AMDAdvanced Micro Devices, Inc.
Industry AI semis Updated 24 Sep 2026
The idea
The expensive one: ~41x FY27 EPS even if Data Center doubles in a year. The price is the option on the second-source thesis, and the OpenAI/Meta warrants (320m shares at $0.01) dilute holders if it works.
Events to watch
- Q3 results early Nov
- Helios and MI500 ramp
- warrant vesting in each 10-Q
- the '>$20 EPS' target timeline
How we got the target
The dossier's base FY27 non-GAAP EPS ~$13.8 × 37.5x, the midpoint of its own bear (30x) and bull (45x) multiples. Bear and bull are the dossier's figures. Time frame: FY2027 (Dec 2027).
Target: worked out from expected earnings. Prices: market data get_latest_stock_price, close of 2026-09-22 (fetched 2026-09-23)
What would prove it wrong
6 warning signsData Center revenue fails to more than double in 2027.
Details
Management guided exactly that on the Q2 FY26 call. It is the whole thesis in one number.
Warrant tranches vest without the purchase volume that was supposed to trigger them, or vest on stock-price tranches rather than purchase milestones.
Details
That is dilution without revenue.
Non-GAAP gross margin falls below ~54% for two consecutive quarters.
Details
Guided ~56% for Q3 FY26; the Instinct mix and rising memory cost both push the other way.
A second MI-series export charge.
Details
MI308 cost ~$800m gross in Q2 2025 with ~$360m reversed. A repeat on MI450/MI500 would show the China exposure was never actually removed.
Company non-GAAP gross margin falls in the first two Helios quarters (Q4 2026, Q1 2027) against the ~56% guided for Q3 2026.
Details
— AMD does not disclose Instinct margin and would not give it on two calls, so this is the only public read on whether the ramp is accretive.
Client and Gaming deteriorates faster than Data Center compensates.
Details
Gaming was −31% y/y in Q2 FY26 and management is "planning for a softer PC market" on memory cost.
How far along
6 of 10 steps done · 4 done in a group review
Next step
track warrant vesting. AMD has issued warrants over 320 million shares at $0.01 to OpenAI (2025-10) and Meta (2026-02). None had vested at 2026-06-27. Each 10-Q's warrant note is the single most valuable recurring read on this name, because vesting is where the customer relationships turn into ~20% dilution.
Needs your decision
S6 product test / expert call, S7 conviction.
Failed checks
none. S4/S6/S7 are not closed by design (cluster stages); S6 is not closed in the cluster file itself. ⚠️ Thread to pull (dodge log): Instinct gross margin was dodged on the Q1 and Q2 2026 calls (C.J. Muse both times) as Helios starts to ramp in Q3 2026. Capital allocation is neutral: flat GAAP EPS since FY21.
Who runs it
The Front OfficeStep by step
- S0How it does as AI gets cheapDone in a group review
carry-forward (cluster: AI-capex complex — cluster pass, 2026-09-20)
the cluster's.
⚠️ SPLIT verdict; AMD sits on the seller side. Owner checks dropped 2026-09-21.
- S1Where the idea came fromDone
watchlist (owner, Apple Stocks, 2026-09-20); no Tier 0 promotion.
- S2Quick deal-breaker checkDone
FY25 rev $34.6bn +34%, GM 49.5%, op margin 10.7%, OCF $7.7bn, FCF $6.7bn, net cash. Survives.
⚠️ the exposure is dilution, not solvency: 320m warrant shares at $0.01 plus 107m plan shares against 1,659m diluted — up to ~26%. · capital allocation:
⚠️ neutral — GAAP diluted EPS $2.57 (FY21) → $2.65 (FY25) on revenue 2.1x; incremental ROIC ~0% GAAP / ~4% ex amortisation on ~$51bn of new capital (mostly the Xilinx stock issue, count +33%). Buybacks cheap (avg ~$99–144 vs $614.61) but small ($8.6bn vs $5.9bn SBC); no dividend; no impairment. The 320m-share warrants are the largest pending allocation item.
- S3Reading the company's reportsDone
(gate-minimum, partial)
FY25 10-K risk factors + MD&A, Q2 FY26 10-Q, DEF 14A ownership, Aug-2026 424B5. Q1 and Q2 FY26 calls read Q&A-first (2026-09-24, closes the one-transcript gap). Comp structure still not read.
Full notes
EQ: 2 flags — inventory days ~130 → ~159 → ~165 (FY23–FY25, ~152 H1 FY26). SBC 123% / 58% of FCF in FY23 / FY24 (24% in FY25). Non-GAAP add-backs falling (19.6% → 11.8% of revenue). MI308 charge kept inside non-GAAP.
- S4Rivals and suppliersDone in a group review
carry-forward (cluster: AI-capex complex — cluster pass, 2026-09-20)
- S5Who owns the stockDone
proxy 2026-03-19, 13F Q2-26, FINRA short interest, buyback + shelf. No ATM captured (Equibles, ~13 months; absence is not proof). New S-3ASR 2026-08-13 drew $4.75bn of notes within a day — see S5 note.
- S6What workers and customers sayDone in a group review
carry-forward (cluster: AI-capex complex — cluster pass, 2026-09-20)
not closed there too.
- S7The case for it, and pricesDone in a group review
carry-forward (cluster: AI-capex complex — cluster pass, 2026-09-20)
conviction open OPEN (user) · valuation (bear/base/bull; moved from S8 2026-09-23): ~77x model FY26 non-GAAP EPS, ~41x model FY27. The most expensive of the three names in this slot on every measure.
- S8Do the bosses keep their word?Done
23 rows over 8 quarters: 6 beat, 10 met, 7 missed
reliable. Revenue 8/8 in or above the ±$300m range, non-GAAP GM 8/8 on guide; non-GAAP opex over Jean Hu's guide 7/7 by 3–8% (an analyst called it out 2026-05-05), tracking revenue beats.
