🧭 MACRO SNAPSHOT

The macro backdrop is increasingly bifurcated. Fed funds sits at 3.63% (FRED) with the 2Y at 4.05% — the curve at the front end still pricing a Fed that is reluctant to cut faster, and CPI YoY running hot (the 332.4 print is the index level; trajectory remains above the 2% target with energy pressure from the euro zone's 3.2% print tied to Iran/oil per CNBC). Real GDP growth at 1.6% confirms we're in a "slow but not breaking" regime — high-yield credit spreads at 272 bps are notably benign, signaling no credit stress despite the geopolitical overhang. The 10Y print is unavailable today, which limits real-rate visibility.

What matters for the book: the AI capital arms race has shifted from a story about chip demand to a story about capital itself. Alphabet's $80B equity raise — with Berkshire taking $10B (Barron's) — is the cleanest signal yet that compute scarcity has become capital scarcity. Stratechery's framing of "The Google Capital Company" is the right lens: in this regime, balance-sheet-rich incumbents ($GOOGL, $MSFT, $BRK-B) and infrastructure beneficiaries ($AVGO, $TSM, $ANET) are the structural winners; pure-play hardware narratives ($NVDA's consumer AI PC) are starting to fray at the edges per Ben Thompson. Manufacturing recovery remains "muddled" (Axios), which keeps a soft ceiling on industrial cyclicals.

₿ BITCOIN DAILY WRITE-UP

Price & Market Structure: BTC trades at $67,013, down 3.56% on the day, -11.59% on the week, and -15.08% over 30 days. We're now ~46.8% off the $126,080 ATH and printing negative YoY returns — the first time this cycle. BTC dominance at 55.9% suggests this is broad crypto risk-off, not just BTC-specific.

Structural Thesis: BTC is held as a long-horizon allocation to a programmatically scarce asset with structurally growing institutional demand (ETFs, treasuries, advisor platforms). The thesis does not require near-term price confirmation — it requires the supply/demand math to remain intact, which it does.

What Happened This Week: Two material developments: (1) Mt. Gox coin movements and ETF outflows have pressured price into the mid-$67Ks (Bitcoin Magazine), and (2) US-Iran military strikes have made BTC trade as a risk asset, not a hedge — a recurring failure mode for the "digital gold" narrative under acute geopolitical stress. Sanders/Warren are also lobbying the Labor Department to abandon the Bitcoin 401(k) rule — a regulatory headwind worth tracking but not yet decisive.

Bull / Bear Scorecard:

Bull:
- ETF wrapper demand is structural and compounds across advisor allocation cycles even when flows are choppy
- Post-halving supply math (~450 BTC/day issuance) remains mathematically tight against institutional demand sinks
- 95.4% of terminal supply already mined — inelasticity grows each cycle

Bear:
- Cycle pattern divergence is now material — month 13-14 post-halving should be parabolic upside, not -45% drawdown
- Macro is actively hostile: hot CPI delays Fed cuts, geopolitics pressures risk assets, BTC correlating with risk not hedging it
- Negative YoY return is a structural signal — not just noise

Conviction Check: Action: HOLD | Conviction: 6/10. Unchanged. The structural thesis is intact but the cycle divergence is real and unexplained — I am not adding here, and I am not selling on macro-driven weakness.

What to Watch:
- ETF net flows turning positive for 10+ consecutive sessions (would re-confirm institutional bid)
- US-Iran de-escalation OR a Fed pivot signal — either would relieve macro pressure on risk
- A decisive break below $60K (which would invalidate the post-halving uptrend technically) or a recovery above $85K (which would re-confirm the cycle)

Community Pulse: Reddit is quiet today — no top threads fetched, which itself reflects a fatigued/uncertain community after the drawdown. The Bitcoin Layer's "Inflation Boogeyman" and "Everything Everywhere All at Once" frames this correctly: macro is the proximate driver, with the Strait of Hormuz, hot ISM/CPI, and the Anthropic IPO competing for capital all weighing on BTC. Bitcoin Magazine's note on $MSTR and other BTC-proxy stocks taking a beating suggests leveraged BTC exposure is being unwound — sentiment is cautious-to-fearful, not yet capitulatory.

