AMD Advanced Micro Devices
Dive Nov 16, 2026. Price at dive $507.91. Priced high Price at the dive was above our base value.
About 15M World Labs shares issued at the Nov 13 close are not in the share count.
Researched Cases
- Bear
- $246.25
- Base
- $278.27
- Bull
- $313.25
The base case is the forecast believed most likely, one path. Bear and bull are the same path with the top three swing factors each moved by a researched step, all three in the same direction. Everything else is identical across cases.
Key Assumptions, Stated
- Discount rate (WACC)
14.56% - Terminal growth
3.5% - Shares
1,661M diluted - Balance sheet
Net cash $8,697M
Identical in all three cases. Built from SEC filings. Equity risk premium: 3.70% (Damodaran, Oct 8, 2026).
Swing Factors
The three swing factors (KPIs: the few numbers the value hinges on) behind the cases. Each row shows the researched base path and the bear and bull flexes.
| Swing Factor | Base Path | Bear Flex | Bull Flex | Why It Matters |
|---|---|---|---|---|
| Revenue growth | 82.0% in FY2027, fading 38%/24%/16%/12%/9.5%/7.5%/6.0%/5.0%/4.5% | each year's growth ×0.95 (FY2027: 77.9%) | each year's growth ×1.05 (FY2027: 86.1%) | Every forecast cash-flow dollar scales with revenue, and the 2027 step of over 80% is the single largest value driver in the model. |
| Operating cash-flow margin | 23% in FY2027 rising to 30% by FY2031, then 30% | 1.0pt lower each year: 22% rising to 29% | 1.0pt higher each year: 24% rising to 31% | Rack-scale systems absorb working capital before they release it, so whether the Q3 inventory build converts or persists swings realized cash flow by billions. |
| Capex intensity (capex as % of sales) | 5.0% in FY2027 fading to 3.0% by FY2031, then 3.0% | 0.5pt higher each year: 5.5% fading to 3.5% | 0.5pt lower each year: 4.5% fading to 2.5% | Capex intensity nearly doubled from 2.8% in FY2025 to 5.2% in 9M 2026, and it is deducted dollar-for-dollar from operating cash flow in every forecast year. |
Run Your Own Version
Each workspace has three sliders, one per swing factor. The sliders need JavaScript; the researched values above work without it. Move the sliders past the bear and bull marks, or type any value you believe. The researched cases stay shown above. Your value updates as you go.
See the research
Summary
Base $278.27 rests on three findings. First, the demand step is guided, not assumed: AMD guided Q4 2026 revenue of about $17B and 2027 growth of more than 80%, backed by named Helios deployments at Anthropic, Meta, Microsoft, OpenAI and Oracle and multi-gigawatt purchase commitments. Second, cash conversion must recover: 9M 2026 operating cash-flow margin was 18.5% after a $3.6B working-capital build (inventory up $3.5B), and the base path assumes that inventory converts as the ramp ships, lifting the margin toward 30%. Third, the discount rate is punishing: beta near 2.6 with a 5.10% risk-free rate gives a 14.56% WACC, which caps the value of even this growth. Evidence that would change the call: Q4 revenue missing the $17B guide or inventory failing to convert; HBM allocation shortfalls forcing a lower 2027 guide; warrant tranches vesting faster than expected (up to 320M shares); a durable fall in the risk-free rate or beta; or Nvidia's Nov 17 earnings resetting the AI demand read. Educational only; no buy or sell recommendation.
