Phase Space AI

Trade Construction

Advanced Micro Devices [AMD]

Advanced Micro Devices, Inc. [AMD] — Trade Construction

Spot $454.54 (2026-07-28) · 12-month target $325 as-reported / $272 fully diluted · Valuation Criteria FAIL

The memo issues no position verdict. The Valuation Criteria is BINDING on the long-only absolute-return strategy and it FAILS here, so what follows is written for two consumers: the strategy that will therefore not initiate a long, and the relative-value fork for which a FAIL plus a MEASURED Short Mechanism score is a live input.

---

Liquidity Criteria — chain pulled live

criteria.md: "Any proposed options structure requires the actual chain pulled first." Pulled from Alpaca 2026-07-28/29.

15-Jan-2027 calls (171 days), IV 73–75%

16-Oct-2026 calls (80 days), IV 76–79%

Liquidity Criteria: PASS. Four-figure open interest across the January-2027 chain with deep two-sided size; the $500 strike quotes $0.53 wide on a $61 premium (0.9%). $741bn market cap. Size is not a constraint.

But note the volatility. Implied volatility of 73–75% on a $741bn company means the January-2027 at-the-money straddle prices roughly a ±37% move. Long premium here is expensive in absolute terms: the $500 call costs 13.5% of notional for 171 days.

---

Why no long is constructed

Three independent reasons, each sufficient on its own:

1. The implied path requires a 58.1% five-year revenue CAGR — 42.8pp above the corrected trailing record, 20.3pp above the best quarter AMD has printed in three years, implying $370bn of FY2031 revenue. No cell in the exit-multiple sensitivity produces a PASS, including at twice the anchor.
2. The 12-month target is 28.5% below spot on the same method that produces +58.5% for AVGO. The multiple is at the 95.9th percentile of AMD's own history.
3. Upside is structurally capped. 320m warrant shares at $0.01 vest on Instinct GPU purchase milestones and share-price targets. Delivering the required path transfers 19.6% of the equity to OpenAI and Meta. The bull case pays for itself out of the shareholders' side of the ledger.

Within the cluster, the same capital buys AVGO — where the mechanism is a $164.6bn firmly committed, non-cancellable order book rather than an "up to 6 gigawatts" announcement, and where the 12-month target is above spot. On slot competition AMD loses to AVGO on the evidence, not on the story.

---

What the Short Mechanism Criteria says — and it says no

criteria.md defines the Short Mechanism Criteria as "decelerating growth and exhausted margin runway (operating margin already high and already expanded, incremental leverage visibly spent)."

Short Mechanism Criteria: NOT PRESENT. Expensive is not a short mechanism, and this is the discipline the criterion exists to enforce. A name can simultaneously be too expensive to own and structurally wrong to short; AMD is that name today. Acted on by nothing on the long-only fork in any case.

---

If the book wants exposure to the theme anyway

Stated for completeness, not recommended. Two constructions that express the AI-compute theme while respecting the Valuation FAIL:

A. Relative value — long AVGO, short AMD, sized on EV/Sales own-history percentile. AVGO sits at its 80.3rd percentile, AMD at its 95.9th. The spread is the trade: same end-market, same capex budget, same foundry, opposite positions within their own valuation histories. Requires the RV fork and borrow work that this long-only memo does not cover; borrow, short interest and days-to-cover are explicitly not part of the Liquidity Criteria on a long-only book.

B. A defined-risk long with a cap below the bull case. Jan-2027 $500/$640 call spread: long $500 at $61.35, short $640 at $30.89 → net debit $30.46, maximum value $140, maximum payoff 4.6x, breakeven $530.46 (+16.7%). Open interest 4,144 and 1,324; both fillable. This caps at $640, above the +7.5% multiple-holds bull case of $489 — so it is a bet on the multiple expanding past its own all-time high, which is not a thesis this memo supports.

Neither is recommended. They are documented so the decision to decline is legible.

---

Sizing, if overruled

Inverse-volatility, per the framework's interim protection on the Downside Criteria.

AMD sizes at two-thirds of an AVGO position on volatility alone, before the Valuation FAIL is considered.

Cluster constraint, and it binds. AVGO, AMD and MRVL are one exposure. All three are fabless AI silicon designers selling into the same hyperscaler capex budgets; MRVL's own 10-K names AMD and Broadcom as direct competitors. They must be sized as a single cluster limit in portfolio_book.json, not as three independent positions. Adding AMD alongside AVGO buys correlation, not diversification.

---

Monitoring triggers — what would reopen the name

---

Ledger entry

Recorded so the call is falsifiable in both directions. No position is recommended; the target and bear case are logged for scoring regardless.

• Ticker AMD · date 2026-07-29 · spot $454.54
• Valuation Criteria FAIL (required 5y CAGR 58.1%; margin −42.8pp vs corrected trailing, −20.3pp vs latest quarter)
• 12-month target $325 as-reported (−28.5%) / $272 fully diluted (−40.2%); bull $584/$489 (+28.5%/+7.5%); bear $260/$217 (−42.8%/−52.3%)
• Bear-case probability 35% over 24 months; named cause: MI450 ramp slips or OpenAI/Meta volumes land below announced scale while warrant tranches dilute anyway
• No position recommended. Reopen on AI-revenue disclosure, a material RPO, or EV/Sales below 8.58x
• Cluster: fabless AI silicon (with AVGO, MRVL) — size as one exposure