🔬 TODAY'S DEEP DIVES

AMD — Advanced Micro Devices, Inc. — ROLLING REVIEW
Conviction: 5/10 | Status: WATCHLIST | Sector: Semiconductors

WHAT THEY DO: AMD designs CPUs, GPUs, and adaptive computing chips (FPGAs, via the Xilinx acquisition). Revenue comes from three customer buckets: (1) hyperscalers buying EPYC server CPUs and MI-series AI accelerators (Microsoft Azure, Meta, Oracle, Google); (2) PC OEMs buying Ryzen client CPUs (Dell, HP, Lenovo); and (3) embedded customers in telecom, automotive, aerospace, and industrial.

WHY IT'S INTERESTING NOW: $AMD is at $521.54 with an ~$850B market cap — meaning the market is now valuing it as a serious AI accelerator contender, not just a CPU player. The question is whether MI300X/MI325 traction at hyperscalers is real and durable, or whether $NVDA's CUDA moat continues to hold despite AMD's price/perf advantage on paper. The May 10 thesis at 7/10 has been downgraded to 5/10 — a meaningful conviction drop that requires explanation.

BULL CASE:
- Genuine second-source dynamic at hyperscalers: every major cloud is publicly committed to multi-vendor accelerator strategy, and AMD is the credible #2
- EPYC server CPU share gains continue to compound — this is the cash cow funding the AI accelerator R&D ramp
- Embedded/Xilinx segment provides counter-cyclical revenue and exposure to defense/aerospace tailwinds independent of the AI cycle

BEAR CASE:
- CUDA software moat at $NVDA remains the dominant gating factor — performance benchmarks are necessary but not sufficient for share gains
- Valuation at $850B already prices substantial AI accelerator success that is not yet visible in the financials
- New $NVDA RTX Spark / consumer chip moves compress AMD's traditional client GPU/CPU pricing power
- The downgrade from 7→5 reflects: deteriorating relative narrative vs. $NVDA, growing custom-silicon competition from $AVGO and hyperscaler in-house designs, and a stock that has run hard without commensurate AI accelerator revenue confirmation

KEY METRICS: Bull case requires MI-series to hit a $10B+ run rate; current valuation implies ~30x forward earnings on a normalized basis assuming AI ramps. Differentiator vs. $NVDA: open-source ROCm software stack and pricing flexibility; differentiator vs. $AVGO: programmability vs. custom ASICs.

BOTTOM LINE: AMD belongs on the watchlist at 5/10 — interesting business, defensible CPU franchise, but the AI accelerator narrative needs revenue proof before conviction can move back up.

Note: Only one deep review on the docket today — no new ideas screened. This is appropriate given the macro uncertainty; I'd rather wait for cleaner setups than force initiations.

📋 TARGET LIST STATUS

No conviction changes today beyond the AMD rolling review (7→5, completed earlier in the week). No new names displaced existing positions — the screen returned no high-conviction setups today, which is itself useful information.

