Everyone won this week. That's the problem.

…all in a FREE Weekly newsletter.

AI & robotics investing — explained so you can actually act on it.

ISSUE 10 // Friday, 15 May 2026

THE HOOK

Nvidia added $900 billion in market cap this week and is three trading days from a binary earnings event. Every chip name moved with it. Every cap-ex play. Every China name. The S&P broke 7,500 for the first time. Everyone won this week. That's the problem — because reflation into a binary print is the part of the cycle nobody wants to remember.

YOUR ONE DECISION THIS WEEKEND

Pick one of three research paths into Wednesday's NVDA print (20 May, 2pm PT):

1. AMAT — the low-catalyst picks-and-axes play. Wins regardless of which chip name beats.

2. BABA / KWEB — the binary China bet, sized to the Trump–Xi summit outcome.

3. Japanese robotics — the watchlist add. Optimus delay gave it a free quarter.

Everything below makes the case. The deliverable: one ticker, one tripwire, by Wednesday ope

WEEKLY SCOREBOARD

Bottom line: every chip, every cap-ex play, every China name moved up. That's not selectivity — that's reflation. Plain English: when everything wins together, the market has spent its surprise budget. Surprises into Wednesday's print now skew downward.

THE THREE STORIES THAT MATTER

1. Nvidia hit $5.7 trillion. The market has spent its surprise budget before Wednesday's print.

(Reported: Bloomberg, 14 May.) NVDA closed Thursday near $236 — up 20% in seven sessions, $900 billion added in a week, market cap $5.7T. The pivot is 20 May, 2pm Pacific. The market is now priced for a clean beat, raised guidance and silence on receivables. What it means: a beat-and-raise has to land outside consensus to push the stock further. A merely-good print can still drag it down 5–10%. Plain English: when a stock has run 20% into earnings, "good" can still be "not good enough."

2. Trump–Xi summit: Huang flew with the delegation. The trip is the signal.

(Reported: CNBC, Euronews, 13–14 May.) Trump arrived in Beijing Wednesday with Jensen Huang, Elon Musk and Tim Cook on the plane. Xi told the US executives China will "open wider." Reportedly on the table (per CNBC/Reuters briefings, not officially confirmed): a formal H200 export licensing framework that flips "presumption of denial" to "presumption of approval" for commercial AI buyers. My read: a paper signed this weekend resurrects a trade left for dead three years. BABA and KWEB reprice Monday. Or not. Both outcomes are tradeable, in opposite directions.

3. AI revenue is no longer a forecast. It's a print.

(Reported: CNBC, Yahoo Finance, 13–14 May.) Cisco raised fiscal-2026 AI revenue guide from $3B to $4B and cut 4,000 jobs to fund the build — stock +15% Wednesday. Applied Materials Q2 the next afternoon: $7.91B revenue (+11.4% YoY), EPS $3.51 vs $2.71 consensus, 30% growth guided for 2026. What it means: the AI ROI question that hovered over the cycle for two years just got two data points hard to argue with — from the picks-and-axes layer, not the headline names. That's why AMAT is Idea 1.

Supporting signals (for portfolio adjustments, not main trades)

● CoreWeave Q1 (reported, CNBC 7 May): revenue +112% to $2.08B and backlog $99.4B — both beat — but stock sold off on light Q2 guide, wider EPS loss, and $31–35B capex with $25B total debt. Validates the Issue 09 named-loser thesis. Oracle reports 11 June — same template, larger balance sheet.

● Robotics layer split (reported, TechTimes 14 May; Tesla earnings call): Figure 03 hit 8-hour autonomous shifts at BMW Spartanburg. Tesla pushed Optimus V3 to late July/August. Same week, opposite trajectories. Shippers (Figure, FANUC, Keyence, Yaskawa) over storytellers. Tilts the Japan trade in Idea 3.

OUR INVESTING ANGLE

Everyone's watching whether NVDA hits $6 trillion. The smarter watch is which of three setups blinks first.

The thesis (my call, calibrate against your own read): the run-up into Wednesday has been sentiment-and-flow-driven, not earnings-driven. NVDA hasn't printed. The Beijing summit hasn't concluded. Nexus hasn't been restructured. The market is pre-pricing three favourable resolutions simultaneously. That's the reflation tell.

The asymmetry: a clean NVDA beat plus a signed framework gives single-name longs roughly 5–8% upside from here. A miss or a summit non-event costs ~12–15%. The risk-adjusted setup is to own the asymmetry, not the catalyst — AMAT (wins regardless), BABA (binary upside, bounded downside), Japan (no catalyst dependency at all).

Who gets hurt: single-name NVDA longs piled in this week. CoreWeave — already showed the leverage problem in the print. Oracle's debt-funded compute backlog — same vulnerability, 11 June reports. PLTR at ~46x sales — the multiple, not the fundamentals, is now the risk.

→ Deep dive: the full payoff matrix (8 joint outcomes, 4 priced), scenario probabilities for the summit, and position-sizing logic for each of the three paths.

Read it →

WHAT COULD GO WRONG? (the bear case)

1. NVDA beats and rallies through $250. Receivables shrink, margins hold, China added back. The asymmetry trade misses the upside.

