Decision time: May 3, 2016 at 9:30 a.m. EDT. Knowledge cutoff: May 2, 2016 at 11:59:59 p.m. UTC. This learner packet contains no post-cutoff outcome evidence.
Decision
AMD must choose among four bounded alternatives: initiate or add exposure; maintain without new gates; defer and re-underwrite; or route a possible reduction to an authorized human. The selected alternative is defer and re-underwrite. Initiate or add no capital, make no unconditional-maintain decision, and do not size or execute a trade. Any existing holding goes to authorized human re-underwriting. judgment.amd.cutoff.defer-reunderwrite judgment.amd.cutoff.valuation-abstention
This is an evidence-readiness decision, not a claim that the roadmap must fail. Roadmap intent, OEM sampling, licensing, and asset-sale proceeds justify continued research. Weak reported economics, negative operating cash flow, funding and supplier constraints, missing independent product proof, and absent valuation inputs prevent a capital action. claim.amd.cutoff.roadmap-intent claim.amd.cutoff.polaris-sampling claim.amd.cutoff.financial-compression claim.amd.cutoff.funding-strain claim.amd.cutoff.technical-validation-missing claim.amd.cutoff.valuation-not-ready
What was observable
AMD identified Zen as a new x86 core intended to support re-entry into high-performance and server computing, while explicitly warning that development, qualification, shipment, customer value, and timing were not assured. AMD also reported that Polaris had been demonstrated and sampled to OEMs during Q1 2016. Sampling is a milestone; it is not production qualification, volume, yield, customer availability, adoption, or product economics. claim.amd.cutoff.roadmap-intent claim.amd.cutoff.polaris-sampling evidence.amd.cutoff.zen-roadmap-risk evidence.amd.cutoff.polaris-sampling source · amd.cutoff.q1-2016-10q
PCWorld contemporaneously relayed AMD's claim of a 40 percent Zen performance improvement and quoted CEO Lisa Su on high-end desktop investment. That article supplies context for the issuer's public narrative, not an independent benchmark. The artifact is a later-captured copy whose embedded publication metadata says May 6, 2015; it is not a cutoff-time archive receipt. claim.amd.cutoff.independent-zen-context evidence.amd.cutoff.pcworld-zen-context source · amd.cutoff.pcworld-zen-2015
The THATIC structure gave AMD majority and minority interests in two entities through contributed patents. AMD stated that the partner would fund operations, AMD had no obligation to fund the ventures, and approximately USD 293 million of license fees were payable over several years only if milestones were achieved. AMD recognized USD 7 million of related operating income in Q1 2016. These are licensing terms and recognized income, not product-market validation. claim.amd.cutoff.thatic-terms fact.amd.cutoff.thatic-license-fees fact.amd.cutoff.q1-2016-licensing-income evidence.amd.cutoff.thatic-terms evidence.amd.cutoff.thatic-recognition source · amd.cutoff.q1-2016-10q
On April 29 AMD disclosed the sale of 85 percent of its Penang and Suzhou ATMP joint ventures for approximately USD 371 million before adjustments, with approximate net proceeds of USD 320 million after taxes and customary expenses; AMD retained 15 percent. The issuer expected the transaction to be P&L cost-neutral and to reduce capital expenditure, but that expectation was not recurring operating cash flow or an audited required-reinvestment schedule. claim.amd.cutoff.atmp-proceeds claim.amd.cutoff.manufacturing-capital-boundary fact.amd.cutoff.atmp-gross-proceeds fact.amd.cutoff.atmp-net-proceeds evidence.amd.cutoff.atmp-proceeds evidence.amd.cutoff.atmp-issuer-expectation source · amd.cutoff.atmp-closing-8k source · amd.cutoff.atmp-closing-release
Reported financial base
| Consolidated measure (USD millions) | FY2015 | Q1 2016 |
|---|---|---|
| Revenue | 3,991 | 832 |
| Gross profit | 1,080 | 269 |
| R&D expense | 947 | 242 |
| Operating income (loss) | (481) | (68) |
| Net income (loss) | (660) | (109) |
| Operating cash flow | (226) | (42) |
| Gross cash PPE purchases | 96 | 26 |
