Week recap · Macro · AI · semis · energy · June 8, 2026
Streak broke at nine. The thesis hits a tougher test.
The S&P 500 broke its nine-week winning streak. Closed Friday at 7,383.74, down 2.6% on the day, down 2.6% on the week.1 Two specific catalysts did the work. Thursday: Broadcom reported $10.8B in AI revenue, +143% Y/Y, but maintained rather than raised its 2027 AI semi outlook — the Street took that as the first crack in the AI capex story and Nasdaq finished down 4.18%.2 Friday: May nonfarm payrolls printed +172K against expectations of 80–85K, the unemployment rate held at 4.3%, and March/April revisions added another +93K to the books. The print effectively killed the last hopes for a 2026 rate cut and drove the 10-year above 4.54% with 20/30s above 5%.3
Our May 29 base case framework was S&P 7,500–7,800 over the next 90 days at 50% probability. Friday’s close at 7,383.74 sits ~115 points below the lower bound of that range — missed our base case by ~1.5%. The 10Y range we gave was 4.25–4.65%; the close at ~4.54% landed inside it. The trip-wire #1 condition (NFP >250K with average hourly earnings >4% Y/Y) did not trigger — the 172K print was hot relative to expectations but below the breakage threshold we named, and AHE held at 3.4% Y/Y.3 Yet the market reacted as if the trip-wire fired anyway. Lesson: the trigger isn’t the absolute level. The trigger is the reaction function.
And then today. Iran and Israel exchanged missile strikes over the weekend, threatening the tentative US–Iran ceasefire memorandum that had been pulling the energy-driven inflation premium out of the curve. WTI gapped above $94 this morning, up from Friday’s $87 close — roughly +8% in a single overnight session.4 That reopens the inflation-via-crude pathway we removed from the trip-wires when we tightened the framework. The Iran de-escalation premium we credited for the 25bp 10Y rally on May 29 is back on the table.
The week ahead is mechanical: May CPI prints Wednesday June 11 at 8:30 AM ET — the dress rehearsal for the FOMC. FOMC June 16–17 is the framing fight (easing-bias language stays or goes), with the SEP/dot-plot update at 2:00 PM ET Wednesday June 17. Our positioning into both is unchanged: light duration, trim concentrated NVDA into rallies, the AI infrastructure breadth basket holds but the entry-window is meaningfully better than a week ago.
1. Where we are: streak broke at nine; chip complex tested
The S&P 500 closed Friday June 5 at 7,383.74, down 200.57 points (−2.6%) on the day and approximately 2.6% lower for the week. The Nasdaq closed at 25,709.43, down 4.18% on the day, ending the week down ~4.7%. The Dow finished at 50,866.78, down 695.15 points (−1.3%) on the day. The VIX exploded +39.7% on Friday to close at 21.51 — its biggest single-day spike of the year so far.1
The semiconductor complex took the structural hit. Single-day moves on Friday alone: NVIDIA −6.20% to $205.10, Broadcom (a second-day move after Thursday’s outlook print) −5.7%, Micron −9.4% to ~$1,004, AMD −10.86% to $466.38, Intel −11.28% to $99.17.5 Dell, our single-name watch from the May 29 piece, closed Friday at $394.39 −6.55% — a clean retest of the $395 lower bound of the $395–$425 consolidation zone we identified. Dell had also hit a fresh all-time closing high of $465.96 on June 1 before the chip rout sold it back through the zone.6
The 10-year Treasury yield finished Friday above 4.54%; the 20- and 30-year yields pushed back above 5%. The bond market reaction was clean: hot NFP, kill the cut, term premium widens.3
Chart 1 — S&P 500 weekly closes through June 5; streak ended at 9
Up nine, then down on the tenth. The break came on a hawkish reaction function, not a trip-wire trigger.
Weekly closes April 17 through June 5. The 9-week up-streak (gold marker at May 29 close of 7,580) ended with a −2.6% week into June 5. The break came on a Broadcom maintained-outlook reaction Thursday and a hot 172K NFP print Friday — not on the trip-wire #1 condition (NFP >250K with AHE >4% Y/Y), which did not trigger. Sources: TheStreet, Bloomberg, BBN Times closing summaries.
