Three tests the supercycle has to pass this week.

1. The week at a glance

Monday is quiet. Tuesday adds Home Depot before the bell and starts the consumer read. Wednesday is the binary day. Thursday cleans up with Walmart and an industrial-production update. Friday is the wrap.

2. Test 1 — The consumer (HD Tue / TJX Wed / WMT Thu)

Three retailers, three income brackets, three formats. By Thursday lunchtime we will know whether the “real wages negative + sentiment at record lows + gas eating discretionary” thesis from last week is showing up in real corporate numbers.

Home Depot — Tuesday before the bell

Consensus: $41.5 billion revenue (+4.2% YoY) and $3.42 EPS (−3.9% YoY).⁵ The setup here is binary on big-ticket discretionary. Mortgage rates are still elevated, housing turnover is low, and large home-improvement projects have been suppressed for four consecutive quarters. We are watching two specific things: (1) comp-store sales by ticket size — a deterioration in the >$1,000 average-ticket bucket would confirm last week’s −2% furniture print is a real trend; (2) Pro vs. DIY mix — Pro has been holding up because of project backlog, but if DIY softens dramatically, that’s the consumer tell. EPS guide cut on FY26 would weigh on the whole homebuilder/durables complex.

TJX — Wednesday after the close

The off-price tell. TJX historically benefits when consumers trade down. If TJX prints a beat-and-raise with comp acceleration, that is corroborating evidence that the income consumer is feeling squeezed and migrating into off-price — bad news for full-price specialty retail (Macy’s, Nordstrom) and the broader cyclical consumer trade. The market may not move on TJX in isolation (it lands the same evening as NVIDIA and gets buried), but the read-through to the consumer complex is real.

Walmart — Thursday before the bell

Consensus: $172.5 billion revenue and $0.65 adjusted EPS.⁶ Mass-market consumer read with the income skew baked in. The single most important data point is the food-vs-general-merchandise mix. If grocery comps are running well above GM comps and the gap is widening, customers are trading down into staples and out of discretionary — the cleanest corporate version of the “wage compression hitting the wallet” story. Watch also for any management commentary on tariff passthrough, which the April CPI components show is starting to bleed through.

What confirms our thesis vs. what breaks it

Confirms: HD guide cut on FY26 EPS, TJX comp acceleration, Walmart food>GM gap widening. Two of three of those landing means the consumer is rolling and the cyclical leg of the supercycle thesis (we don’t need a strong consumer for AI capex to keep going) gets its real-world signature.

Breaks: All three retailers print clean beat-and-raise. That would suggest the Michigan sentiment number (48.2 record low) is a sentiment phenomenon rather than a spending one — the consumer is grumpy but not actually pulling back. That would force a recalibration of the “cooling under semis” framing, because the consumer would still be carrying its share of the load.

3. Test 2 — The Fed (minutes Wednesday 2 PM ET)

The April 28–29 FOMC meeting voted 8–4 to hold the federal funds rate at 3.50–3.75%. That is the most divided FOMC since October 1992. Stephen Miran dissented for a 25bp cut. Beth Hammack, Neel Kashkari, and Lorie Logan dissented against the existing easing bias in the post-meeting language — they wanted a more neutral or hawkish tone.² The minutes will show how that fight actually played out behind closed doors, and after the hot April CPI/PPI prints, the market is positioned for the hawks to have had the better of it.

What we’re watching for in the minutes:

• Distribution of cut/hold/hike views. A formal “dot plot” substitute reading. If 3+ members are characterized as comfortable with a hike at the next meeting, the 28% December-hike probability moves higher and the 10-year tests 4.60%.
• Language on tariff and energy passthrough. The April PPI was the largest monthly gain since March 2022. If the minutes acknowledge that explicitly — treating it as a one-time shock vs. a persistent inflation source — that tells you whether the Fed sees this as a transitory or structural problem.
• Discussion of de-anchoring inflation expectations. Michigan year-ahead inflation jumped from 3.8% to 4.8% in the May survey — the largest one-month spike in over a year. If multiple voters flag the expectations move as a concern, the door for cuts effectively closes for 2026.

