Question: Does CRS's own 10-K (Form 10-K for fiscal year ended 2025-06-30, filed by Carpenter Technology Corporation, CIK 0000017843) name HWM/GE Aerospace/RTX-Pratt/Rolls-Royce/Safran as customers with explicit concentration, or does CRS exhibit the same OEM-mediated diversification that ATI's FY2025 10-K disclosed (zero 10%+ customers)?

Verdict: CRS exhibits the same no-10%+-customer-concentration disclosure shape as ATI — no single customer ≥10% of net sales for FY2025, FY2024, or FY2023. The "alloy layer to engine OEMs" framing in the 2026-05-11 CRS deep-dive is therefore OEM-mediated, not direct-customer-of-record, exactly the bear case the deep-dive flagged as open. However, a partial refinement: ~40% of CRS net sales sit under firm-price sales arrangements with annual purchasing commitments and consumption schedules — the structural-LTA evidence that is in the 10-K text and that ATI's 10-K did not equivalently disclose. The thesis weakens on customer-relationship-specificity but hardens on operational-LTA-share. Net: no thesis break; the 2026-05-11 CRS deep-dive should carry a 2026-05-26 correction block.

What we're asking

Per TASKS-RESEARCH.md:105, the 2026-05-11 CRS deep-dive rested on three converging-but-indirect signals:

1. ATI's FY2025 10-K named CRS as a "legacy nickel-based alloys and superalloys" competitor (not customer-of-record).
2. The 2026-05-06 blind research-trace surfaced CRS at 2/3 convergence as "nickel-superalloy powders for turbine parts."
3. ATI's own Q1 2026 print showed flat consolidated growth (+0.6%) while HWM's Gas Turbines segment did +39% YoY — by elimination, ATI was not capturing the IGT pull-through, so the alloy-layer-to-HWM relationship had to sit elsewhere.

The deep-dive's bear case explicitly flagged the missing direct test:

"CRS could be selling alloy primarily to engine OEMs (GE/PW/RR/Safran) who specify both alloy and caster separately — the same OEM-mediated structure ATI has. If so, CRS's 'alloy layer above HWM' framing has the same caveats (not direct supplier-to-HWM, but supplier-to-OEM-who-specifies-HWM-as-caster). Read CRS 10-K Item 1 to confirm. ~10 min."

This investigation runs that test against the actual filed document.

What we found

1 — Customer-concentration disclosure (the direct test)

From the CRS FY2025 10-K, Note 1 (Significant Accounting Policies) — Concentrations of Credit Risk:

"We are not significantly dependent upon a single customer, or very few customers, such that the loss of any one or more particular customers would have a materially adverse effect on our consolidated statement of operations. No single customer accounted for 10 percent or more of total net sales for the years ended June 30, 2025, June 30, 2024 and June 30, 2023. No single customer accounted for 10 percent or more of the accounts receivable outstanding at June 30, 2025 and 2024."

And restated in segment disclosure (Note 20):

"On a consolidated basis, no single customer accounted for 10 percent or more of net sales for the fiscal years ended June 30, 2025, 2024 and 2023. No single customer accounted for 10 percent or more of the accounts receivable outstanding..."

This is functionally identical to ATI's FY2025 10-K disclosure. Both companies sell diversified-customer specialty alloys; neither has a single customer ≥10%.

The deep-dive's bear case is confirmed as the actual structure: CRS is OEM-mediated, not direct-supplier-of-record to HWM specifically. Selling alloys to many parties — engine OEMs who specify both alloy and caster separately, casters who source alloy directly, distributors, etc.

2 — End-use market concentration (the indirect test)

From Item 7 MD&A "Sales by End-Use Markets" (FY2025 reported figures):

Two key reads:

(a) CRS is structurally A&D-heavy. 62% of FY25 revenue is Aerospace and Defense, up from 51% three years prior — A&D mix has been growing (51% → 56% → 62%). The market-level concentration is real even though customer-level concentration is not.