Full notes
Nothing withdrawn; current team (Su, Hu) made every row. 4 dodges. Instinct gross margin dodged two quarters running.
- S9What to watch nextDone
trigger written. Tier 0 EDGAR sweep covers the feed (cik 2488)
Sources and notes
Sources 7 documents
- Full notes, 24 Sep 2026
- Full notes, 21 Sep 2026
- AI-capex complex — cluster pass, 2026-09-20
- EDGAR filings CIK 2488
| Short name | Document | Where it came from | Read on |
|---|---|---|---|
| 10K-FY25 | AMD 10-K for FY ended 2025-12-27, filed 2026-02-04 | https://www.sec.gov/Archives/edgar/data/2488/000000248826000018/amd-20251227.htm | 2026-09-21 |
| 10Q-Q2FY26 | AMD 10-Q for the quarter ended 2026-06-27, filed 2026-08-05 | https://www.sec.gov/Archives/edgar/data/2488/000000248826000123/amd-20260627.htm | 2026-09-21 |
| DEF14A-26 | AMD DEF 14A filed 2026-03-27 (ownership table based on shares outstanding at 2026-03-19) | https://www.sec.gov/Archives/edgar/data/2488/000119312526129057/d85856ddef14a.htm | 2026-09-21 |
| 424B5-AUG26 | AMD 424(b)(5) prospectus supplement, $4,750,000,000 of senior notes, filed 2026-08-14 | https://www.sec.gov/Archives/edgar/data/2488/000119312526352628/d173126d424b5.htm | 2026-09-21 |
| CALL-Q2FY26 | AMD Q2 FY2026 earnings call, 2026-08-04 (machine transcript, speaker-labelled) | Equibles GetEarningsCallTranscript AMD, turns 1-4 of 56 | 2026-09-21 |
| 13F-Q2-26 | Institutional holders, 13F report date 2026-06-30 | Equibles GetTopHolders AMD | 2026-09-21 |
| 8K-EXEC-26 | AMD 8-K executive and director changes, filed 2026-05-15 / 2026-08-19 | Equibles GetExecutiveChanges AMD (each row links its own 8-K on sec.gov) | 2026-09-21 |
Links to other companies 25 links
| From | Link | To | Date | Where it says so |
|---|---|---|---|---|
| Lisa T. Su | officer of | AMD | 2026-08-04 | CALL-Q2FY26 — speaker turn 3, 'Lisa Su (CEO)'; introduced in turn 2 as 'Dr. Lisa Su, our Chair and CEO' |
| Jean Hu | officer of | AMD | 2026-08-04 | CALL-Q2FY26 — speaker turn 4, 'Jean Hu (CFO)' |
| Lisa T. Su | director of | AMD | 2026-05-13 | 8K-EXEC-26 — 8-K filed 2026-05-15, Item 5.07 voting results |
| Tim Ryan | director of | AMD | 2026-08-19 | 8K-EXEC-26 — 8-K filed 2026-08-19, Item 5.02 |
| Joe Householder | governance structure | AMD | 2026-08-19 | 8K-EXEC-26 — 8-K filed 2026-08-19, Item 5.02 |
| AMD | issuer of | Warrant to OpenAI OpCo, LLC over up to 160,000,000 AMD shares at $0.01, issued October 2025 | 2026-06-27 | 10Q-Q2FY26 — Notes, 'Warrants' subsection following the Repurchase Program paragraph |
| AMD | issuer of | Warrant to Meta Platforms, Inc. over up to 160,000,000 AMD shares at $0.01, issued February 2026 | 2026-06-27 | 10Q-Q2FY26 — Notes, 'Warrants' subsection, same paragraph |
| OpenAI OpCo, LLC | warrant holder | Warrant to OpenAI OpCo, LLC over up to 160,000,000 AMD shares at $0.01, issued October 2025 | 2025-12-27 | 10K-FY25 — Item 7 MD&A, Overview, paragraph beginning 'In October 2025, we entered into a product purchase agreement' |
| Meta Platforms, Inc. | warrant holder | Warrant to Meta Platforms, Inc. over up to 160,000,000 AMD shares at $0.01, issued February 2026 | 2026-06-27 | 10Q-Q2FY26 — Notes, 'Warrants' subsection |
| AMD | partner and competitor | OpenAI OpCo, LLC | 2026-09-21 | 10K-FY25 — Item 7 MD&A, Overview, OpenAI product purchase agreement paragraph; cross-read against NVDA and AVGO dossiers of 2026-09-21 |
| Anthropic | customer of | AMD | 2026-08-04 | CALL-Q2FY26 — Lisa Su prepared remarks, paragraph beginning 'In addition to our multi-generation gigawatt scale deployments' |
| Microsoft Corporation | customer of | AMD | 2026-08-04 | CALL-Q2FY26 — Lisa Su prepared remarks, sentence following the Anthropic paragraph |
| Microsoft Corporation | supplier dependency | AMD | 2025-12-27 | 10K-FY25 — Item 1A, Risk Factors Summary, Operational and Technology Risks bullet |
| Top customer at 11% of FY25 accounts receivable (unnamed in the 10-K) | channel concentration | AMD | 2025-12-27 | 10K-FY25 — Note 11 — Concentrations of Credit Risk |
| AMD | acquired | ZT Systems | 2025-03-31 | 10K-FY25 — Item 7 MD&A, Overview, paragraph beginning 'In March 2025, we completed the acquisition of ZT Systems' |
| Sanmina Corporation | acquired | ZT Systems | 2025-10-31 | 10K-FY25 — Item 7 MD&A, Overview, same paragraph, second sentence |
| Bureau of Industry and Security, U.S. Department of Commerce | regulator of | AMD | 2025-12-27 | 10K-FY25 — Item 1A, risk factor 'Government actions and regulations such as export regulations, import tariffs, and trade protection measures may limit our ability to export our products to certain customers.' |