Assumptions, Stated
| Input | Researched Value | Decision | Evidence |
|---|---|---|---|
| Risk-free rate | 5.10% | kept, the tool default matches the latest published observation; no adjustment. | FRED DGS10 printed 5.10% on Nov 12, 2026, up from 4.98% on Nov 10 and 4.84% a week earlier; the Yahoo Finance ^TNX cross-check of 5.12% in the tool payload agrees within 0.3%. Source: FRED DGS10. Source |
| Equity risk premium | 3.70% | kept, the labeled fallback equals the verified current headline figure; the stale 4.20% variant is rejected. | Damodaran's current headline implied ERP is 3.70% from his Oct 8, 2026 computation: expected return on stocks 8.99% less the then 10-year Treasury 5.29%, trailing-12-month cash-flow basis. No fresher headline figure was found in this research. The 4.20% figure dated Oct 1, 2026 is the adjusted-payout variant, an older and different computation; presenting it as the current headline would be an error under the fallback rule, and it is not used. The tool payload's labeled fallback already carries the correct 3.70% headline value. Source |
| Beta | 2.58 | kept, widening the vendor set does not move the median; beta near 2.6 is a genuine property of this stock, not a vendor artifact. | Four readings: managed channel/Finnhub 2.62, StockAnalysis beta (5Y monthly) 2.54, MarketBeat 2.49, and the tool's self-computed 5-year monthly regression vs the S&P 500 at 2.61. Dispersion is 0.13 across vendors. The four-vendor median is 2.58, identical to the tool median. Source |
| Cost of debt (pre-tax) | 5.05% (observed) | replaced, observed recent-note coupons replace the synthetic coverage-based rate; after tax at 25% the cost of debt is 3.79%, and at a 0.8% debt weight the WACC effect is under 0.01pt. | AMD issued $4.75B of senior notes on Aug 17, 2026: $1.25B 4.600% due 2029, $1.5B 5.000% due 2031, $1.0B 5.250% due 2033, $1.0B 5.500% due 2036 (424B5 prospectus supplement). Principal-weighted coupon is 5.05%. An observed issuance rate outranks any coverage formula. Source |
| Discount rate (WACC) | 14.56% | adjusted, immaterial 0.01pt move from the observed cost of debt; the WACC is dominated by the equity leg. | CAPM: 5.10% + 2.58 x 3.70% = 14.65% cost of equity. Market weights: equity about $843.6B (price $507.91 x 1,661M diluted shares) vs debt $7,079M, giving 99.2% equity and 0.8% debt. WACC = 14.65% x 99.2% + 3.79% x 0.8% = 14.56%. |
| Revenue path | FY2026E $52,166M; growth 82% (FY2027) fading to 4.5% (FY2036) | replaced, a revenue-anchored path from filed actuals and company guidance replaces the fragile EPS-conversion bridge (3-year conversion history 0.38x, 0.70x, 1.20x). | 9M 2026 revenue of $35,166M is filed in the Q3 10-Q; Q4 guidance of about $17B (+/-3%) gives FY2026E $52,166M (+50.6% vs FY2025 $34,639M). Management guided 2027 revenue growth of more than 80%, server revenue doubling, and Data Center over 80% of revenue; the base uses 82% for FY2027, the floor plus 2pts for Q4 exit-rate momentum (Q4 Data Center guided up about 50% sequentially) and named first Helios deployments at five customers. Growth then fades 38%/24%/16%/12%/9.5%/7.5%/6.0%/5.0%/4.5% as gigawatt deployments mature and the PC market declines double digits in 2027 per management. FactSet 2027 consensus near $90.8B (pre-print, on a lower 2026 base) is a cross-check only. Risk: the HBM shortage constrains shipments and AMD pays a reported premium for HBM4. Source |
| Operating cash-flow margin | 23% (FY2027) rising to 30% (FY2031), then 30% | replaced, an explicit margin path from filed cash-flow history replaces the unstable conversion ratio. | OCF margin history from filings: 7.4% FY2023, 11.8% FY2024, 22.3% FY2025, 18.5% in 9M 2026, 20.1% TTM. The 9M 2026 dip reflects a $3.6B working-capital outflow (inventory +$3.5B to support the Data Center ramp, receivables +$2.9B, supply prepayments +$2.1B, partly offset by payables +$4.5B); Q3 alone ran 9.0%. Non-GAAP operating margin was 29% in Q3 and the Q4 guide implies about 33%, and management said opex growth will run at less than half of revenue growth. The base path assumes the inventory build converts as Helios ships, so the margin recovers past the FY2025 level and levels at 30% as growth decelerates. Source |
