What a $154 stock actually buys you tonight — a governance trial dressed up as an earnings call.
The call — Hold the staged value long, add nothing into the print. Tranche 1 ($148–152) filled and working; new money only on a washout to $138–142 or panic to $118–125. Target $200, 6–12 months, SPECULATIVE.
Why now — Tonight (~4:20 p.m. ET results, 5 p.m. ET call) is the first Nvidia print without Huang in 33 years, with the stock down ~24% this month and options pricing an 8.5% post-print swing (~$335B).
The disagreement — Wall Street says the CEO exit breaks the AI-banker deal machine, so it cut targets 10–20%. My read: the moat (CUDA + systems + backlog) survives the man; the real break-risk is China revenue, not charisma.
The level that changes everything — Q4 revenue guide $111B. Below it, the staged-long thesis is dead and I exit. At or above $115B with FY28 reiteration, the bottom is likely in.
First sentence: judge Colette Kress on three sentences tonight, not on the $109B. Consensus (LSEG) expects EPS of ~$2.47 and revenue of ~$109.1B on company guidance of $108B (±2%), i.e. ~91% growth — and even a clean beat will not move the stock if Kress sounds like a caretaker.
The facts are stark. Jensen Huang resigned effective November 12 after admitting extramarital affairs, one reportedly involving a woman U.S. intelligence believes to be a Chinese agent, plus a second involving a direct report. CFO Colette Kress (CFO since September 2013) is interim CEO and CFO, Harvey C. Jones is the board's first-ever chair, and the company's own transition release promises operations and roadmap "continue without interruption." The board has hired Wachtell Lipton to lead a probe, is not considering bringing Huang back per CNBC sourcing, and Huang is slated to appear before House panels December 1 with Nvidia's lawyer testifying December 2.
What that means for pricing: more than 20 analysts tracked by FactSet cut targets in nine days — Mizuho to $220 from $250 ("6–7 month management overhang"), Morgan Stanley to $267 from $300 with calendar 2027–28 data-center estimates cut ~10%, Evercore slashing far deeper — yet nearly all kept Buy ratings. Morningstar cut fair value 19% to $250 from $310, kept a wide moat and 4 stars, and called shares "moderately undervalued." Translation: the Street still believes the silicon, it just no longer trusts the multiple without the dealmaker.
My opinion: that is the correct discount applied to the wrong risk. Kress does not need to be Huang the visionary tonight. She needs to be Kress the allocator — confirm the $92B / $87B / $88B supply commitments convert to shipments, confirm hyperscaler acceptance of Vera Rubin, and name a permanent-CEO search with guardrails on neocloud guarantees. If she does that in the first ten minutes, the stock can rally on a mere in-line guide. If she reads a compliance statement, no beat saves it.
First sentence: Q3 is already printed in the supply chain — Q4 guidance is the entire trade. The Oanda preview frames it precisely: revenue of $106–108B would sit inside management's $108B ±2% band and still disappoint, because consensus sits at the top of the range at $108.8–109B with EPS of $2.47–2.48.
Run the arithmetic a numerate reader should demand. At my 11 a.m. ET check Nvidia traded ~$153.70 (day range $150.33–157.46, down ~1.6%), putting market cap near $3.7T. A $109B quarter at ~74% gross margin (company guide: 74% ±50bp, ~$9B non-GAAP opex) is ~$80B of gross profit — the machine works. But free cash flow is the tell the bulls skip: Q2 GAAP net income was $59.7B against operating cash flow of only $24.1B and free cash flow of $21.3B, with ~$7.8B of pre-tax equity-investment gains flattering earnings. When revenue grows 106% and FCF grows 59%, the gap lives in receivables, inventory and advance payments — exactly where neocloud vendor financing would hide.
So grade tonight on my pre-set ruler (from open calls DAILY-NVDA-1112/1116): bear needs Q4 guide below $111B — that kills the staged long and I exit; base is $111–114B with FY28's ~70% growth "supply-constrained outlook" reaffirmed (stock chops $150–175); bull needs Q4 at or above $115B plus FY28 reiteration (stock reclaims $180–190 fast). Probability weights into the print: bear 30%, base 50%, bull 20% — because Kress inherits Huang's commitments but not his ability to conjure demand with a phone call, and Morgan Stanley already trimmed forward data-center sales 10% for that reason.
First sentence: up to one-fifth of Nvidia's compute revenue may be China-adjacent, and Washington closed the rental loophole yesterday. This is the input that flips my whole model if I am wrong, so state it plainly.
The chain: Nvidia insists it gets "essentially no data-center sales from China." Culper Research alleges ~20% of fiscal 2026 compute revenue was "driven by China" via Singapore-area intermediaries. On Monday the Commerce Department issued an interim final rule blocking Chinese firms from renting advanced chips through overseas data centers — the exact Megaspeed pattern. Justice and Commerce confirmed a joint probe of Megaspeed (probing since 2024; Nvidia put it on "no sell" in June 2025), and the board today stood up a National Security and Export Control Oversight Committee reporting quarterly.
Consequence math: 20% of a ~$194B FY26 data-center base is ~$39B of at-risk-adjacent revenue. Even if only half is truly exposed and only half of that is lost, that is ~$10B of annual revenue — roughly 2.5 points of FY28 growth — gone at 75%+ margin. That alone justifies Morningstar's $60 fair-value cut far more than any "leadership premium." For readers: above $115B Q4 guide, management is telling you the rental ban does not bite; below $111B, it is telling you it does. Listen to the guide, not the prepared remarks on "no evidence of diversion."