The annual and quarterly columns have different durations and are not directly comparable or annualized. Every value above is an as-reported consolidated fact. The gross cash-PPE line remains a positive cash-use measure; subtracting it from operating cash flow is only a scale check, not an authoritative free-cash-flow definition. fact.amd.cutoff.fy2015-revenue fact.amd.cutoff.fy2015-gross-profit fact.amd.cutoff.fy2015-rd fact.amd.cutoff.fy2015-operating-income fact.amd.cutoff.fy2015-net-loss fact.amd.cutoff.fy2015-ocf fact.amd.cutoff.fy2015-gross-cash-ppe fact.amd.cutoff.q1-2016-revenue fact.amd.cutoff.q1-2016-gross-profit fact.amd.cutoff.q1-2016-rd fact.amd.cutoff.q1-2016-operating-income fact.amd.cutoff.q1-2016-net-loss fact.amd.cutoff.q1-2016-ocf fact.amd.cutoff.q1-2016-gross-cash-ppe table.amd.cutoff.reported-results table.amd.cutoff.cash-flow claim.amd.cutoff.financial-compression source · amd.cutoff.fy2015-10k source · amd.cutoff.q1-2016-10q
Segment economics
| Segment measure (USD millions) | FY2014 | FY2015 | Q1 2015 | Q1 2016 |
|---|---|---|---|---|
| Computing and Graphics revenue | 3,132 | 1,805 | 532 | 460 |
| Computing and Graphics operating income (loss) | (76) | (502) | (75) | (70) |
| Enterprise, Embedded and Semi-Custom revenue | 2,374 | 2,186 | 498 | 372 |
| Enterprise, Embedded and Semi-Custom operating income | 399 | 215 | 45 | 16 |
The annual figures show Computing and Graphics revenue contraction and a materially wider operating loss. The year-over-year quarterly comparison remained negative in revenue, while the quarterly operating loss narrowed modestly. Enterprise, Embedded and Semi-Custom remained profitable in the reported periods but also declined. These segment definitions do not isolate Zen, Polaris, licensing, product units, ASP and mix, gross margin, or roadmap-product contribution. fact.amd.cutoff.fy2014-cg-revenue fact.amd.cutoff.fy2014-cg-operating-income fact.amd.cutoff.fy2014-eesc-revenue fact.amd.cutoff.fy2014-eesc-operating-income fact.amd.cutoff.fy2015-cg-revenue fact.amd.cutoff.fy2015-cg-operating-income fact.amd.cutoff.fy2015-eesc-revenue fact.amd.cutoff.fy2015-eesc-operating-income fact.amd.cutoff.q1-2015-cg-revenue fact.amd.cutoff.q1-2015-cg-operating-income fact.amd.cutoff.q1-2015-eesc-revenue fact.amd.cutoff.q1-2015-eesc-operating-income fact.amd.cutoff.q1-2016-cg-revenue fact.amd.cutoff.q1-2016-cg-operating-income fact.amd.cutoff.q1-2016-eesc-revenue fact.amd.cutoff.q1-2016-eesc-operating-income table.amd.cutoff.segments claim.amd.cutoff.segment-economics claim.amd.cutoff.product-proof-separation
Balance sheet, funding, and manufacturing boundary
| Balance-sheet measure (USD millions) | Dec. 26, 2015 | Mar. 26, 2016 |
|---|---|---|
| Cash and cash equivalents | 785 | 716 |
| Current assets | 2,320 | 2,226 |
| Current liabilities | 1,403 | 1,328 |
| Short-term debt | 230 | 230 |
| Long-term debt | 2,032 | 2,006 |
The balance sheet supplies scale, not a conclusion that liquidity was sufficient. AMD also reported USD 208 million of Q1 2016 GLOBALFOUNDRIES purchases and approximately USD 300 million of currently known obligations. It could not meaningfully quantify future obligations beyond that amount while negotiating another WSA amendment and expected future purchases to remain material. fact.amd.cutoff.fy2015-cash fact.amd.cutoff.fy2015-current-assets fact.amd.cutoff.fy2015-current-liabilities fact.amd.cutoff.fy2015-short-debt fact.amd.cutoff.fy2015-long-debt fact.amd.cutoff.q1-2016-cash fact.amd.cutoff.q1-2016-current-assets fact.amd.cutoff.q1-2016-current-liabilities fact.amd.cutoff.q1-2016-short-debt fact.amd.cutoff.q1-2016-long-debt fact.amd.cutoff.q1-2016-gf-purchases fact.amd.cutoff.q1-2016-gf-known-obligations table.amd.cutoff.balance-sheet table.amd.cutoff.transaction claim.amd.cutoff.funding-strain claim.amd.cutoff.manufacturing-capital-boundary evidence.amd.cutoff.gf-obligations source · amd.cutoff.fy2015-10k source · amd.cutoff.q1-2016-10q
AMD's FY2015 filing also described Moody's and Standard & Poor's downgrades and warned that an inability to finance or generate sufficient operating cash could force project abandonment or lower R&D and strategic investment. Approximate ATMP proceeds oppose an immediate cash-exhaustion narrative, but they do not resolve maturity timing, interest, commitments, working capital, required reinvestment, or recurring cash generation. evidence.amd.cutoff.fy2015-ratings-cash-risk evidence.amd.cutoff.atmp-proceeds conflict.amd.cutoff.proceeds-versus-funding-strain assumption.amd.cutoff.funding-resilience
Deterministic checks and their limit
The cutoff model performs only Decimal-based comparisons: revenue changes, gross and segment margins, current ratios, debt and net-debt scale, cash change, approximate ATMP proceeds relative to debt, and operating cash flow after gross cash-PPE purchases. Inputs are reported facts; no prose-generated number is authoritative. The run has no frozen model-registry execution receipt and is therefore non-authoritative. Its outputs cannot establish free cash flow, financing capacity, product performance, intrinsic value, target price, or a trade. table.amd.cutoff.deterministic-checks judgment.amd.cutoff.valuation-abstention