| Capex intensity (capex as % of sales) | 5.0% (FY2027) fading to 3.0% (FY2031), then 3.0% | adjusted, an explicit fading path replaces the implicit history; the level is set by the recent run-rate, not the older fabless-light history. | Capex intensity from filings: 2.4% FY2023, 2.5% FY2024, 2.8% FY2025, 5.2% in 9M 2026 ($1,845M), 4.5% TTM, 7.0% in Q2 2026 and 4.8% in Q3 2026. The step-up coincides with the rack-scale build-out and accrued-but-unpaid equipment purchases of $505M at quarter-end. The base path starts at the recent run-rate (5.0%) and fades to 3.0% as revenue scale outruns lab and test capacity spending. Source |
| Terminal growth | 3.5% | kept, 3.5% is a fair mature-state rate; the memo's sensitivity span shows +/-0.5pt explicitly. | No company outgrows the economy forever; 3.5% sits below the 5.10% risk-free rate (implying modest real growth) and below the year-10 forecast growth of 4.5%, so the terminal year steps down rather than jumping. The AI runway is long, but terminal growth must describe a mature AMD facing Nvidia, custom silicon and cyclical memory markets, not the current ramp. |
| One-time and below-the-line items | Warrant remeasurement $1,097M, investment gains $616M (9M), tax benefit $555M (Q3): all excluded from operating read | adjusted, items are identified and kept out of the operating read; the warrant charge is non-cash and excluded from non-GAAP results, but the underlying dilution is carried in the sensitivity lenses, not ignored. | Q3 2026 other income (expense) of -$902M includes a $1,097M non-cash change in fair value of the customer warrant liability (10-Q Note 14: warrants to OpenAI and Meta, up to 160M shares each at $0.01, weighted-average fair value $178.58 vs $14.25 grant date; no shares vested). The 9M cash-flow statement adds back $1,179M of non-cash warrant impact and strips $616M of gains from long-term investments. Q3 also carries a $555M income tax benefit on $1,868M pre-tax income. Amortization of acquisition-related intangibles was $1,629M in 9M 2026 and is non-cash. The DCF is built on OCF less capex from filed statements, and the $4.5B unrecognized warrant grant-date value that will reduce future revenue is disclosed as a drag of under 1% of annual revenue spread over the deals. Source |
| Diluted share count | 1,661M | kept, the filed latest-quarter diluted count is the divisor, as specified; future dilution is disclosed, not buried. | Q3 2026 10-Q income statement: weighted-average diluted shares 1,661M for the quarter (1,657M for 9M), basic 1,636M; cover shares outstanding 1,640,087,251 as of Oct 28, 2026. Buybacks are paused: $221M repurchased in 9M 2026 vs $1,316M a year earlier, while $1,836M went to tax withholding on employee equity plans and 65M treasury shares were retired. Q4 guidance assumes about 1.67B diluted shares. Excluded from the count, by GAAP: up to 320M warrant shares (none vested) and the pending all-stock World Labs acquisition (about $8.2B) plus the Taalas deal (about $2.9B cash and stock); both are flagged in the sensitivity and world-context lines rather than added to today's divisor. Source |
| Net debt (EV-to-equity bridge) | -$8,697M net cash | kept, the filed composition reconciles; contingent customer-support items are real risk but are maximum, contingent amounts, not debt owed today. | Q3 10-Q balance sheet: long-term debt $7,079M (current portion nil after the $875M 4.212% notes matured in September 2026, refinanced in part by the $4.75B August notes), cash $7,404M plus short-term investments $8,372M = $15,776M. Long-term operating lease liabilities of $1,837M are excluded from net debt because lease payments already run through operating cash flow; adding them would double-count. Off-balance-sheet items are disclosed, not added: $48.8B unconditional purchase commitments ($33.4B due by end FY2027), $6.2B cloud compute capacity commitments, $13.3B data-center leases not yet commenced (to be assigned or subleased), $8.5B maximum guarantees (no partner in default, nothing paid), and up to $5.1B investment commitments. Source |
| Stock-based compensation | About $1,999M TTM; captured via diluted shares, SBC lens in sensitivity | kept with lens, no double deduction in the headline; the SBC cost is quantified in the sensitivity span instead of hidden. | SBC was $1,513M in 9M 2026 (Q3 $523M) per the 10-Q cash-flow statement; TTM is about $1,999M (FY2025 $1,638M less 9M 2025 $1,152M plus 9M 2026 $1,513M), about 4.4% of TTM revenue. SBC is a real shareholder cost: it is added back inside OCF, so headline FCF (OCF less capex) overstates owner earnings. The base case follows the specified bridge (diluted shares carry the dilution from awards already granted) and the sensitivity span carries an explicit SBC lens deducting 4.0% of revenue each year, worth $235.87 per share. Source |
