🧭 MACRO SNAPSHOT

The story today is the long end of the curve. The 10Y is at 4.59% and the 30Y has hit its highest level since 2023 (Bloomberg) on renewed inflation concern — CPI is still running hot, the Fed is at 3.64%, and the 2s/10s spread has steepened to ~50bp. Translation: the market is pricing fewer cuts and more term premium, which is a direct headwind to long-duration equities and anything trading on 2027+ free cash flow. That's the cleanest explanation for why $ANET (-15% MoM), $AVAV (-18% MoM), $MP (-14% MoM), and $FCX (-14% MoM) have all bled even as $SPY prints new highs on 40% below-average volume (per community data). The rally is narrow.

Two cross-currents matter for the watchlist: (1) China's investment data is weakening again (Bloomberg) — incrementally negative for $BABA and $PDD, neutral-to-positive for U.S. defense names like $AVAV and $RKLB given Rinehart's $100M defense bet (Bloomberg); and (2) the AI-to-power-grid trade is consolidating fast — NextEra/Dominion's $67B all-stock merger explicitly cites AI data center load as the rationale, which is structurally supportive for $VST, $GEV, $ETN, $CARR, and the optical/networking complex ($GLW, $LITE, $COHR, $CIEN). Credit spreads at 2.8% and VIX at 18 say no stress yet, but the small-UK-lender story (CNBC) is the kind of pebble that bears watching.

₿ BITCOIN DAILY WRITE-UP

Price & Market Structure. BTC is at $76,656, down 0.72% on the day, -5.1% on the week, and +2.2% on the month. We remain ~37% below the $126K ATH printed earlier in this post-halving cycle. Dominance is 58.2%, which historically signals capital still concentrated at the top of the stack rather than rotating to alts — a "risk-off-within-crypto" tell.

Structural Thesis. We are ~13 months past the April 2024 halving, in what has historically been the strongest phase of prior cycles (2013, 2017, 2021). The supply side is structurally tight (~450 BTC/day issuance against $1.59T cap), and the demand side now has a permanent new buyer — spot ETFs — that didn't exist in prior cycles. The reason to hold is that this is the first cycle with institutional plumbing, sovereign interest, and a programmatic supply shock all converging.

What Happened This Week. The single biggest development is the White House signaling that a formal Strategic Bitcoin Reserve announcement is "imminent" — Patrick Witt said "the hard part is done" (Bitcoin Magazine). Separately, Iran launched a BTC-based maritime insurance platform for Strait of Hormuz shipping to bypass SWIFT — another data point in the "nation-state monetary routing" trend. Standard Chartered is absorbing Zodia Custody (Bitcoin Magazine), continuing the institutional consolidation of crypto infrastructure. Price action is weak, but the structural news flow is meaningfully bullish — a divergence worth noting.

Bull / Bear Scorecard.
- Bull: Halving supply pressure at all-time lows; demand impulse meets thinner sell-side.
- Bull: Strategic Bitcoin Reserve is now a known-unknown catalyst with a likely near-term resolution — asymmetric to the upside if announced.
- Bull: Sovereign use cases (Iran maritime insurance) keep delegitimizing the "pure speculation" framing.
- Bear: A 37% drawdown at month 13 post-halving is OFF the historical playbook — prior cycles peaked late in this window, not early.
- Bear: ETF flow data is the swing variable and price action suggests outflows are plausible; if confirmed, the marginal-buyer thesis breaks until reversed.
- Bear: Citi's quantum-computing risk note (Reddit thread today) is a slow-burn structural concern, not a near-term driver, but worth filing.

Conviction Check. Action: STRONG HOLD | Conviction: 8/10. No change. The price weakness is uncomfortable but the structural thesis — institutional plumbing + supply shock + sovereign legitimization — is intact, and the Strategic Reserve announcement is a near-term asymmetric catalyst.