| BlackRock, Inc. | holds | AMD | 2026-06-30 | 13F-Q2-26 — rank 1 of 4,213 rows |
| Vanguard (Vanguard Capital Management LLC / Vanguard Portfolio Management LLC) | holds | AMD | 2026-06-30 | 13F-Q2-26 — ranks 2 and 7 of 4,213 rows |
| State Street Corp | holds | AMD | 2026-06-30 | 13F-Q2-26 — rank 3 of 4,213 rows |
| Jane Street Group, LLC | context | AMD | 2026-06-30 | 13F-Q2-26 — ranks 9 and 12 of 4,213 rows, position type Put and Call |
| Susquehanna International Group, LLP | context | AMD | 2026-06-30 | 13F-Q2-26 — rank 11 of 4,213 rows, position type Put |
| AMD | governance structure | Lisa T. Su | 2026-03-19 | DEF14A-26 — Beneficial ownership table, 'All current directors and executive officers as a group (16 persons)' row and footnote (2) |
| AMD | issuer of | AMD senior notes, four tranches due 2029-2036, $4.75bn, August 2026 | 2026-08-14 | 424B5-AUG26 — cover page, 'Filed pursuant to Rule 424(b)(5) File No. 333-298288' |
| AMD | event | AMD senior notes, four tranches due 2029-2036, $4.75bn, August 2026 | 2026-08-14 | 424B5-AUG26 — cover page, aggregate principal amount line, '$4,750,000,000' |
History 3 entries
- 2026-09-21 — first pass, slot
names-ai-capex-1, model-drafted unattended by the cloud research routine. S1, S2, S3 (gate-minimum), S5 and S9 closed; S0/S4/S6/S7 carried forward from the 2026-09-20 cluster file; S8 modelled and left open. Evidence and access limits in dossier-2026-09-21.md. Nothing here is human-verified. - 2026-09-23 — stage redefinition, owner decision (WATCHLIST_PLAN.md §25): S8 is now Management credibility and is not yet run for this name; the old S8 valuation figures moved to the
valuation:line under S7, and any proposed buy price or position size was removed (the pipeline no longer drafts either). No research was redone. S2 capital allocation and S3 earnings quality are also pending, in the same backfill slot. - 2026-09-24 — backfill, slot
backfill-2026-09-23-1(manually started session). Wrote S2 capital allocation (⚠️ neutral), S3 earnings quality (2 flags: inventory days, historic SBC/FCF) and S8 management credibility (✅ reliable, 23 rows) in dossier-2026-09-24.md; added the Dodge log and a Helios-margin kill criterion (the thread from two dodged quarters). S3 now has two transcripts. No valuation haircut: S8 is not over-promises. Other stages unchanged. Nothing here is human-verified.
Full notes, 24 Sep 2026 9 min read
AMD — dossier, 2026-09-24 (backfill)
Slot backfill-2026-09-23-1, backfill pass, manually started research session following .claude/skills/stock-sourcing-pipeline/SKILL.md §0 (backfill slots). Nothing here is human-verified. Only the three checks added on 2026-09-23 are written here — S2 capital allocation, S3 earnings quality and S8 management credibility (WATCHLIST_PLAN.md §25). Every other stage stands as written in dossier-2026-09-21.md.
Transcript quotes were located by exact substring in the archived market data files under ledger/research/_transcripts/AMD/, with the speaker label read back at that position. market data labels turns by name only; roles come from relationships.yaml (Lisa Su, Chair and CEO; Jean Hu, EVP, CFO and Treasurer, both as of 2026-08-04). Call date is the locator.
Price reference: $614.61 close 2026-09-23 (market data get_latest_stock_price NASDAQ:AMD).
Sources
| Key | Document | Locator |
|---|---|---|
PR-Q2-24 … PR-Q2-26 | Nine quarterly results releases, 8-K Item 2.02 Ex. 99.1, 2024-07-30 → 2026-08-04 | https://www.sec.gov/Archives/edgar/data/2488/<acc>/q<q><yyyy>991.htm — e.g. 000000248826000121/q22026991.htm (2026-08-04), 000000248826000072/q12026991.htm (2026-05-05), 000000248826000014/q42025991.htm (2026-02-03), 000000248825000009/q42024991.htm (2025-02-04), 000000248824000161/q32024991.htm (2024-10-29) |
10K-FY25 | 10-K for FY ended 2025-12-27, filed 2026-02-04 | https://www.sec.gov/Archives/edgar/data/2488/000000248826000018/amd-20251227.htm |
10Q-Q2-26 | 10-Q for the quarter ended 2026-06-27 | https://www.sec.gov/Archives/edgar/data/2488/000000248826000123/amd-20260627.htm |
XBRL | EDGAR companyfacts, CIK 2488 | https://data.sec.gov/api/xbrl/companyfacts/CIK0000002488.json |
CALL-Q3-24 … CALL-Q2-26 | Eight earnings calls, 2024-10-29 → 2026-08-04 | _transcripts/AMD/FY2024-Q3 … FY2026-Q2 (sha256 in _transcripts/index.json) |
Equibles: 0 calls for this name. market data: one price call.