(b) A&D revenue grew +15% YoY in FY25 ($1,538.8M → $1,768.6M); excluding surcharge revenue +20%. This is operationally consistent with the HWM Gas Turbines +39% YoY Q1-2026 pull-through narrative — different magnitude, different cadence (HWM Q1 calendar vs CRS full-year), but same direction. CRS is capturing meaningful A&D-cycle lift through a diversified customer base.

3 — Long-term agreement disclosure (the partial-supplier-of-record evidence)

From Item 7 MD&A "Impact of Raw Material Prices and Product Mix":

"Approximately 40 percent of our net sales are sales to customers under firm price sales arrangements. Firm price sales arrangements involve a risk of profit margin fluctuations, particularly when raw material prices are volatile. ... Firm price sales arrangements generally include certain annual purchasing commitments and consumption schedules agreed to by the customers at selling prices based on raw material prices at the time the arrangement is entered into."

And:

"These firm price sales arrangements are expected to continue as we look to strengthen [supplier-customer] relationships by expanding, renewing and, in certain cases, extending to a longer term, our customer long-term arrangements."

40% of $2,877.1M = ~$1,151M of net sales under structural LTAs with annual volume commitments. This is the evidence the deep-dive's bull case needed: CRS does have structural multi-year supplier-of-record relationships, even though no individual customer is 10%+. The relationships are denominated by grade + volume commitment + firm price rather than concentrated %-of-revenue.

This disclosure is materially different from ATI's. ATI's FY2025 10-K notes diversification but does not equivalently quantify a 40%-of-sales LTA share with explicit volume commitments. The structural-supplier evidence sits more clearly at CRS than at ATI.

4 — Customer naming (the specificity test)

Grepping the full 10-K plaintext for HWM, Howmet, GE Aerospace, GE Aviation, RTX, Raytheon, Pratt & Whitney, Rolls-Royce, Safran, Airbus, Boeing:

• Boeing named once — but only in risk-factor language: "challenges affecting the commercial aviation industry or key participants including, but not limited to production and other challenges at The Boeing Company." Generic forward-looking-statement boilerplate, not a customer disclosure.
• HWM / Howmet / GE / RTX / Pratt / Safran / Rolls-Royce / Airbus — zero direct customer mentions anywhere in the filing.

CRS does not name specific aerospace customers in its 10-K. The customer base is treated as confidential and aggregated under end-use market labels.

5 — SAO segment financial validation

From the deep-dive's bull case, the prediction was that CRS's Specialty Alloys Operations segment should be the alloy-layer-leverage segment. The 10-K confirms:

SAO operating margin expanded from 16.7% (FY24) to 23.0% (FY25) — +630bp of segment-level operating margin in 12 months. The shape is consistent with mix-shift into premium nickel-superalloy and high-temperature alloys, which is exactly the IGT/jet-engine pull-through chain the thesis described. PEP is the downstream powder/parts segment and is structurally lower margin.

6 — Risk-factor language (the cyclicality caveat)

The 10-K Item 1A explicitly flags:

"The commercial aerospace and defense markets are historically cyclical due to both external and internal market factors. These factors include general economic conditions, airline profitability... The cyclicality of those markets can adversely affect our current business and our expansion objectives."

The thesis's bear case (specialty-metals cyclicality) is acknowledged by management. Not new information; verifies bear-case framing.

Verdict + reasoning

• The direct test confirms the bear case. CRS has zero 10%+ customer concentration, same shape as ATI's FY2025 10-K. The "alloy layer above HWM" framing should NOT be read as "CRS is HWM's direct supplier." CRS supplies many parties (engine OEMs, casters, distributors, end-user customers), and the relationship to HWM specifically is OEM-mediated.

• But the partial refinement holds. ~40% of CRS net sales are under firm-price LTAs with annual volume commitments — explicit, quantified, multi-year structural supplier relationships. This is materially more disclosure than ATI provides on equivalent terms. The supplier-of-record evidence does not exist at the individual-customer level, but exists at the grade × volume × duration level for ~$1.15B annualized.