💼 YOUR PORTFOLIO

• $GOOGL | STRONG HOLD | 9/10 — Core long-horizon position. The $80B equity raise with Berkshire participation validates the "capital is the new moat" thesis. Buffett's stamp of approval matters.
• $AVGO | STRONG HOLD | 8/10 — Custom silicon franchise is the structural winner of the hyperscaler diversification trend. New broadband chip with AI capabilities extends the moat. Aschenbrenner's bearish bet noted but not actionable without more data.
• $BABA | STRONG HOLD | 8/10 — Down 9.9% in 3 weeks, -34.5% from 52W high. Position cost basis is favorable; the China tech price-war directive ("stop fighting on price, invest in AI") is actually constructive for $BABA's margin profile.
• $ISRG | BUY MORE | 8/10 — Near 52-week low at $423.63 (-23.3% 1Y). This is exactly the kind of high-quality compounder mispricing the long-horizon mandate is built to exploit.
• $MKL | HOLD | 7/10 — Holding gain ~+103%. No reason to act on price; underlying insurance franchise intact.
• $MP | HOLD | 7/10 — Rare earth thesis intact; selective trim opportunity above $70.
• $UNH | STRONG HOLD | 7/10 — Bernstein reiterating Outperform on PBM transparency; recovery thesis playing out.
• $AVAV | HOLD | 7/10 — Ignore the securities class action noise — these are standard ambulance-chaser filings. Defense unmanned tailwind intact.
• $FSLR | TRIM | 6/10 — Up +37.9% in 2 weeks, +59.2% MoM. Take some chips off the table; momentum has outrun fundamentals.
• $SYM | HOLD | 6/10 — Revenue beat, software/automation growth confirmed; conviction stable.
• $AAPL | HOLD | 5/10 — WWDC 2026 next week is the catalyst; conviction has drifted lower as AI narrative gaps remain unresolved.
• $TSLA | TRIM | 4/10 — Position is +5.5% above cost basis; AI/robotaxi narrative remains unconfirmed in financials. Take size down.

⚠️ WATCH LIST

• $MRVL — Up 46% in a week, 77% in a month on AI chip demand. Either this is a step-function rerating to be added on conviction, or it's overheated. Need Q1 print and customer concentration data before moving from 6/10.
• $NOW — Up 25% in a week, 38.8% MoM. ServiceNow has fundamental AI tailwinds, but the velocity demands a fundamental refresh before acting.
• $PANW — Up 19.6% / 61% in a month on the Portkey acquisition and NATO contract. Cybersecurity AI gateway thesis getting validated — candidate for upgrade if the next earnings print confirms ARR acceleration.
• $FTNT — Up 16.4% / 66.8% MoM. Same setup as PANW — watch for Q2 confirmation before upgrade.
• $RKLB — Down 17.9% on the week; SpaceX IPO and Blue Origin explosion are both noise. Fundamental thesis intact at 5/10; would need a successful Neutron milestone to consider upgrade.

🔁 RE-REVIEW QUEUE

Eight names are due for re-review today:

• $CEG | Was 6/10 | Dropped 2026-05-13 | Dropped because conviction fell below threshold for a 50-name slot. Power-demand thesis from AI data centers has only strengthened since — worth a fresh look given $VST and $GEV peer strength.
• $DDOG | Was 6/10 | Dropped 2026-05-07 | Observability/monitoring SaaS displaced by higher-conviction names. AI-driven log volume growth is real; merits a refresh.
• $TTD | Was 6/10 | Dropped 2026-05-07 | Ad-tech leader; dropped on valuation/competitive concerns. Walmart/Amazon DSP competition is the key question to re-examine.
• $BWXT | Was 6/10 | Dropped 2026-05-07 | Nuclear/naval propulsion exposure. With AI power-demand narrative intensifying and uranium ($UUUU) remaining on the list, $BWXT deserves a fresh dive.
• $NET | Was 6/10 | Dropped 2026-05-07 | Cloudflare's edge/AI inference positioning has gotten more interesting as the AI workload distribution debate evolves.
• $KTOS | Was 6/10 | Dropped 2026-05-13 | Defense drone/unmanned systems. $AVAV peer; would re-examine if defense unmanned procurement trends continue.
• $CRWD | Was 6/10 | Dropped 2026-05-08 | CrowdStrike's AI-native security positioning relevant given $PANW/$FTNT strength.
• $NTRA | Was 6/10 | Dropped 2026-05-09 | Genetic testing; less obvious catalyst — lower re-review priority.

To run a fresh dive on any of these, ask Meridian in the chat.