S2 — capital allocation and returns (FY2021–FY2025, $m)
| FY21 | FY22 | FY23 | FY24 | FY25 | 5-yr | |
|---|---|---|---|---|---|---|
| Operating cash flow | 3,521 | 3,565 | 1,667 | 3,041 | 7,709 | 19,503 |
| Capex | (301) | (450) | (546) | (636) | (974) | (2,907) |
| Free cash flow | 3,220 | 3,115 | 1,121 | 2,405 | 6,735 | 16,596 |
| Acquisitions, net of cash | 0 | Xilinx all-stock; net line not read | (131) | (548) | (1,760) | — |
| Buybacks | (1,762) | (3,702) | (985) | (862) | (1,316) | (8,627) |
| Tax-withholding repurchases | — | — | (427) | (728) | (607) | — |
| Dividends | none | none | none | none | none | 0 |
| Stock-based compensation | 379 | 1,081 | 1,384 | 1,407 | 1,638 | 5,889 |
| Diluted shares (m) | 1,229 | 1,571 | 1,625 | 1,637 | 1,636 | +33% |
Source: XBRL (NetCashProvidedByUsedInOperatingActivities, PaymentsToAcquirePropertyPlantAndEquipment, PaymentsForRepurchaseOfCommonStock, ShareBasedCompensation, WeightedAverageNumberOfDilutedSharesOutstanding); FY23–FY25 acquisitions and tax-withholding lines from the 10K-FY25 cash-flow statement. FY25 also received $1,356m from selling the ZT manufacturing business: 10K-FY25 — "In March 2025, we completed the acquisition of ZT Systems for $3.2 billion in cash and 8.3 million shares of our common stock", then "sold the ZT data center infrastructure manufacturing business (ZT Manufacturing Business) to Sanmina Corporation (Sanmina) for $2.4 billion in cash". H1 FY26 (XBRL, 10-Q): OCF 5,321, capex 1,197, buybacks 221, SBC 990.
- The window is dominated by one decision paid in stock. Xilinx (FY22) added $23,888m of goodwill (
XBRLGoodwillAcquiredDuringPeriod) and took the diluted count from 1,229m to 1,571m. Since FY23 the count has been flat — buybacks plus tax-withholding repurchases have offset SBC, and done no more. - Buybacks at what price. Cash over shares repurchased (
XBRLStockRepurchasedDuringPeriodShares): FY21 ~$104, FY22 ~$103, FY23 ~$99, FY24 ~$144, FY25 ~$110 — all a fraction of today's $614.61. Well-priced, but small: $8.6bn over five years against $5.9bn of SBC. - Returns. Invested capital = equity + debt − cash − short-term investments. FY21: 7,497 + 313 − 3,608 = 4,202; NOPAT 3,648 × (1 − 14%) ≈ 3,137 → ROIC ~75%. FY25: 62,999 + 3,250 − 10,552 = 55,697; NOPAT 3,694 × 0.87 ≈ 3,214 → ~5.8% on GAAP, ~9% adding back the $2,254m of acquisition-intangible amortisation. Incremental return FY21→FY25: ~0% on GAAP (ΔNOPAT ~77 on ΔIC ~51,500), ~4% with amortisation added back.
- Per share: GAAP diluted EPS $2.57 (FY21) → $2.65 (FY25) on revenue that went $16.4bn → $34.6bn; non-GAAP EPS $3.31 (FY24) → $4.17 (FY25) (
PR-Q4-25). - Goodwill impairments: none tagged in the window.
- A fourth use of capital, off the cash-flow statement: warrants over 320m shares at $0.01 to OpenAI and Meta (prior dossier) — equity given for purchase commitments. Not in this table because no cash moved; it is the largest potential capital-allocation item on the name.
Verdict: ⚠️ neutral. Not a destroyer — no impairment, and the per-share non-GAAP line is growing — but five years of capital, most of it the Xilinx stock issue, have produced flat GAAP earnings per share and an incremental return near zero on the GAAP measure. The buybacks were cheap and too small to matter. Whether the 2025–26 AI ramp finally earns the Xilinx equity is the open question, and the warrants would add up to ~20% more shares against it.
S3 — earnings quality
- Non-GAAP bridge. Operating-income add-backs: FY23 4,453 (19.6% of revenue), FY24 4,238 (16.4%), FY25 4,074 (11.8%) — SBC, acquisition-intangible amortisation (2,811 → 2,394 → 2,254), acquisition- related costs, restructuring (FY24 only) (
PR-Q4-24,PR-Q4-25reconciliations). Falling in dollars and as a share of revenue. Two points in its favour: the ~$800m MI308 export charge in Q2 2025 was kept inside non-GAAP gross margin (43% guided and reported), and investment gains ($365m in FY25) are removed from non-GAAP. Minor: "Acquisition-related and other costs" recur in 3 of 3 years (262, 186, 228) — small, but a recurring "one-off". - SBC. FY25 1,638 = 4.7% of revenue, 24% of FCF. ❗ FY23 123% and FY24 58% of FCF — FCF after SBC was negative in FY23 and 41% of reported FCF in FY24. Cleared in FY25 and H1 FY26 (24%).
- Cash conversion. OCF / NI: FY23 1.95, FY24 1.85, FY25 1.78 (amortisation is non-cash); H1 FY26 1.45. Accruals (NI − OCF) / average assets FY25: −4.6%. Clean.
- Working capital. DSO (year-end): FY24 88 days, FY25 67; quarterly Q2 FY26 57 — falling. ❗ Inventory days: FY23 ~130 → FY24 ~159 → FY25 ~165 (inventory 4,351 → 5,734 → 7,920 against cost of sales; FY23/FY24 cost of sales from reported GAAP gross margin, FY25 from
XBRL17,487) while revenue grew — two rising periods, the rule's flag. Eased to ~152 on H1 FY26 annualised (inventory 8,468). - Costs moved onto the balance sheet: not checked at this depth.
EQ: 2 flags — inventory days 130 → 159 → 165 (FY23–FY25, easing in 2026); SBC above half of FCF in FY23–FY24 (cleared in FY25). Non-GAAP bridge shrinking and conservative on the export charge.