First sentence: CUDA, networking and the $267B three-year supply book are institutional assets, not founder magic. Data-center revenue grew from $47.5B (FY24) to $193.7B (FY26), ~94% of total growth and ~90% of sales; networking alone did $31.4B in FY26 (+141%). Morningstar's wide-moat case — GPU hardware plus switching costs around CUDA plus a full AI-systems stack — is why it still models $385B of FY27 data-center revenue and $672B the year after, with mid-70s gross margins. I share that direction but not that slope: my working model needs only ~$340–350B FY27 DC revenue to justify $200, because at 19.5× trailing earnings ($153.69 / $7.91 TTM EPS) the stock already prices a growth scare.
Kress is the right custodian for exactly this moment, and the market misreads her. Thirteen years as CFO through revenue compounding from $4.1B (FY14) to $215.9B (FY26), 25 years at Cisco/Microsoft/Texas Instruments before that, lead negotiator on Mellanox — this is the executive who checked Huang's bets, per Reuters' profile, not the one who dreamed them. An interim CFO-CEO cannot replicate Huang's neocloud deal-juggling (the $500B financing push, the equity-plus-purchase-commitment loops with CoreWeave/Nebius/OpenAI), but she can do the thing that actually protects equity now: slow the off-balance-sheet creativity, collect receivables, and let hyperscaler capex (still accelerating per every Q3 print) carry the guide. For a tech-native reader: underwrite the balance sheet, not the keynote.
Downstream consequences I would position around: (1) neocloud equity (CoreWeave at ~$48, down from my $54–62 buy zone — tranche 1 underwater, no second half until a daily hold above $54) stays uninvestable until Kress explicitly reaffirms guarantees; (2) memory and optical (Micron, Lumentum/Coherent — the latter two jumping today on the China equipment-curb news) inherit the AI-capex bid if Nvidia's multiple stays capped; (3) Anthropic's Thursday IPO (range $70–78, New Street Buy at $105) becomes the clean sentiment read on AI demand if Nvidia's print is muddied by governance — watch its pricing-night demand as your cross-check on my bull case.
First sentence: this is a WATCH-into-the-print on top of a SPECULATIVE staged position — not a new buy. Open-call honesty first: DAILY-NVDA-1112 tranche 1 ($148–152) filled (low $148.80 Nov 12; high $159.89 Nov 13) and sits in a small profit at ~$154; DAILY-NVDA-1116 said no chase at $159 into tonight; DAILY-HIKE-1113 short QQQ $672–688 is working (QQQ ~$653 intraday). Tonight either validates or kills the equity leg of that book.
| Flagship | NVDA staged value long (existing — no new buying tonight) |
|---|---|
| Entry | T1 $148–152 FILLED · T2 $138–142 washout only · T3 $118–125 panic/leadership-only; average ~$137 if all fill; HALF size, survivable to ~$95 |
| Target | $200 (Morningstar $250 is the stretch frame); trim half at $180–190 on a $115B+ bull guide |
| Invalidation | Q4 revenue guide below $111B — exit the trading long; also exit on charges vs. Nvidia the company, interim mgmt pulling neocloud guarantees with customer cuts |
| Timeframe | 6–12 months |
| Conviction | SPECULATIVE (one signal: price vs. moat value; unconfirmed until guide + receivables print) |
What each reader does tomorrow: holders of T1 do nothing tonight — your stop is the guide, not the after-hours wick. Cash sideliners wait for T2/T3 prints; an 8.5% down-move from $154 is ~$141, i.e. the market will hand you T2 on a soft guide if you are patient. QQQ-short holders keep the hedge through the Anthropic pricing Wednesday night; a disorderly NVDA break re-tests S&P 7,300, at which point my book rule halves all longs. No options lottery: implied 8.5% both ways with a governance overhang is fairly priced — edge is in the staged equity, not the straddle.
First sentence: three measurable conditions, any one of which retires this thesis. (1) Q4 guide below $111B or FY28 growth walked down from ~70% — demand thesis broken, I exit and say so. (2) Q3 receivables up >30% quarter-on-quarter or auditor-flagged related-party exposure above $10B — the cash-conversion story I am underwriting is false. (3) Interim management pulls neocloud purchase guarantees and CoreWeave/Nebius cut forward capex within 30 days — the financing loop I assume merely pauses is actually unwinding. Conversely, the bull I am under-positioned for: Q4 guide ≥$115B with FY28 reiteration and a named permanent CEO who is not a politician's pick — that would force me to upgrade conviction to HIGH and pull T2 up to $148.
Opinion disclosure: staged-long assessment is the author's judgment under uncertainty; China-exposure and CEO-search outcomes are unconfirmed as of publication. Not investment advice.
Prior calls referenced: DAILY-NVDA-1112 (staged long, T1 filled), DAILY-NVDA-1116 (no chase), DAILY-HIKE-1113 (QQQ short working), DAILY-CRWV-1111 (underwater, awaiting $54), DAILY-BTC-1110 (closed flat — acknowledged in this morning's crypto daily).
Chinese edition: 中文版 (same data, same levels)