Hawkish minutes (more than 3 voters flagged as hike-comfortable, or explicit de-anchoring concerns) is bearish for long-duration growth equity. Dovish minutes (Miran’s view characterized as gaining traction with two more members) probably extends the rally regardless of what NVIDIA prints two hours later.

4. Test 3 — NVIDIA (Wednesday after the close)

The print of the quarter, on the most asymmetric setup we have seen this year. Stock at $223 entering the print with consensus $272 PT (+22% implied upside).³ The setup is rich, the bar is high, and the optionality from the H200 China license is genuinely material.

What we’re watching specifically

• Q2 revenue guide. Above $87B is the line where the supercycle entrenches. The whisper number is closer to $90B. Below $85B is where the “growth deceleration” narrative wins and the stock takes 10–15% off the multiple.
• Non-GAAP gross margin commentary. Holding ~75% as the Blackwell mix scales is the structural case. Any softening — even just management hedging on “production yields stabilizing” or “mix headwinds” — is more important than a revenue beat. Goldman flagged this as the single most important thing in the print.
• Rubin platform color. Rubin is the next architecture after Blackwell, targeted for H2 CY26. Consensus has Rubin starting to contribute revenue next quarter and assumes $38B for the full year. Any concrete timeline or customer-design-win commentary from Huang feeds directly into FY28 estimates.
• H200 China utilization. The U.S. cleared shipments to ten Chinese firms (Alibaba, Tencent, ByteDance, JD, Lenovo, Foxconn) at up to 75,000 chips each — potentially 750,000 H200 chips of incremental TAM.⁴ Beijing has so far blocked actual deliveries. Any explicit color from Huang on which firms have completed orders, whether the Trump-Xi summit produced movement, and the FY27 revenue contribution math is the single biggest source of upside optionality in the print.
• Q1 data-center revenue mix. Blackwell drove ~70% of data-center compute revenue last quarter and the Hopper transition is essentially complete. Networking is the second derivative — up 64% sequentially last quarter on NVLink and Ethernet AI Switch adoption. Continued networking acceleration is the read-through to Cisco’s $9B AI guide and the entire data-center power complex.

Positioning into the print

We are not adding to NVDA-specific exposure into the print. The asymmetry from $223 with consensus PT $272 is real, but the implied move from the options market is large enough that the post-print drawdown on a sub-$85B guide would be in the 8–12% range. Position sizing matters more than directional conviction here. Our preferred expressions remain at the platform level — long the AI complex through diversified exposure (NVDA + AVGO + MU + CSCO + the power names) rather than concentrated single-name through the print.

If the print delivers (Q2 guide $87B+, GM held at 75%, constructive China commentary), the entire AI complex repriices higher into Friday. If it disappoints, the cooling-under-semis thesis gets stress-tested for the first time. Either way, we are sized for it.

5. What last week taught us, going into this one

Friday closed down 1.24% on the S&P and 1.54% on the Nasdaq — the first real sell day in seven weeks.¹ The driver was profit-taking in tech (Intel −6%, AMD −5.7%, Micron −6.6%) and disappointment that the Trump-Xi summit ended without an explicit H200 delivery breakthrough. But the week still finished green by ~0.3% on the S&P and Nasdaq, with both indices booking fresh ATHs midweek.

The takeaway for this week: the seven-week win streak has its first real wobble, the rate-sensitive equity bid is starting to feel the 4.5%+ 10Y yield, and the AI complex specifically is over-positioned into a single-day binary catalyst. Entering Wednesday with the post-Friday pullback is actually constructive — some of the “chase” froth came out of the names on Friday, leaving the print to do real work instead of just confirming an extended rally.

6. Positioning into Wednesday

Same framework as the last two weeks, reinforced by the Friday pullback and the catalyst density of this week.

Overweight (no change)

Structural AI infrastructure across memory (MU), compute (NVDA, AMD), networking (AVGO, CSCO), and data-center power (electrical equipment, cooling, utility-interconnect). The Cisco $9B AI orders guide from last week confirmed the demand backdrop. The H200 China optionality is real even if no shipments have flowed yet.