S8 — management credibility
AMD guides one quarter: revenue ±$300m and non-GAAP gross margin in the release; non-GAAP opex on the call (Jean Hu, turn 4 or 5 of each call). Eight quarters scored, Q3 2024 → Q2 2026.
| # | Quarter guided | Revenue guide → actual | Score | Non-GAAP GM guide → actual | Score | Non-GAAP opex guide (Hu, call) → actual | Score |
|---|---|---|---|---|---|---|---|
| 1 | Q3 2024 (guided 2024-07-30) | $6.7bn → $6.819bn | met | ~53.5% → 54% | met | — (Q2 2024 call not archived) | — |
| 2 | Q4 2024 (2024-10-29) | $7.5bn → $7.658bn | met | ~54% → 54% | met | ~$2.05bn → $2.125bn | missed (+3.7%) |
| 3 | Q1 2025 (2025-02-04) | $7.1bn → $7.438bn | beat | ~54% → 54% | met | ~$2.1bn → $2.213bn | missed (+5.4%) |
| 4 | Q2 2025 (2025-05-06) | $7.4bn → $7.685bn | met | ~43% incl. ~$800m MI308 charge → 43% | met | ~$2.3bn → $2.429bn | missed (+5.6%) |
| 5 | Q3 2025 (2025-08-05) | $8.7bn → $9.246bn | beat | ~54% → 54% | met | ~$2.55bn → $2.754bn | missed (+8.0%) |
| 6 | Q4 2025 (2025-11-04) | $9.6bn → $10.270bn | beat (incl. ~$390m MI308 China not in the guide) | ~54.5% → 57% (~55% ex a $360m reserve release and China) | beat | ~$2.8bn → $3.001bn | missed (+7.2%) |
| 7 | Q1 2026 (2026-02-03) | $9.8bn → $10.253bn | beat | ~55% → 55% | met | ~$3.05bn → $3.145bn | missed (+3.1%) |
| 8 | Q2 2026 (2026-05-05) | $11.2bn → $11.536bn | beat | ~56% → 56% | met | ~$3.3bn → $3.394bn | missed (+2.8%) |
Locators: each guide from the prior quarter's release ("Current Outlook") and call (Hu, "Non-GAAP operating expenses to be approximately $…"); each result from the next release's headline and its "Non-GAAP operating expenses" reconciliation line; quarterly revenue from XBRL (Q4s derived from the 10-K year less three quarters). Who: the release outlooks are company (written); the opex guide is Jean Hu's, read off her speaker label in CALL-Q3-24 through CALL-Q1-26. Revenue "met" = inside the ±$300m range; opex "missed" = spent above the guide.
The opex pattern is the finding, and analysts see it. Seven for seven above the guide, by 3–8%. Stacy Rasgon put it to management on CALL-Q1-26 (2026-05-05): "it sort of continues to blow past the targets". Hu's answer tied the overrun to revenue beats and aggressive AI investment — which the table supports: revenue beat or met in every quarter the opex overran, so operating income was not missed. It is a guide that understates spend, not a margin that disappoints.
Open, not scored: server CPU "will grow by over 70% year-over-year" in 2027 and the Data Center segment "more than double" (Lisa Su, CALL-Q2-26); non-GAAP EPS "significantly higher than $20" (Hu, CALL-Q2-26, restating the Financial Analyst Day target); Q3 2026 revenue ~$13bn ±$300m, GM ~56%, opex ~$3.65bn (PR-Q2-26; Hu, CALL-Q2-26).
Are the promisers still there? Su and Hu hold their roles throughout; no row is departed.
Dodge log rolled up (§3.1; CALL-Q1-26 and CALL-Q2-26 read Q&A-first): four dodges; one topic dodged in two consecutive quarters — Instinct (data-center GPU) gross margin.
- Q1 2026 · Instinct gross-margin path · C.J. Muse (Cantor). "How are you thinking about that either today or in the coming 1, 2, 3 years?" Hu: "at this stage, we really focus on driving the topline revenue growth on our Instinct family of product".
CALL-Q1-26, turns 39–40. - Q2 2026 · Instinct gross margin entering and exiting 2027 · C.J. Muse. "how should we think about beginning 2027 and then exiting?" Hu answered the company-level mix — "it's really determined by both the pace of server CPU business ramp and the data center AI ramp".
CALL-Q2-26, turns 47–48. Tom O'Malley asked the same about 2027 gross margin earlier in the call and got "puts and takes". - Q1 2026 · Data Center AI sequential growth ex-China · Stacy Rasgon. Hu: "The China, for our business, in Q1, it's not material." Rasgon: "But you didn't answer my question."
CALL-Q1-26, turns 44–45. - Q2 2026 · 2027 gigawatts in line of sight and revenue per gigawatt · Vivek Arya (BofA). "how many gigawatt of compute does AMD have line of sight into for 2027 and what is AMD's monetization per gigawatt". Su answered segment growth rates ("server CPU ... will grow by over 70%"), not gigawatts or content.
CALL-Q2-26, turns 17–18.
The thread: management has never guided an Instinct margin, so this is not a dodge of its own guide — but it is the one number that decides whether the MI450/Helios ramp is accretive, the ramp starts in Q3 2026, and two analysts asked in two quarters. Carried to the scorecard ⚠️ line and to an S9 trigger.
Verdict: ✅ reliable. 23 scored rows: 6 beat, 10 met, 7 missed. Revenue 8/8 in or above range, gross margin 8/8 on guide (including an export charge the company put in the guide rather than adjusting out), and non-GAAP opex overrun 7/7 by a small, stable margin that tracks revenue. No first guide was cut and nothing was withdrawn. Current team made every row. No valuation haircut.
Gates this changes
- S2 ✅ stays; capital allocation: ⚠️ neutral (flat GAAP EPS since FY21, incremental ROIC ~0–4%). Not a destroyer, so no new kill criterion; the warrant criterion already covers the dilution risk.
- S3 ✅ (gate-minimum, partial) stays partial — two transcripts now read (closes the gap); comp structure still not extracted. EQ line added.
- S8 ⬜ → ✅ reliable.