2. May 29 framework, scored against actuals
Eight days ago we set out a 90-day base case (50%), a 10Y range, three trip-wires, and one forward call. Here is how each item closed by Friday 4 PM and where the carry-over sits going into CPI.
| Framework item | May 29 view | June 5 outcome | Verdict |
|---|---|---|---|
| S&P 90-day base | 7,500–7,800 (50%) | Friday close 7,383.74 — ~1.5% below the lower bound | JUST BELOW |
| 10Y range | 4.25–4.65% through June | Friday close above 4.54%; 20/30s back above 5% | INSIDE |
| Trip-wire #1 (labor) | NFP >250K + AHE >4% Y/Y | NFP +172K, AHE +3.4% Y/Y — below the named threshold | DID NOT TRIGGER, MARKET REACTED ANYWAY |
| Trip-wire #2 (hyperscaler capex cut) | META/MSFT/AMZN/GOOGL pre-announces cut | No cut. Broadcom maintained-not-raised was the proximate selloff catalyst | ADJACENT, NOT EXACT |
| Trip-wire #3 (credit event) | Private credit halt / regional bank CRE disclosure | None observed in the tape | QUIET |
| Forward call (CapEx raise window) | META/MSFT/AMZN raises 2026 CapEx Jun 10–25 | Window opens this week (June 10). Status: open. | OPEN |
| Iran / geopolitics | Removed from trip-wires as "partially playing out" | Today Iran/Israel missile exchange; WTI +8% overnight from $87 to $94+ | CALL OFF — deserved to stay a trip-wire |
Three observations matter here. First: the base case missed by ~1.5%. We did not size the framework for a single-week 2.6% drawdown, which means our position-sizing call (light duration, AI infrastructure breadth on) ran into headwinds but did not break. Second: trip-wire #1 didn’t fire on absolute thresholds. The 172K print was hot to expectations but well below the 250K we named. The market reacted to the surprise vs. expectations, not the level vs. our line. That is a calibration error in the trip-wire definition itself. We will tighten that in the refreshed list below. Third: removing the Iran trip-wire was a process error. We called it "partially playing out" when we tightened, and today it is the catalyst on the wire. Lesson: don’t prune a risk just because it has shown up once.
3. The Broadcom signal: record AI revenue, maintained outlook
Broadcom reported fiscal Q2 results Tuesday after the close. AI semiconductor revenue: $10.8B, up 143% year-over-year — a record.2 The number itself was a clear validation of the AI infrastructure thesis we had carried forward through the May 29 piece. CEO Hock Tan’s decision to maintain, rather than raise, the company’s 2027 AI semiconductor outlook is what sold the stock and the complex with it. The Street read the unraised guide as the first explicit signal that hyperscaler-driven AI demand growth may be plateauing on the planning horizon — even though Q2 itself printed as fast as any historical period.
That reading dovetails with the May 29 NVDA paradox we wrote (broadening leadership, NVDA-specific positioning unwind) but it inverts the implication. We argued the breadth basket (Dell, MRVL, AVGO, ANET, VRT) would re-rate higher as leadership rotated away from NVDA. The Broadcom report suggests the breadth basket also has multiple-compression risk when the supply-side stops surprising to the upside. Our pivot: the basket is still the cleanest expression of long AI infrastructure for the next 90 days — but the entry-window after this selloff is tighter, and we want at least one of META, MSFT, or AMZN to confirm with a CapEx raise before we add. The forward call we put on May 29 (CapEx raise between June 10 and June 25) is the exact gate.
4. May NFP: hot to expectations, not hot enough to trigger
The Bureau of Labor Statistics reported May 2026 nonfarm payrolls at +172,000, against consensus expectations in the 80,000–85,000 range — roughly double the high end of the consensus. The unemployment rate held steady at 4.3%. Average hourly earnings rose 0.3% month-over-month and 3.4% year-over-year, in line with expectations. Labor force participation held at 61.8%. Revisions to prior months added a combined +93,000: March was revised up by 29,000 (from +185,000 to +214,000) and April was revised up by 64,000 (from +115,000 to +179,000).3
The number itself does not meet the trip-wire #1 condition we named in the May 29 piece (NFP >250K with AHE >4% Y/Y). The labor market is firm rather than overheating. But two characteristics of the print made it functionally hawkish for markets: (a) the surprise magnitude versus consensus — a roughly 2× beat on the headline is the kind of single-data-point shock that re-prices the front end immediately; (b) the revisions, which retroactively make April look stronger than the “cooling labor market” narrative had priced. Together those two characteristics killed the residual probability of a Fed cut in 2026 — current pricing is at zero cuts through year-end, against 1 cut prior to the print.
Chart 2 — May NFP: actual vs. consensus, with revisions
+172K beat consensus by ~2×. With +93K of prior-month revisions, the labor read got materially stronger overnight.
May headline NFP of +172K printed roughly 2× consensus expectations of ~85K (left panel). Prior-month revisions added another +93K combined — March from +185K to +214K, April from +115K to +179K (right panel). Trip-wire #1 (NFP >250K with AHE >4%) did not trigger on the absolute condition, but the magnitude of the surprise + revisions was sufficient to kill the cut narrative. Sources: BLS Employment Situation, June 5 release; Bloomberg; CNBC.