What to Watch.
1. Strategic Bitcoin Reserve announcement — confirms or denies the single biggest near-term catalyst.
2. Spot ETF flow data — if net outflows are sustained for 4+ weeks, marginal-buyer thesis is broken.
3. Break below $70K — would signal the cycle peak ($126K) is in and a multi-quarter bear phase has begun.

Community Pulse. The Bitcoin community is in a notably split mood. The two dominant threads are the Iran maritime insurance story (1,085 upvotes — "geopolitical legitimization") and the White House Strategic Reserve news (367 upvotes — anticipation/skepticism mix). Underneath that, there's clear fatigue showing — threads like "At least you have bitcoins" and "$20k in btc stolen" suggest holders are processing the drawdown. The Bitcoin Layer newsletter is focused on "Are Global Bond Bailouts Next?" — tying BTC to the buckling global bond narrative, which is the right macro frame. Sentiment: cautiously bullish on catalysts, exhausted on price.

🔬 TODAY'S DEEP DIVES

SMAR — Smartsheet Inc. — NEW IDEA (REJECTED)
Conviction: 1/10 | Status: WATCHLIST | Sector: Technology (formerly)

WHAT THEY DO: Smartsheet was a collaborative work-management SaaS platform — think of it as a spreadsheet-first project management tool used by enterprise teams. It monetized through per-seat subscriptions and tiered enterprise plans.

WHY IT'S INTERESTING NOW: It isn't. $SMAR was taken private in January 2025 by Blackstone and Vista Equity Partners at $56.50/share (~$8.4B). The screener pulled an empty data payload because there is no public security to analyze.

BULL CASE: N/A — not a publicly traded equity.

BEAR CASE: N/A — not a publicly traded equity.

KEY METRICS: All data null; ticker is dead from a public-equity standpoint.

BOTTOM LINE: Auto-screener artifact — does not belong on the list and was correctly not added. Flagging the screener logic should suppress delisted/private tickers going forward.

UPST — Upstart Holdings — NEW IDEA
Conviction: 5/10 | Status: MONITORING (added then dropped same day) | Sector: Financials / Fintech

WHAT THEY DO: Upstart runs an AI-driven lending platform that originates unsecured personal loans, auto loans, small-dollar loans, and HELOCs. Critically, they don't primarily hold the loans — they underwrite via a machine-learning model (claiming better risk discrimination than FICO) and route originations to a network of bank and credit-union partners, earning fee income on volume.

WHY IT'S INTERESTING NOW: Two things. First, the forward P/E of 8.2x vs. TTM P/E of 68x (yfinance) implies the sell-side is modeling a sharp earnings inflection — if forward EPS of $3.40 hits, this is a 44% revenue grower at 8x forward earnings. Second, a declining-rate environment (which is the consensus path, though today's long-end action complicates it) would reopen institutional loan-buyer demand and dramatically expand Upstart's funding capacity.

BULL CASE:
- Forward P/E of 8.2 vs. TTM P/E of 68.1 (yfinance) implies dramatic earnings inflection. Combined with 82% gross margins and 44% revenue growth, the setup is extraordinarily cheap IF the forward number is real.
- Secular tailwind from declining rates would re-open institutional loan-buyer demand, expanding funding capacity and origination volume.

BEAR CASE:
- Debt/Equity of 269.8 (yfinance) is alarming for a "platform" company — reflects warehoused loans plus corporate debt. If credit quality deteriorates, equity holders absorb losses well before fee income recovers. This is not a clean SaaS balance sheet.
- Negative FCF of -$310M (yfinance) means they're burning cash even at scale. The forward EPS estimate is sell-side hope, not reality.
- The "AI underwriting moat" is unproven through a full credit cycle — the last downturn forced them to retain loans on balance sheet when buyers stepped away.

KEY METRICS: Revenue growth 44%, gross margin 82%, debt/equity 269.8, TTM P/E 68 vs. forward P/E 8.2, FCF -$310M.