- ⚠️ line: Instinct gross margin dodged in Q1 and Q2 2026.
- S9 trigger added: company non-GAAP gross margin in the first two Helios quarters (Q4 2026, Q1 2027) against the ~56% guided for Q3 2026 — the only public read on Instinct margin, since AMD does not disclose it. Next earnings date not sourced this pass.
Process support, not investment advice.
Full notes, 21 Sep 2026 12 min read
AMD — dossier, 2026-09-21
Slot names-ai-capex-1, name pass. Drafted unattended by the cloud research routine following .claude/skills/stock-sourcing-pipeline/SKILL.md. Nothing here is human-verified. Every figure carries the document it was read from; anything from background knowledge is marked [background]. Never edited after today — corrections go in scorecard.md's log.
Price reference: $559.82 close 2026-09-18 (market data). Shares outstanding 1,632,475,042 as of 2026-07-29 (EDGAR dei:EntityCommonStockSharesOutstanding, Q2 FY26 10-Q cover). Implied market capitalisation ≈ $914bn; ≈ $1.09tn if the 320m warrant shares are treated as outstanding.
S1 — Source
watchlist (owner, Apple Stocks, 2026-09-20), cluster ai_semis, coverage: sec_domestic, cik 2488. No Tier 0 promotion: queue.json promoted was empty at the start of this slot.
S2 — Kill test
EDGAR XBRL companyconcept. Fiscal years end late December; FY2025 ended 2025-12-27.
| $bn unless noted | FY2021 | FY2022 | FY2023 | FY2024 | FY2025 |
|---|---|---|---|---|---|
| Revenue | 16.4 | 23.6 | 22.7 | 25.8 | 34.6 |
| Gross profit | 7.9 | 10.6 | 10.5 | 12.7 | 17.2 |
| Gross margin (GAAP) | 48.2% | 44.9% | 46.1% | 49.4% | 49.5% |
| Operating income | 3.65 | 1.26 | 0.40 | 1.90 | 3.69 |
| Operating margin | 22.2% | 5.4% | 1.8% | 7.4% | 10.7% |
| Operating cash flow | 3.52 | 3.57 | 1.67 | 3.04 | 7.71 |
| Capex | 0.30 | 0.45 | 0.55 | 0.64 | 0.97 |
| Free cash flow | 3.22 | 3.12 | 1.12 | 2.41 | 6.74 |
| Cash and equivalents | 2.54 | 4.84 | 3.93 | 3.79 | 5.54 |
| Long-term debt (noncurrent) | 0.001 | 2.47 | 1.72 | 1.72 | 2.35 |
| Diluted shares (bn) | 1.229 | 1.571 | 1.625 | 1.637 | 1.636 |
Share count over five years. 1.229bn → 1.636bn, +33%. Almost all of that is the 2022 Xilinx acquisition [background: the step from 1.229 to 1.571 coincides with FY2022 and is acquisition-driven; the filings read this pass do not restate it]. Since 2022 the count has been flat to slightly down — AMD has been holding its share count, not growing it.
That is the fact about to change, and it is the whole risk. From the Q2 FY26 10-Q:
In October 2025 and February 2026, the Company issued warrants to OpenAI OpCo, LLC (OpenAI) and Meta Platforms, Inc. (Meta), each entitling the holder the right to purchase up to 160 million shares of the Company's common stock at an exercise price of $0.01 per share (the OpenAI Warrant and the Meta Warrant, respectively).
320 million shares at one cent, against 1,659m diluted in Q2 FY26 — ~19% dilution, and that is before the equity plan. AMD amended its 2023 plan on 2026-05-13, raising authorised shares by 65 million to 153 million, with 107 million still available at 2026-06-27. Total identified overhang: 427 million shares, ~26% of the current diluted count.
The warrants vest on Instinct purchase milestones plus stock price targets, and none had vested at 2026-06-27 — so nothing has hit the P&L yet. They are classified as liabilities until the conditions for equity classification are met.
Cash, debt and the maturity wall. Cash, cash equivalents and short-term investments were $10.6bn at FY25-end (from $5.1bn a year earlier) and $13.1bn at 2026-06-27 (Q2 call). Aggregate principal debt was $3.3bn at FY25-end (from $1.8bn). Then, in August 2026, AMD registered a new shelf (S-3ASR, 2026-08-13) and drew $4.75bn of senior notes off it the same day. Pro-forma debt ≈ $8.05bn against ≈ $17.8bn of cash and short-term investments — still net cash of roughly $9.8bn. Net debt/EBITDA is negative; there is no maturity wall.
Cash runway: not applicable. AMD is profitable and generated $6.7bn of FCF in FY25.
Analyst coverage count: not captured this pass. Thin.
Verdict — survives or dies: AMD survives comfortably. Revenue +34% in FY25 and +50% y/y in Q2 FY26, free cash flow $6.7bn, net cash, no going-concern issue. GAAP operating margin of 10.7% in FY25 is the weak number — an order of magnitude below NVIDIA's 60.4% — but it is rising and reached 17.2% in Q2 FY26. The kill risk here is not insolvency, it is that the equity is being handed to the customers: 320m warrant shares at $0.01 is, at today's price, roughly $179bn of value contingently transferred to two buyers in exchange for their purchase commitments.
S3 — Filings, at gate-minimum depth
Read this pass: FY25 10-K (Item 1A risk factors and Item 7 MD&A only), Q2 FY26 10-Q (statements, warrant note, segment MD&A), DEF 14A 2026-03-27 (ownership), the August-2026 424(b)(5), and the Q2 FY26 earnings call. One transcript, not two — the Equibles budget for this run was 25 calls across three names. Not read: the FY24 10-K, older 10-Qs, comment letters, critical audit matters, auditor tenure and fees. Note also a 10-K/A filed the same day as the 10-K (2026-02-04); the amendment itself was not read this pass — flagged for the next.