Six locked gates
| Gate | What must be established | Cutoff status |
|---|---|---|
| Technical validation | Independently reproducible workload, performance per watt, reliability, security, and total-system cost | Unknown; roadmap statements and sampling only |
| Roadmap execution | Tape-out, qualification, schedule, production volume and yield, customer availability and adoption | Partial; Polaris sampling observed, conversion unverified |
| Segment economics | Stable revenue, ASP and unit mix, gross margin, and operating contribution excluding licensing and restructuring | Unknown at product level; reported segments insufficient |
| Manufacturing and capital | Foundry access, WSA and supply commitments, concentration, gross cash PPE, and required reinvestment | Partial; material purchases and approximate obligations disclosed, forward schedule incomplete |
| Funding resilience | Cash, debt, maturities, interest, commitments, operating cash flow, and gross cash PPE | Incomplete; cash, debt, and cash flow reported, full downside schedule absent |
| Valuation readiness | Exact price and security, diluted capitalization bridge, approved forecast, mandate, exposure, and risk budget | Missing; valuation and action abstention required |
No gate can be substituted for another. A product sample is not qualification; a license receipt is not recurring product economics; an asset-sale proceed is not operating cash flow; a current ratio is not downside funding sufficiency; and a promising roadmap is not price attractiveness. judgment.amd.cutoff.six-gates table.amd.cutoff.gate-gaps claim.amd.cutoff.technical-validation-missing claim.amd.cutoff.product-proof-separation claim.amd.cutoff.valuation-not-ready
Alternatives and scenarios
The four alternatives are deliberately action-bounded:
- Initiate or add exposure — rejected at the cutoff because all six gates did not clear.
- Maintain without new gates — rejected because it would erase the unresolved technical, economic, manufacturing, funding, and valuation boundaries.
- Defer and re-underwrite — selected status quo; initiate or add no capital and route any existing holding to authorized human review.
- Human-directed reduction review — available if an approved limit is breached, but the agent does not determine reduction size or execute.
The scenario split—0.25 roadmap-clears, 0.45 products-arrive-but-economics-mixed, and 0.30 delay-or-funding-strain—is an explicit analyst assumption, not an empirical reference-class forecast or a valuation input. The branch logic depends on technical validation, execution, segment conversion, manufacturing and capital, and funding assumptions that remain unverified. assumption.amd.cutoff.scenario-weights assumption.amd.cutoff.technical-validation assumption.amd.cutoff.roadmap-execution assumption.amd.cutoff.segment-conversion assumption.amd.cutoff.manufacturing-capital assumption.amd.cutoff.funding-resilience
Strongest disconfirmers and reversal conditions
Evidence opposing deferment includes Polaris sampling, the THATIC terms, approximate ATMP proceeds, and the issuer's expected capital-expenditure relief. Evidence supporting deferment includes reported losses, negative operating cash flow, ratings pressure, supplier obligations, missing independent validation, unisolated product economics, and incomplete valuation inputs. The conflicts remain disclosed and unresolved rather than averaged away. conflict.amd.cutoff.roadmap-versus-validation conflict.amd.cutoff.product-versus-reported-segment conflict.amd.cutoff.proceeds-versus-funding-strain
Reconsider initiating or adding only when every gate has complete evidence under stable definitions and an authorized human approves valuation and sizing. Escalate a possible reduction review when a roadmap, manufacturing, or funding measure breaches an approved limit without a credible funded remedy. Re-underwrite whenever segment definitions, security terms, capitalization, supply commitments, or forecast assumptions change materially. judgment.amd.cutoff.defer-reunderwrite judgment.amd.cutoff.six-gates judgment.amd.cutoff.valuation-abstention
The central transferable lesson is procedural: preserve the difference between a roadmap, a sample, qualified volume, customer adoption, product economics, financing resources, and valuation readiness. Each is a different proposition with different evidence and a different failure mode.