[Research Brief] April 26, 2026 — NVDA Reclaims $5T Cap Amid Trump-Driven Tape Volatility; Yield Curve Steepens

🧭 MACRO CONTEXT

Risk assets remain firm into the weekend with the S&P 500 at 7,165.08 and VIX at 19.31 (source: FRED), suggesting a complacent but not euphoric tape. The yield curve has un-inverted meaningfully — 10Y at 4.34% vs. 2Y at 3.83%, a +51bp positive slope (source: FRED) — historically a constructive setup for cyclicals and financials post-easing cycles, though I'd caution this often reflects late-cycle term premium re-pricing rather than pure growth optimism. Fed funds at 3.64% (source: FRED) implies the market has already digested most of the cutting cycle; further duration tailwinds for long-duration tech (NVDA, MSFT, NET, DDOG, ADBE in our book) are likely capped from here. HY credit spreads at 286bp (source: FRED) remain benign — no stress signal.

The dominant overnight tape driver is NVIDIA closing at a record, market cap back above $5T (source: Yahoo Finance). This is directly supportive of our NVDA thesis (Conviction 8/10, RECOMMEND) and indirectly supportive of MSFT, CEG (AI power demand), and BWXT (nuclear-for-AI narrative — note Pentagon awards for nuclear reactors on Air Force bases per BWXT news feed). I'd also flag the Bloomberg headline "He Has the Market in a Chokehold': Stocks Swing as Trump Posts" (source: Bloomberg) — confirms ongoing headline-driven volatility regime. This is a process risk, not a thesis risk: short-term swings driven by Truth Social posts are explicitly excluded as a basis for thesis change per our hard rules.

Two macro data caveats: GDP growth and real GDP both null in today's FRED pull — we are flying without a current quarter print. CPI YoY shows 330.293 (this is the index level, not the YoY rate — likely a feed labeling issue; flagging for data hygiene, do not use as an inflation signal today). Other macro headlines (Roundup at SCOTUS, Spirit Airlines bailout discussions, crypto at Mar-a-Lago) are noise for our portfolio.

📊 PORTFOLIO THESIS UPDATE

NVDA (RECOMMEND, 8/10) — Thesis Reinforced. Record close pushing market cap >$5T (source: Yahoo Finance). News flow includes Oklo/LANL partnership for nuclear-powered AI factories — supports the durable compute demand thesis and creates a secondary linkage to BWXT and CEG. 1M +21.62%. No conviction change — but I am explicitly NOT raising conviction on price action alone. Watching Q2 earnings cadence and any signs of hyperscaler capex digestion.

MSFT (RECOMMEND, 8/10) — Stable. 1M +16.03%, 1W +1.57%. UK cloud licensing lawsuit (source: news feed) is a tail risk worth tracking but historically these regulatory matters resolve in low-single-digit % revenue impact. No change.

MELI (RECOMMEND, 7/10) — Stable. No new news. 1M +12.52% confirms LatAm e-commerce/fintech narrative continues to work. No change.

LLY (MONITORING, 7/10) — Stable. 1M -1.45%. Adjacent longevity/cellular rejuvenation news is interesting peripheral context but not material to GLP-1 thesis. No change.

ADBE (MONITORING, 7/10) — Headline noise, no fundamental shift. "Adobe Will Collapse Because of AI" headlines (source: news feed) reflect the consensus bear narrative I already addressed in the prior thesis. Until ARPU or net-add data deteriorates, this is sentiment, not signal. No change.

KNSL (MONITORING, 7/10) — Stable. 1W -3.63% in line with broader specialty insurance pullback. No change.

MEDP (MONITORING, 5/10) — Conviction Cut Watch. 1W -22.31%, 1M -12.01%. Multiple securities class action announcements (Schall, Rosen, Robbins Geller per news feed). This is the most actionable negative development in the book. The flagged red flags in the initiation thesis are now materializing. Lowering conviction to 4/10, status remains MONITORING — leaning toward removal from watchlist pending review of underlying allegations. Will not recommend until litigation overhang clarifies.

KTOS (MONITORING, 5/10) — Drawdown Continues. 1M -19.25%. No fundamental data to support thesis change in either direction. Holding at 5/10.

RYAN (MONITORING, 6/10) — 1W -8.66%. No company-specific news driving the move. Watching but not adjusting.

FOUR (MONITORING, 6/10) — 1W -8.17%, 1M -6.44%. Payments group weakness, no FOUR-specific news. Holding.

CSGP (MONITORING, 6/10) — 1M -11.77%. BTIG reiterated Buy (source: news feed) — analyst rating noted but not used as basis for action. No fundamental update.

CEG (MONITORING, 6/10) — 1W +9.03%. Likely riding the AI/nuclear power tailwind alongside NVDA news. Watching for inclusion as RECOMMEND if I can validate PPA pipeline economics.

BWXT (MONITORING, 6/10) — Pentagon micro-reactor awards (source: news feed) directly supportive of thesis. Will dig into specifics next week — potential conviction upgrade candidate.

All other positions (ITUB, CELH, CAVA, AAON, EXPO, NET, DOCS, SPSC, PCTY, TTD, DDOG, AXON, COST): no material new data. Status quo.

🔍 NEW RESEARCH / WATCHLIST ADDITIONS

No new initiations today. Pipeline focus for next week:
1. BWXT — re-underwrite given Pentagon micro-reactor news; potential upgrade to RECOMMEND.
2. CEG — validate AI/data-center PPA economics; if I can get conviction on pricing power, candidate for RECOMMEND.
3. MEDP — review securities litigation primary documents to determine if conviction should go to 0 (remove).

Watchlist unchanged: ITUB, CELH, CAVA, RYAN, FOUR, ADBE, MEDP, AAON, EXPO, KNSL.

⚠️ RISKS & RED FLAGS

1. MEDP litigation overhang — escalating. Three separate securities class action announcements in one news cycle (source: news feed). This is a flashing yellow light. Need to read the underlying complaints and 10-Q to determine if there is restatement risk vs. typical opportunistic plaintiff-bar activity post-drawdown.
2. NVDA concentration / valuation risk. $5T market cap (source: Yahoo Finance) is now ~7%+ of the entire S&P. Any whisper of hyperscaler capex moderation will be violent. Thesis remains intact, but position sizing discipline matters more than ever.
3. Headline-driven volatility regime. Bloomberg flags Trump posts moving the tape (source: Bloomberg). I am holding the line on Hard Rule #7: short-term volatility is not a thesis-change trigger absent fundamental shift.
4. Macro data gap. GDP and real GDP nulls in FRED feed today. CPI YoY field appears to contain index level, not YoY rate — flagging for data ops. Do not act on inflation reads until corrected.
5. Software cohort weakness. "Terrible week in software stocks" (source: DDOG news feed) — DDOG, NET, ADBE, SPSC, PCTY, DOCS all in our book. Watching for any sign of demand softening in Q1 prints; if it's just multiple compression without fundamental break, this is opportunity, not exit signal.

📋 POSITIONS SUMMARY

Active RECOMME