The three biggest risks management itself discloses
Management's own words, located by exact substring and copied with their locator.
1. Customer concentration — stated as a dependency, deliberately unquantified on revenue. 10-K FY25, Item 1A:
We depend on a small number of customers for a substantial portion of our business and we expect that a small number of customers will continue to account for a significant part of our revenue and receivables in the future.
The only number AMD actually gives is on receivables, in Note 11:
One customer accounted for approximately 11 % and another customer accounted for 24 % of the total consolidated accounts receivable balance as of December 27, 2025 and December 28, 2024, respectively.
No single customer above 10% of FY25 revenue is disclosed. That makes AMD materially less concentrated than either of the other names in this slot — NVDA discloses 22% and 14% of revenue in two customers, AVGO 32% in one. It is the one structural advantage AMD has here, and the warrants are in the process of trading it away: the whole point of the OpenAI and Meta agreements is to concentrate.
2. Export control, with a cost already booked. 10-K FY25, Item 1A:
In April 2025, the U.S. government implemented a new license requirement for the export of certain semiconductor products to a D5 Country, and to companies headquartered in, or with an ultimate parent located in such D5 Country. This restriction impacts our AMD Instinct™ MI308 products. As a result of the restriction, we incurred approximately $800 million in inventory and related charges in the second quarter of 2025.
Roughly $360m of that was reversed in Q4 2025 once licences were granted, for a net ~$440m — which is the entire difference between AMD's reported 50% FY25 gross margin and what it would otherwise have been.
3. Dependence on third-party manufacturing, now compounded by its own divestiture. 10-K FY25, Item 1A Risk Factors Summary:
We rely on third parties to manufacture our products, and if they are unable to do so on a timely basis in sufficient quantities and using competitive technologies, our business could be materially adversely affected.
AMD is fabless [background], and in the same year it bought ZT Systems for $3.2bn plus 8.3m shares and then sold ZT's manufacturing arm to Sanmina for $2.4bn plus 1.2m Sanmina shares and up to $450m contingent — keeping the designers, selling the factories, and naming Sanmina its preferred partner. The rack-scale ambition (Helios) rests on a contract manufacturer.
Latest quarter — Q2 FY26 (ended 2026-06-27)
| $m | Q2 FY26 | Q2 FY25 | y/y |
|---|---|---|---|
| Net revenue | 11,536 | 7,685 | +50% |
| Gross profit (GAAP) | 6,203 | 3,059 | +103% |
| GAAP gross margin | 53.8% | 39.8% | +14.0 pts |
| Operating income (GAAP) | 1,990 | (134) | — |
| GAAP operating margin | 17.2% | −1.7% | — |
| Net income | 2,297 | 872 | +163% |
| Diluted EPS | $1.38 | $0.54 | +156% |
| Diluted shares (m) | 1,659 | 1,630 | +1.8% |
The Q2 FY25 comparison flatters: it carried the ~$800m MI308 charge. Segments: Data Center $6.7bn, +107% y/y, 58% of revenue (42% a year ago), operating income $2.1bn against a $155m loss; Client and Gaming $3.8bn (+6%), within which Gaming fell 31% to $779m; Embedded $977m (+19%). Stock-based compensation was $990m in the first half, versus $733m.
Guidance given on the call (machine transcript, wording mangled in places): Q3 FY26 revenue ~$13.0bn ±$300m (+41% y/y), non-GAAP gross margin ~56%, opex ~$3.65bn, tax 13%, diluted share count ~1.66bn. Beyond that: server revenue +>80% y/y in H2 2026 and +>70% in 2027; Data Center segment revenue to more than double year-over-year in 2027; and an explicit statement that AMD expects to "significantly exceed our $20 annual EPS target within our strategic timeframe."
New customer announcements on the same call: Anthropic for up to 2 GW of MI450 in Helios, first gigawatt in H1 2027; Microsoft deploying Helios at scale on Azure. Neither is in a filing read this pass.
Proxy skim (DEF 14A filed 2026-03-27)
Based on 1,630,338,779 shares of our common stock outstanding as of March 19, 2026.
All 16 current directors and executive officers together: 8,107,566 shares, marked "*" — less than one percent. Insider alignment at AMD is through compensation, not ownership. That is a genuine contrast with NVIDIA (3.94%) and Broadcom (1.9%) and it matters more here, because AMD's board is the body approving warrants that hand ~19% of the company to two customers. Compensation structure and what the bonus is paid on: not extracted this pass — thin.
Insider pattern (12 months to 2026-09-21)
Zero open-market purchases by any insider (Equibles GetInsiderTransactions AMD, transactionType: Buy, fromDate: 2025-09-21 → "No insider transactions found"). 612 Form 4s on the current EDGAR submissions page, newest 2026-09-17; none of it is conviction buying.
Red-flag sweep
- Going concern: no flag. Newest examined filing is the 10-Q filed 2026-08-05; coverage there starts 2026-05-06, so earlier filings are unexamined.
- Executive changes (8-K Item 5.02, via Equibles): board only. Tim Ryan appointed 2026-08-19; Joe Householder retired the same day, two days after giving notice. Eight directors elected at the 2026-05-13 annual meeting. No CEO or CFO change — Lisa Su and Jean Hu are both in seat.
- Auditor change / restatement / NT 10-K: not run this pass — the
efts.sec.govfull-text sweep was dropped for time. Thin; flagged for the next pass. The 10-K/A of 2026-02-04 should be read then. - Leases not yet commenced: $1.3bn of data center and other real estate leases signed but not commenced at FY25-end, expected to start in 2026 with terms of 1 to 10 years. AMD is taking on data center property of its own — small against the balance sheet, new in character.
