-statement notes): "The Company markets its products to a diverse customer base, principally throughout the U.S. No single customer accounted for more than 10% of sales for any year presented." — No customer concentration. HWM is NOT a top ATI customer; ATI did not have 10%+ exposure to ANY customer in any year shown. The "alloy layer above HWM" framing implied at least some concentration; the disclosure rules that out.

2. ATI HPMC segment competitor list (FY2025 10-K, Item 1 business description): "Principal competitors in the HPMC segment include: Berkshire Hathaway Inc., for nickel-based alloys and superalloys and specialty steel alloys, titanium and titanium-based alloys, and precision forgings through its ownership of Precision Castparts Corporation and subsidiaries; Howmet Aerospace Inc., for titanium and titanium-based alloys; Carpenter Technology Corporation for legacy nickel-based alloys and superalloys and specialty steel alloys; and Aubert & Duval for precision forgings." — HWM is explicitly listed as an ATI competitor in titanium-based alloys, not a customer. CRS (Carpenter) is the nickel-superalloy competitor to ATI.

Effect on thesis (significant structural correction):

• The "3-layer stack ATI → HWM → GEV" framing from the 2026-05-06 entry was structurally incorrect. ATI is not a supplier to HWM in the dominant alloy class (titanium); they compete head-to-head in that market. HWM has internal melt capacity for titanium alloys and casts the blades — vertical-integration ATI does not have. The corrected picture: ATI, PCC (Berkshire), HWM, CRS, and Aubert & Duval are five overlapping participants in different cross-sections of the engine-alloy market. Specifically:
  • Nickel-superalloy feedstock (the IGT-blade-relevant class): ATI, PCC, CRS, Aubert & Duval all compete. HWM is a consumer of nickel superalloy from these vendors AND/OR has internal melt for some grades.
  • Titanium-based alloys (commercial-jet engine + airframe class): ATI, HWM, PCC compete head-to-head.
  • Casting and forging (downstream): HWM, PCC, MTU Aero Engines, Rolls-Royce, Doncasters compete.
• HWM's value-capture story is better under the corrected framing. If HWM has internal titanium melt + nickel-superalloy casting + finished-blade assembly, HWM captures more vertical margin than a pure caster sourcing alloy from ATI. The +39% Gas Turbines YoY isn't just demand pull-through to an alloy supplier — it's pull-through to a vertically-integrated player. HWM key_ticker conviction strengthens under this re-read.
• ATI key_ticker conviction softens materially. Two reasons compound: (a) ATI is NOT the alloy layer above HWM for IGT (Q1 2026 data); (b) ATI competes WITH HWM in titanium (10-K disclosure). The original "picks-and-shovels above HWM" framing is gone. ATI is a standalone specialty-alloy play with broad end-market exposure (aerospace + defense + energy + electronics + medical + automotive) — sensitive to commercial-airframe cycle, conventional-energy spend, and electronics demand. Q1 2026 showed those mixes net to flat. Not a sell signal but a re-grade from "key_ticker in stacked thesis" to "watchlist member with diversified specialty-alloy exposure." The 2026-05-06 deep-dive at 2026-05-06-ati-deep-dive was framed around the 3-layer stack; that framing needs revision.
• CRS coverage decision (TASKS.md:576) becomes the primary alloy-layer question. Per the ATI 10-K, Carpenter Technology is the named competitor for "legacy nickel-based alloys and superalloys" — which is the IGT-blade-relevant alloy class. If we want a pure-play alloy-layer name above HWM for the IGT thesis, CRS is the named candidate. The 2026-05-06 blind research-trace surfaced CRS at 2/3 convergence; today's ATI structural correction raises the priority of resolving the CRS coverage decision. User-decision item, ~10 min.
• PCC remains the closest comparable to HWM (vertically-integrated nickel-superalloy + titanium + casting + forging). PCC is private inside Berkshire so not investable directly; the existing PCC-vs-HWM divergence tracker (TASKS.md:432) becomes the proxy for "is HWM gaining share from the largest competitor."
• Methodology note: the structural correction came from a 10-K read that took ~5 minutes once asked — the original 2026-05-06 framing was based on the @Gaurab/@Ren_aramb capture which named "3 single-crystal casters" but didn't define the supply-chain structure ABOVE the casters. The followup question "is HWM actually an ATI customer or competitor?" should have been asked in the deep-dive sequence and wasn't. Filed as a meta-research observation: deep-dive sequences for vertically-organized theses should explicitly include a "is the vertical real or am I inferring it from the narrative?" check via competitor / customer disclosures in each layer's 10-K.

Per-ticker: ATI ↓↓ (structural framing was wrong — competes with HWM rather than supplies it; conviction reduced to "specialty-alloy diversified exposure" without the IGT-stack story; deep-dive needs revision). HWM ↑ (re-grading the vertical-integration angle strengthens value-capture story). CRS ↑ (becomes the primary alloy-layer question for the IGT thesis; coverage decision priority elevated). PCC reframed (closest HWM peer; PCC-vs-HWM divergence tracker becomes more useful as the share-shift proxy).