BOTTOM LINE: A real business with a real story, but the balance sheet leverage and unproven cycle resilience make this a 5/10 — it earned a brief watchlist slot then was correctly displaced. Worth re-examining if (a) FCF turns positive sustainably or (b) we get a recession stress-test that they survive.

ANET — Arista Networks — ROLLING REVIEW
Conviction: 7/10 | Status: RECOMMEND | Sector: Technology / Networking

WHAT THEY DO: Arista builds high-performance Ethernet switches and the network operating system (EOS) that runs them, sold primarily to hyperscale cloud operators (Meta, Microsoft historically the biggest customers) and large enterprises. They're the premium alternative to Cisco in data center networking, with a software-defined architecture that runs on commodity merchant silicon (Broadcom, mostly).

WHY IT'S INTERESTING NOW: The stock is down ~15% over the past month while fundamentals haven't changed — a valuation/sentiment de-rate, not a thesis break. More importantly, two strategic narratives are converging in Arista's favor: (1) the AI back-end fabric is shifting from InfiniBand (Nvidia/Mellanox monopoly) toward Ethernet via the Ultra Ethernet Consortium, where Arista is a charter member, and (2) Piper Sandler's 2026-05-16 note explicitly flagged that AI inference workloads are expanding the addressable market beyond training — which matters because inference scales with deployment, not just model training.

BULL CASE:
- AI back-end Ethernet TAM expansion: the industry move away from InfiniBand toward Ethernet for AI fabrics is a multi-year tailwind. Arista's Etherlink portfolio and Jericho3-AI-based platforms are positioned to take share. Piper Sandler's 2026-05-16 inference note materially expands the runway.
- Barclays reiterated Overweight (Yahoo, 2026-05-16) — secondary signal, but the analyst community is staying constructive through the drawdown.
- The de-rate to ~31.8x forward P/E (from ~40x at peak) improves risk/reward without any fundamental deterioration.

BEAR CASE:
- Financials are exceptional — 35.1% revenue growth on $9.7B TTM, 42.7% op margin, 63.5% gross margin, 31.5% ROE, $4.36B FCF, essentially no debt (yfinance / latest 10-Q). The bear case isn't financial; it's customer concentration (Meta + Microsoft = majority of revenue), and the risk that hyperscalers eventually disintermediate via in-house networking silicon (Microsoft is already doing this with SONiC).
- Long-duration tech is exposed if the 10Y keeps grinding higher; today's macro tape is a headwind even if the business is fine.

KEY METRICS: Revenue growth 35.1%, op margin 42.7%, gross margin 63.5%, ROE 31.5%, FCF $4.36B, forward P/E ~31.8x, debt essentially zero.

BOTTOM LINE: Holding at 7/10 RECOMMEND — the 15% drawdown is a re-rating not a thesis break, and the Ethernet-for-AI tailwind is a 3-5 year structural story that the long-horizon mandate is designed to capture.

📋 TARGET LIST STATUS

Changes this week: $UPST was added at 5/10 and then dropped today — it cleared the screener but couldn't displace higher-conviction names. $VRT (6/10) was abandoned earlier in the week for the same reason — conviction floor on the list has tightened. Net: the bar to enter the 50 is now effectively 6/10 minimum, with 7/10 ideas getting any defensible "recommend" status.

💼 YOUR PORTFOLIO

⚠️ WATCH LIST

• $ENPH — Up 32.6% in a week, 46.7% in a month, on the AI data center / battery / EV charging roadmap announcement. This is sentiment-driven; I want to see if the product roadmap converts to revenue before raising conviction from 6/10. Trigger: next earnings print confirming AI data center pipeline orders.
• $FTNT — Up 53% MoM after Q1 earnings; conviction is still 6/10 because valuation has run ahead of the growth deceleration concern flagged earlier. Trigger: any sign of growth re-acceleration would push to 7/10; multiple compression without growth would push to 5/10.
• $PANW — Up 46% MoM with no