Overweight (added emphasis)

Cybersecurity and defense as the higher-rate / higher-geopolitical-risk hedge. Both have low macro beta, durable demand, and historically work in stagflationary regimes.

Underweight (deeper conviction after last week’s data)

Broad cyclicals tied to consumer discretionary big-ticket (autos, furniture, department stores, full-price specialty). The April retail mix — furniture −2%, autos −0.5%, department stores −3.2%, clothing −1.5% — was the leading indicator. Home Depot Tuesday and Walmart Thursday will be the confirming corporate prints.

Rate-sensitive growth — long-duration software (still rich on multiples), unprofitable growth, anything that needs cheap capital. 10Y at 4.49% and 30Y above 5% is the regime, and the FOMC minutes Wednesday may push both higher.

Hedge layer

Long-end Treasuries if 30Y crosses 5.10% (mean-reversion play on rate-shock). Gold as the structural hedge while inflation expectations are de-anchoring — the May Michigan 4.8% one-year reading is doing real work here. We are not running a tactical short on the AI complex going into the print — the asymmetry on a beat is too large — but we are sized so that a 10–15% drawdown on a bad guide doesn’t change our underlying thesis sleep.

7. What changes the rating

Coming into Wednesday, three specific things would shift our positioning bias from “continue with conviction” to “trim and reassess.” Probability-ordered.

1. NVIDIA guides Q2 below $85B with any non-GAAP GM softening. That is the single highest-probability trigger and the cleanest one. It would re-rate the AI complex 10–15% lower and confirm that the supercycle was ahead of the fundamentals on the recent leg. Probability of this exact combination: low (~15%), but the impact is high enough to size for it.
2. FOMC minutes reveal three or more voters comfortable with a hike at the next meeting. That pushes the 28% December-hike probability above 40%, takes 10Y to 4.60%, and breaks the long-duration growth bid. Probability of this characterization: moderate (~25%), given the April vote was already 8–4 with three hawkish dissents.
3. Walmart prints a soft guide on FY27 EPS with food-vs-GM mix widening. That confirms the consumer rolling and pulls the cyclical complex (already underweight in our book) into a deeper drawdown. Probability: low-to-moderate (~20%), as Walmart has historically guided conservatively and reset later.

None of these have triggered yet. Two of three triggering this week pulls our rating to neutral. Three of three triggering is the “cooling-under-semis thesis just got tested in real time” scenario — we’d trim the supercycle names hard and reposition more defensively into June.

8. Closing

Last week we said the supercycle was winning by a wider margin than anyone modeled. This week it has to defend that margin against three specific tests in five trading days. Wednesday is the day that resolves the most important questions. We are positioned for it; we are not chasing into it.

Trade the catalyst calendar, not the headlines. Wait for Wednesday’s minutes at 2 PM ET. Wait for NVIDIA’s Q2 guide after the close. Re-rate Thursday morning on the Walmart confirmation. Friday is for repositioning, not reacting. That’s the whole week in one paragraph.

Sources

1. TheStreet, “Stock Market Today: Dow futures fall following record close (May 15, 2026).” thestreet.com ↵ ↵
2. The Conference Board, “FOMC Decision: Three Hawkish Dissents Signal Fed Moving Closer to Neutral Stance.” conference-board.org ↵ ↵
3. Benzinga / Goldman Sachs, “Nvidia’s $1 Trillion Promise May Just Be The Start (Q1 FY27 preview).” benzinga.com ↵ ↵
4. CNBC / Reuters, “U.S. clears H200 chip sales to 10 China firms as Nvidia CEO looks for breakthrough.” cnbc.com ↵ ↵
5. Zacks / Trading View, “Home Depot to Post Q1 Earnings: Is Now the Right Time to Invest?” tradingview.com ↵
6. Invezz, “Walmart earnings preview: analyst says WMT’s premium multiple is justified.” invezz.com ↵

Nothing on this page is investment advice. See disclaimer.