• Operational evidence is consistent with the bull case. A&D revenue +15% YoY ($1,538.8M → $1,768.6M), SAO op margin +630bp YoY (16.7% → 23.0%), A&D mix-share growing (51% → 56% → 62% over three years). The operational shape matches a specialty-alloy producer riding an A&D-cycle upswing.

• The thesis does not break, it sharpens. The 2026-05-11 CRS deep-dive should be re-read as:

  • "CRS captures the IGT / commercial-jet / defense pull-through via diversified A&D mix" (correct, confirmed)
  • NOT "CRS is the alloy-supplier-of-record to HWM" (incorrect — OEM-mediated, same as ATI)
  • Adds: "~40% of revenue under firm-price LTAs with volume commitments" (new, supports structural-supplier framing without naming customers)

• Comparison to ATI is now sharper. Both ATI and CRS have identical no-10%+ disclosures. The operational divergence (ATI consolidated +0.6% vs CRS consolidated +4% with A&D +15%) reflects mix and execution, not customer-concentration structure. ATI is more titanium/aerospace-aluminum; CRS is more nickel-superalloy/specialty-alloy. The same A&D market wave benefits both, but mix and timing differ.

• One follow-up filed: add a 2026-05-26 correction block to the May 11 CRS deep dive reflecting the OEM-mediated structure (mirroring the 2026-05-10 ATI correction). Keep watchlist routing unchanged. No re-rate of conviction — the 10-K read confirms the thesis at the operational level even as it constrains the customer-relationship specificity.

Sources

• Primary: https://www.sec.gov/Archives/edgar/data/17843/000001784325000021/crs-20250630.htm — CRS FY2025 10-K (fiscal year ending 2025-06-30, filed via accession 0000017843-25-000021). Downloaded 2026-05-26 to /tmp/crs-10k-fy2025.htm (3.1MB) and converted to plaintext for grep.
• CIK 0000017843 — confirmed CARPENTER TECHNOLOGY CORP via EDGAR company search (Philadelphia, PA; SIC 3312 STEEL WORKS / BLAST FURNACES / ROLLING MILLS).
• Direct citations from the 10-K plaintext (paraphrased and verbatim-quoted above): Note 1 Concentrations of Credit Risk; Note 20 Segment Reporting; Item 7 MD&A Sales by End-Use Markets; Item 7 MD&A Impact of Raw Material Prices and Product Mix; Item 1A Risk Factors (cyclicality).
• 2026-05-11-crs-deep-dive — the deep-dive whose bear case explicitly flagged the missing direct test.
• 2026-05-10-ati-10k-hwm-competitor-not-customer-three-layer-stack-was-wrong — ATI 10-K correction; the structural-parallel anchor for this CRS read.
• Company profile — CRS current price ($434.12, mcap $21.6B), business description.
• CRS quarterly income statements — financials confirming Q3 FY26 +11.6% revenue, +35.3% op income YoY.
• TASKS-RESEARCH.md:105 — source row that scoped this investigation.

Carrier notes

• This is a primary-source read against a filing already on the public record (FY2025 10-K filed ~Aug 2025). The financials are as of 2025-06-30 (fiscal year end), the period the 10-K covers; read on 2026-05-26.
• The 10-K download is to /tmp/crs-10k-fy2025.htm — not committed to the repo (large binary HTML file with embedded inline-XBRL). The plaintext extract at /tmp/crs-10k-fy2025.txt (381k chars) is also temp-only. Anything that needs to persist should be re-extracted directly from EDGAR; the URL above is permanent.
• All percentages and dollar figures quoted from the 10-K are verbatim-pulled from the plaintext; no math was performed by Claude. Op-margin %s and YoY %s are reported directly by Carpenter in the filing.
• No customer name was redacted by Claude — the 10-K simply does not name HWM/GE/RTX/etc. The grep was exhaustive against the plaintext.