5. Iran reignites: today’s gap higher in crude
Over the weekend, Iran and Israel exchanged missile strikes, threatening the tentative 60-day US–Iran ceasefire memorandum that had been awaiting Trump’s sign-off. WTI gapped above $94 per barrel overnight, up from Friday’s ~$87 close — roughly +8% in a single session. Brent crude futures pushed back above $97. The Strait of Hormuz tail-risk that markets had partially priced out during the May 22–29 de-escalation rally is back in the front of the curve.4
The mechanism is clean: every $5 sustained move higher in WTI adds roughly 8–12 basis points to one-year forward inflation expectations through energy pass-through, which the Fed will read in the Cleveland nowcast before CPI prints on Wednesday. If the Iran/Israel exchange is contained and WTI back through $90 by the weekend, the inflation impulse is transient and the Fed’s dovish camp still has the M/M PCE trajectory on its side. If WTI holds >$95 through CPI on Wednesday, the trajectory argument we leaned on for the dovish camp at the May 29 piece materially weakens.
This is the trip-wire we should not have pruned. Adding back to the list in Section 9.
6. Thesis check: AI infrastructure breadth, after the test
The structural call we have carried since the May 8 cooling-under-semis piece is unchanged: the AI infrastructure supercycle is multi-year and broadening across the stack. The chip selloff did not invalidate it. What changed last week is the rate-of-change story — Broadcom’s maintained-not-raised outlook is the first explicit data point on the supply side that does not extrapolate up and to the right. The pivot from May 29 (rotate within the AI complex, not out of it) survives, but with one calibration: the breadth basket needs a CapEx-side confirmation in the next two to three weeks or the multiple compression starts pricing through Dell, Marvell, AVGO, ANET, and VRT alongside the prior leader.
The single-name framework lives in the standalone pieces. The MU after UBS piece and the MU June outlook carry the memory thesis; FQ3 prints Wednesday June 24 after the close, which is now the next major single-name catalyst in the book. The Dell pitch — with the breakout level math — remains forthcoming.
7. Outlook & framework into CPI and the FOMC
The forward window is now exactly nine days to the FOMC. May CPI prints Wednesday June 11 at 8:30 AM ET — the dress rehearsal. June 17 brings the FOMC statement at 2:00 PM ET with the updated Summary of Economic Projections and dot plot, followed by Chair Powell’s press conference at 2:30 PM ET.
Chart 3 — S&P 500 scenarios into the June 17 FOMC, refreshed
Base case 7,300–7,600 (50%). Bull 7,800+ (20%). Bear 6,900–7,200 (30%).
Refreshed scenarios over the next nine days to the FOMC. The May 29 framework anchored on a longer 90-day horizon; this one tightens to the FOMC outcome itself. Probability weights shifted bearish (bear now 30%, bull now 20%) to reflect the hot NFP + Iran energy impulse, both of which push toward an FOMC where the easing-bias language is at material risk. Base case still requires the bias to hold and CPI to print in-line.
10Y range, refreshed
4.45–4.85% through the FOMC. The post-NFP move brought the 10Y to ~4.54%, and the Iran-driven energy impulse plus a potential hot CPI Wednesday could push the long-end through 4.70% before the meeting. We size below benchmark on duration into June 17.
8. Trip-wires, refreshed
The May 29 trip-wire list missed the Broadcom-style guide-not-raise risk and the Iran reignition. We are not adding more cards just because we missed; we are recalibrating the existing ones and putting Iran back. Total trip-wires: still three. Probability of any one tripping in the next 30 days: meaningfully higher than the May 29 framework implied.
June NFP at 200K+ or AHE acceleration above 3.7% Y/Y.
Recalibration: the May 29 trip-wire used 250K + 4% AHE as the threshold. The May print of 172K with 3.4% AHE did not meet either condition, yet the market priced it as if both had triggered. We are tightening the threshold to capture surprise-vs-expectations behavior the original missed. The next June NFP prints on July 3.
Our response: If the June NFP repeats the May surprise — meaning consensus is below 100K and the actual prints 150K+ — cut AI infrastructure gross another 15%; light duration at the front of the curve only when 10Y backs up through 4.85%.
META, MSFT, AMZN, or GOOGL hosts a mid-quarter analyst event between now and the Q2 print and maintains rather than raises 2026 CapEx.
Recalibration: the May 29 trip-wire was set to a CapEx cut. The Broadcom signal showed the market reacts to a "not raised" the same way it would react to a cut. Adjusting the trigger condition accordingly. The behavior we are protecting against is multiple compression on lack of upward surprise, not just on explicit cuts.
