MEMORANDUM

MEMORANDUM

TO: The Bull Case
FROM: The Data
DATE: June 5, 2026
RE: We need to talk

I've been patient. You've had your run — nine consecutive winning sessions on the S&P, the Nasdaq up 8.43% in May, Q1 earnings growth at 28.6% year-on-year, the strongest margin print since Q4 2021. I let it happen. I watched the rotation into semiconductors, the MRVL surge, the Jensen Huang quotes moving markets more reliably than any central bank forward guidance. Fine.

But some things have accumulated and they require your attention.

ON THE SUBJECT OF THE RALLY'S ACTUAL ARCHITECTURE

May's S&P gain was 5.26%. Sounds broad. Isn't. Technology jumped 19.76% and did nearly all the index's work. Eight of eleven sectors finished the month in negative territory. Energy dropped 5.63%. Utilities shed 5.19%. Financials, real estate, industrials, consumer staples — all red. This was not a bull market. This was a Nvidia market with a bull market's name tag on.

Bridgewater estimates Alphabet, Amazon, Meta, and Microsoft will collectively spend roughly $650 billion on AI-related infrastructure this year. Meta has guided to $125–145 billion in capex for 2026, including finance leases. Microsoft put $31.9 billion on the table in a single fiscal quarter. Alphabet dropped $35.7 billion in Q1 property and equipment purchases. These are not investments with a clear return horizon. They are bets, made by companies large enough that the bet itself moves the market — which in turn validates the bet, which encourages more of them. The circularity of this should concern you.

Hyperscalers are now collectively spending more than four times what the entire publicly traded US energy sector spends to drill, extract, deliver, and run chemical plants. Amazon's capex alone exceeds the full US energy sector. At some point, someone is going to ask what the return on invested capital actually is, and "the future" will not be a sufficient answer.

ON THE SUBJECT OF INFORMATION SERVICES LOSING 9,000 JOBS IN MAY

The ADP report, which printed 122,000 overall — above the 110,000 consensus, strongest month since January 2025 — had a line item that didn't get much airtime. Information services lost 9,000 jobs. The sector most directly exposed to AI adoption is shedding headcount. The sector being credited with saving the economy is also eliminating jobs within it. Hold both of those thoughts simultaneously and see how comfortable the AI narrative feels.

The broad-based nature of the May ADP gains is real — eight of ten sectors added workers, geography and firm size spread evenly. That's not fake. But the information services number is a preview of something, and markets are choosing not to read it.

ON THE SUBJECT OF PCE AT 3.8%

April CPI came in at 3.8% year-on-year. Core PCE, the Fed's preferred gauge, at 3.2% — accelerating from 2.9% in February. Energy feeding into headline from the Iran war. The Fed held at 3.50%–3.75% at April's FOMC in an 8–4 vote, the most divided since 1992. Prediction markets are now pricing a 96.9% probability of a hold at the June 16–17 meeting. Traders are also — quietly, at the margin — pricing an 11% probability of a rate hike in July, up from 0.9% a month ago.

Read that again. Eleven percent. A month ago it was less than one.

Bank of America doesn't see cuts until July 2027. The market is pricing "nothing happening soon" while simultaneously holding high-multiple tech at valuations that require cheap money or uninterrupted earnings compounding to justify. You cannot have both a hawkish Fed and a Nasdaq that prices in infinite tomorrows. Eventually the math notices.

ON THE SUBJECT OF TRUMP ACCOUNTS

This one is at least novel. The One Big Beautiful Bill Act, enacted July 4, 2025, created Section 530A accounts — informally known as Trump Accounts. The federal government seeds each with $1,000. Households can contribute beyond that. Funds go exclusively into low-fee broad equity index funds. Bloomberg Intelligence puts the initial flow at roughly $12 billion annually, rising toward $21 billion if households max out contributions.

Treasury Secretary Scott Bessent has called it a simple, secure on-ramp for households to build long-term wealth. Account activation is underway through a mobile app attributed to Bank of New York Mellon and Robinhood.

Here is what this actually is: a structural bid for US equities, legislated and permanent, managed by the federal government, beginning July 4. Twelve billion dollars a year flowing exclusively into index funds. Not into bonds. Not into money markets. Equities. This is a policy choice to reflate equity markets using household savings that wouldn't otherwise be in equities at all. The long-term distributional implications are interesting. The short-term market mechanics are simple: more money, same shares.

ON THE SUBJECT OF TOMORROW'S NUMBER

May nonfarm payrolls drop at 8:30am tomorrow. Wall Street consensus sits at 80,000. April came in at 115,000 — itself well above the 62,000 forecast. March was revised up to 185,000. The trend is running hotter than expected, and ADP's beat this week has tilted positioning toward an upside surprise.

If payrolls beat: 10-year yields test 4.60%, the hawkish repricing continues, rate-cut hopes for 2026 are effectively buried, and high-multiple tech faces an arithmetic problem it can't narrative its way out of.

If payrolls miss: brief window of cut expectations, Brent softens, equities bounce, everyone breathes — until the next overnight flash from the Gulf, or until next month's print.

The market wants the miss. The economy is not particularly interested in cooperating.

IN SUMMARY

The bull case is not wrong about earnings. It is not wrong about AI infrastructure spending or semiconductor demand. Where it is making a category error is in treating a highly concentrated, energy-shock-distorted, rate-sensitive, geopolitically-threaded market as though it were a clean cycle expansion. It isn't. The foundation is partially real and partially a very expensive act of collective agreement.

The data remains available for review at your convenience.

Ends.