| Demand and competitive context | Reflected in the revenue fade; no separate premium or discount | adjusted, context shapes the fade and the risk disclosure; no speculative market-share gain from a competitor's leadership change is baked into the numbers. | Nvidia's founder-CEO Jensen Huang resigned Nov 12, 2026, with CFO Colette Kress as interim CEO ahead of Nov 17 earnings; Nvidia remains far larger (Q2 FY2027 revenue $96.2B) and no leadership-change benefit is assumed for AMD. The industry faces an HBM and memory shortage: suppliers report HBM capacity sold out through 2027, AMD flagged memory costs and guided Q4 gross margin down 1pt to 55%, and the PC market is expected to fall double digits in 2027. Hyperscaler capex of roughly $650-750B in 2026 is expected to rise toward $1T-plus in 2027, which supports the demand side of the 2027 guide. AMD also disclosed about $11B of pending acquisitions (Taalas and World Labs) expected to close by end of 2026. Source |
What Would Have to Be True
For the $507.91 price to be right on this base cash-flow path, the discount rate would have to be 9.9% instead of the researched 14.6%, or perpetual terminal growth would have to be 10.5% instead of 3.5%, or the revenue, margin and capex path would have to beat the base case by more than the full stacked-favorable variant. In plain terms, the price requires AMD to be priced as a much safer company than its beta and the current Treasury yield imply, or to grow forever at a rate no mature company sustains.
Sensitivity Span (Detail Only)
$217.37 to $359.77. Full variant span on the base path: low $217.37 is the stacked adverse variant (WACC +0.5pt, growth 2pts lower each year, terminal growth 3.0%, OCF margin 1pt lower, capex 0.5pt higher); high $359.77 is the stacked favorable mirror. Single-factor variants inside the span: WACC +/-0.5pt $265.02/$292.82, growth +/-2pts $246.24/$314.81, terminal growth +/-0.5pt $272.48/$284.62, beta 2.49/2.62 $287.71/$274.24, SBC lens (FCF less 4.0% of revenue) $235.87, warrant dilution lenses +160M/+320M shares $253.82/$233.32. Detail only; the headline range is bear-to-bull. The span is detail only, never the headline range.
Cross-Checks (Tool Only)
Cross-checks only; the headline is the researched DCF. Graham $443.60 uses the 62.7% average analyst EPS growth rate and inherits its optimism. Multiples $162.98 uses a basket reduced to MSFT and ORCL after NVDA and AVGO were excluded, with the P/B leg excluded as reference-only, so it is a weak comp for a hypergrowth accelerator ramp. The tool's own DCF $222.76 and blend $227.37 run on the fragile 0.86x earnings-to-FCF conversion ratio (3-year history 0.38x, 0.70x, 1.20x), which this memo replaces with a revenue-anchored cash-flow path. All three sit below the $507.91 price; none is the answer.
Smell Tests
- Implied P/E at the base value is 60x trailing GAAP EPS of $4.61, above the 40x flag. Trailing GAAP EPS is depressed by the $1.1B warrant remeasurement charge and $2.2B of TTM acquired-intangible amortization, and the base value is 17x FY2027 consensus EPS of $16.04, so the flag reflects a depressed trailing base plus priced-in hypergrowth rather than an arithmetic error. The growth still has to arrive.
- Market-implied discount rate on the base path is 9.9% at the $507.91 price, vs the researched 14.56%: a 4.7pt gap. The market prices AMD as far less risky than CAPM with a 2.58 beta and a 5.10% Treasury yield implies.
- Market-implied terminal growth at the $507.91 price is 10.5% on the base path, more than double the 5% cap the model refuses to exceed. The price requires perpetual growth no mature company sustains.
- The Gordon terminal value is 44% of enterprise value, inside the roughly 70% institutional comfort line. The value does not hinge entirely on perpetuity; the explicit 10-year cash flows carry the majority.
Earlier Dives
A new dive never rewrites an old row. Earlier dives stay here, marked superseded.
- Superseded: dive Nov 14, 2026. Base $269.40, bear $237.71, bull $304.07. Price at dive $534.06. Superseded by dive #3 (Nov 16, 2026).
- Superseded: dive Nov 4, 2026. Base $270.35, bear and bull cases not carried in the ledger. Price at dive $483.90. Superseded by dive #2 (Nov 14, 2026).
Educational only. Not financial advice.