2. Trump–Xi produces nothing concrete. No formal framework. BABA + KWEB unwind Monday. Size the China side assuming this is the modal case.

3. Macro shock — oil, Iran, Taiwan. WTI at $96 already. VIX above 22 changes the math on every chip name. Risk-off doesn't ask whether your thesis is right.

Size for the possibility that the priced-in good outcome lands without the upside you bought it for.

YOUR THREE RESEARCH PATHS

Not recommendations — starting points. Pick one. Tighten it before Wednesday.

Path 1 — Applied Materials (AMAT). The picks-and-axes name to research this weekend.

Why now: Q2 reported Thursday (reported, AMAT IR): $7.91B revenue (+11.4%), EPS $3.51 vs $2.71 consensus, 30% growth guided for 2026.

Thesis: Every wafer for every AI chip first sits on an AMAT etch tool. At ~22x forward earnings vs LRCX at 24x and chip names at 30x+, this is the cheapest exposure to the $725B AI capex cycle.

Risk: About 25% of revenue is China. Summit cuts both ways.

Tripwire: Q3 guide above $9.5B at the August earnings call = capex super-cycle extends through 2027. Below $8.5B = the cycle is rolling over.

How to research: Ticker AMAT. Compare LRCX, KLAC on equipment-cycle leverage. SMH/SOXX for diversified.

Path 2 — Alibaba (BABA). The binary China play, sized to the summit outcome.

Why now: I keep coming back to this one. Trump–Xi summit reportedly negotiating an H200 export framework. Even partial liberalisation re-rates the whole China AI complex.

Thesis: BABA at ~12x forward (vs US AI peers at 25–30x). Owns the cloud, the model (Qwen), and a 38%-margin commerce business that funds the buildout. The "uninvestable" discount collapses fast when the structural reason disappears.

Risk: Summit produces nothing concrete. Taiwan re-flare.

Tripwire: A formal H200 licensing framework announcement before Monday US open = trade is real, Monday opens up 8–15%. No announcement by Wednesday US open = unwind back to mid-week levels.

How to research: BABA directly, or KWEB (KraneShares China Internet ETF) for diversified exposure.

Path 3 — Japanese industrial robotics. The watchlist add that just got a free quarter.

Why now: Tesla pushed Optimus V3 reveal to late July/August (reported, Tesla earnings call), with external sales slipping to late 2026. Fresh free quarter for the names already shipping.

Thesis: 28x forward for proven industrial cash flows vs ~58x for US names still months from serial-numbered shipments. Same secular trend, half the multiple.

Risk: Tesla actually delivers in Q3. Japan goes nowhere.

Tripwire: Confirmed Optimus serial-numbered shipments leaving Fremont by 31 August = Japan trade pauses. No Tesla shipments by end-September = Japan re-rates.

How to research: BOTZ (43% industrials, heavy Japan). FANUC (6954.T), Keyence (6861.T) directly.

AI INVESTMENT FRAMEWORK

Living portfolio framework. Not financial advice — research starting points only.

Changes this week

● Global raised DEVELOPING ↔ → DEVELOPING ↑↑. H200 framework is the asymmetric catalyst. Sized as binary.

● Physical AI held DEVELOPING ↑. Sharpened: shippers (Figure, Japanese names) over storytellers (Tesla).

● Infrastructure held HIGH ↑↑ — but trim into the print, don't add.

Next two dates that matter

Disclaimer: This newsletter is for informational and educational purposes only and does not constitute financial advice. iPrompt Signals is not a registered investment advisor. Always conduct your own research and consult a qualified financial professional before making investment decisions.

YOUR MOVE

Pick one of the three research paths from the top of this issue. 45 minutes. Latest 10-Q. Your own tripwire. Hit reply and tell me which one you're digging into.

Two reminders before Wednesday:

● A beat is priced; a miss isn't. Size accordingly.

● Have a plan for both summit outcomes before Monday open. Both are tradeable, in opposite directions.

→ Deep dive: the full payoff matrix (8 joint outcomes, 4 priced), scenario probabilities for the summit, and position-sizing logic for each of the three paths.

Read it →

Stay curious — and stay qualified.

— R. Lauritsen

Editor, iPrompt Signals

Know someone building an AI position? Forward this — they'll thank you by Friday.

P.S. — Last Friday I called the custom-silicon trade "delayed, not dead." Six days later AVGO is up 5.5% with a $545 Wells Fargo PT. So "delayed" might actually mean "priced in faster than I thought." The market's discount rate on AI ROI has compressed from months to weeks. Size your conviction accordingly.

TERMS USED THIS WEEK

Binary event — A print with only two believable outcomes. NVDA on Wed 20 May is one.

H200 — Nvidia's slightly throttled GPU built for the Chinese export market. The chip under negotiation at the Trump–Xi summit.

Picks and axes — A side play. AMAT, LRCX, ASML are the AI shovels — sell to every chip name regardless of which one wins.

Receivables — Money customers owe but haven't paid. Above $40B at NVDA's 20 May print = Burry's stress signal flashes.

Reflation — Everything rising together. Selectivity (winners + losers) is bullish; reflation (everyone wins) is late-cycle.