Our response: Net out the AI infrastructure breadth basket; rotate the proceeds into AI software where the demand-side monetization is cleaner (CRM, ORCL, SNOW). The structural long stays on at neutral weight in NVDA.
WTI sustains above $95 through May CPI on Wednesday June 11 or US-Iran ceasefire memorandum formally collapses.
We removed this trip-wire in the May 29 tightening on the framing that Iran was "partially playing out." That was a process error. The weekend missile exchange and the +8% gap in WTI overnight today is exactly the catalyst we should have kept the wire on. Reinstating with a specific threshold linked to CPI day.
Our response: Long energy (XOM, CVX) added at the morning levels; short broad-equity beta into the Wednesday print only if WTI is still >$95 at 8:00 AM Wednesday; the supercycle thesis still survives but the multiple compresses on rate-side stress.
9. The forward call: status check
The May 29 forward call was specific: META, MSFT, or AMZN raises 2026 CapEx guidance at a mid-quarter analyst event between June 10 and June 25, with falsification by June 25 = call dies. Today is June 8. The window opens this week. Status: still open.
Window opens June 10. Falsification date June 25. No raise observed yet.
The CapEx data we have in hand — META’s prior $115B–$135B 2026 range (raised earlier in the year to the $125B–$145B range), Microsoft’s "Azure AI demand exceeding capacity" language, Amazon’s ~$200B 2026 plan — describes baseline commitments. The forward call is about an explicit incremental raise inside our window, not the existence of high CapEx in absolute terms. Broadcom’s maintained outlook last Thursday makes the catalyst more important: without an explicit raise on the demand side, the breadth basket faces the same multiple-compression risk the supply side just took.
What we are watching: any June investor day, mid-quarter analyst event, or pre-announcement from the four hyperscalers. Falsification: June 25 with no raise = call dies; we do not double down. Trade: long the AI infrastructure breadth basket (DELL, MRVL, AVGO, ANET, VRT) at the post-selloff entry levels; trim post-announcement on the gap-up.
Sources & footnotes
- S&P 500 close June 5, 2026 at 7,383.74 (−2.6%, −200.57 pts); Nasdaq close 25,709.43 (−4.18%); Dow close 50,866.78 (−695.15 pts, −1.3%); VIX +39.7% to 21.51. Sources: Yahoo Finance, Washington Post Friday close summary, TheStreet Friday recap. ↩
- Broadcom fiscal Q2 2026 earnings (reported Tuesday June 2 after close): AI semiconductor revenue $10.8B (+143% Y/Y, record). 2027 AI outlook maintained rather than raised. Stock −5.7% Friday June 5; the maintained guidance was the proximate catalyst for the Thursday June 4 chip selloff that wiped ~$1.3T from semiconductor market caps. Sources: CNBC, Intellectia, 24/7 Wall St. ↩
- May 2026 Employment Situation Report, BLS, released Friday June 5: nonfarm payrolls +172,000 (vs. consensus of 80–85K); unemployment rate 4.3% (unchanged); average hourly earnings +0.3% M/M, +3.4% Y/Y; labor force participation 61.8%. Prior-month revisions: March revised up by 29K (from +185K to +214K); April revised up by 64K (from +115K to +179K). 10-year Treasury yield closed above 4.54%; 20-year and 30-year yields above 5.0%. Sources: BLS Employment Situation; Bloomberg; CNBC; Verified Investing. ↩
- WTI crude futures: Friday June 5 close ~$87.18 (down ~1.9% for the day, −16.5% MTD as of Friday); Monday June 8 gap higher to above $94 per barrel on Iran/Israel weekend missile exchange. Brent above $97. Tentative US–Iran 60-day ceasefire memorandum awaiting Trump sign-off and now in doubt. Sources: TradingEconomics, CNBC weekend updates. ↩
- Semiconductor sector single-day moves Friday June 5: NVIDIA close $205.10 (−6.20%); Broadcom (AVGO) −5.7%; Micron (MU) close ~$1,004 (−9.4%); AMD close $466.38 (−10.86%); Intel close $99.17 (−11.28%). Sources: Yahoo Finance NVDA history, Intellectia, 24/7 Wall St. ↩
- Dell Technologies (DELL) close June 5, 2026 at $394.39 (−6.55% on the day); all-time closing high reached June 1 at $465.96; ~15% pullback from June 1 peak. Friday close at the bottom of the $395–$425 consolidation zone identified in the May 29 piece. Sources: Yahoo Finance, public.com analyst snapshot. ↩
Nothing on this page is investment advice. Forward-looking statements are scenarios, not promises. See disclaimer.