S5 — Ownership
| Insider ownership | <1% — 8,107,566 shares across 16 directors and executive officers, DEF 14A at 2026-03-19 |
| Largest 13F holders, 2026-06-30 | BlackRock 150.3m sh (10.77% of institutional total), Vanguard Capital Mgmt 106.5m (7.63%), State Street 74.4m (5.33%), Invesco 53.2m (3.81%), JPMorgan 40.3m (2.89%) |
| Notable option lines | Jane Street puts over 34.2m shares (rank 9) and calls over 25.2m (rank 12); Susquehanna puts over 25.3m (rank 11). Market-making books at notional value, not directional views — but they rank above FMR's common stock line |
| 13F breadth | 4,213 rows across 4,024 institutions, 1.395bn shares total |
| Short interest (FINRA) | 41,712,630 shares at 2026-08-31 settlement, ≈2.6% of shares outstanding, 2.5 days to cover |
| Short trend | 41.6m (06-30) → 39.8m → 37.8m → 40.1m → 41.7m (08-31) — drifting back up off the July low |
| Buyback | $9.2bn remaining as of 2026-06-27 ($9.4bn at FY25-end); $1.3bn / 12.4m shares repurchased in FY25 |
| Dividend | none |
| Shelf | S-3ASR filed 2026-08-13 and drawn for $4.75bn of notes on 2026-08-13/14. A live, freshly used shelf hangs over the name |
| ATM | No ATM program captured in ~13 months of 10-K/10-Q/8-K (Equibles). Absence there is not proof the company runs none — and it would be the wrong place to look anyway, because AMD's equity dilution is coming through warrants, which are disclosed in the notes rather than as an ATM |
Who is on the other side: short interest is 2.6% with 2.5 days to cover — the highest of the three names in this slot, but still not a short case. The more interesting answer is that OpenAI and Meta are on the other side of the equity, holding 320m shares' worth of one-cent warrants. They are not sellers, but they are the counterparties whose exercise sets the share count.
S8 — Valuation and sizing (model; both decisions are ⏳ OPEN (user))
AMD guides on a non-GAAP basis and the gap to GAAP is wide (Q2 FY26: 56% non-GAAP gross margin versus 53.8% GAAP; 27% non-GAAP operating margin versus 17.2% GAAP), almost all of it acquisition-intangible amortisation and stock-based compensation. Both are modelled below.
| Model, $bn except per share | FY2026E | FY2027E |
|---|---|---|
| Revenue | ~49.8 | ~79 |
| basis | H1 actual 21.8 + Q3 guide 13.0 + Q4 modelled ~15.0 | Data Center "more than double" per guidance, Client/Gaming and Embedded modelled flat to slightly up |
| Non-GAAP gross margin | ~56% | ~57% |
| Non-GAAP operating income | ~13.7 | ~28 |
| Non-GAAP net income | ~12.2 | ~24.5 |
| Diluted shares (bn) | ~1.66 | ~1.78 (assumes ~100m of the 320m warrant shares vest) |
| Non-GAAP EPS | ~$7.3 | ~$13.8 |
| P/E at $559.82 | ~77x | ~41x |
| P/E if all 320m warrant shares vest | — | ~45x |
| At management's ">$20 EPS" | — | <28x |
EV: market cap ~$914bn less ~$9.8bn pro-forma net cash → EV ≈ $904bn, ≈ 18x FY2026E revenue and ≈ 11x FY2027E revenue.
- Bear: Data Center does not double in 2027 — it grows 50% — while Gaming keeps shrinking and memory cost holds gross margin at 54%. FY27 revenue ~$63bn, non-GAAP EPS ~$9. At 30x that is ~$270, roughly half today's price. And the warrants still dilute on the stock-price tranches.
- Base: guidance is met. ~$13.8 FY27 EPS at ~41x today.
- Bull: the $20 EPS target lands early on Helios and MI500, and 45x on $20 is ~$900.
The honest summary of the valuation: AMD is the expensive one. NVIDIA trades at ~14x model FY28 EPS and Broadcom at ~12x management's own FY28 EPS target; AMD needs its Data Center segment to double in a single year merely to get to ~41x. It is priced as the option on the second-source thesis, and the warrants mean that if the thesis works, holders own ~19% less of it.
Liquidity: average daily volume 16,772,687 shares (FINRA, 2026-08-31 settlement) at ~$560 ≈ $9.4bn traded per day. No sizing constraint.
⏳ Buy price: OPEN (user). Size: OPEN (user). Both are owner-only under the skill.
S9 — Monitoring
- Next earnings: Q3 FY2026 ends ~2026-09-26; AMD has reported that quarter in the first week of November in recent years. Derived from the reporting cadence, not from a sourced calendar — market data's earnings calendar is behind a paid plan. Thin.
- Filing feed: Tier 0 EDGAR sweep (
watch_edgar.py, 06:40 UTC) covers cik 2488. - Leading indicator between quarters: the Warrants note in each 10-Q. It states, in one sentence, how many warrant shares have vested and become exercisable. Vesting keyed to purchase milestones is a direct, filed read on whether OpenAI and Meta are actually buying — better than any revenue commentary, because it is the customers' own behaviour reported by AMD. Second: Data Center segment revenue against the "more than double in 2027" guide.
- Trigger that converts this to a buy: owner's call. What this routine can say is the condition that would resolve the central objection — warrant tranches vesting on purchase milestones (not on stock-price targets) while Data Center revenue tracks the doubling guide. That would mean the dilution bought real volume. The adverse case is tranches vesting on price targets while revenue lags, which is dilution paid for nothing.
Access and connector notes for this run
Same as the NVDA dossier of this date: SEC EDGAR reachable with the required User-Agent; Equibles reachable, 22 calls spent across the slot against a 25 cap; market data reachable for prices but its earnings calendar is gated; EDINET DB and eu_data not needed for sec_domestic names and not called; Alpha Vantage, Bigdata.com and FMP not called, per the skill.
Process support, not investment advice.