TheMoneyFlowJournal

Alphabet AI talent exodus: specific names confirmed — Noam Shazeer to OpenAI, John Jumper to Anthropic STRUCTURAL
The Alphabet AI talent crisis has specific confirmed exits: Google VP of Engineering and Gemini AI co-lead Noam Shazeer left for OpenAI last Wednesday. Then, DeepMind VP and Engineering Fellow John Jumper left for Anthropic on Friday. Shazeer was one of the co-authors of "Attention Is All You Need" (the Transformer paper that underpins all modern AI). Jumper won the Nobel Prize in Chemistry for AlphaFold. These are not ordinary departures — they represent the two most decorated AI researchers in Google's history leaving in the same week. CNBC confirmed: Alphabet fell 5% Monday then another 2% Tuesday. This is a structural moat erosion event, not a one-day headline.

FedEx beat, fell 6% AH — the "sell the news" template repeating for June 2026 FDX
FedEx EPS $6.31 beat $6.02. Revenue $25.0B beat $24.03B. Full-year FY26 EPS $20.24 (+11%). Network 2.0 on track for $1B in cost savings. Healthcare vertical at $10B revenue. Data centre vertical growing double-digit. And yet: FDX fell 6.16% AH to $297.70 — back below $300. The reasons: (1) stranded costs from FedEx Freight spinoff create calendar year 2026 near-term headwind; (2) new pilot contract costs; (3) a "transition year" framing in guidance. TheStreet: "FedEx is emerging stronger without the weight of its FedEx Freight business, even if its stock doesn't suggest that." $13.3 billion in cash and equivalents, $4.1 billion cash dividend from the Freight spinoff, double-digit data centre logistics growth. The business is fine. June 2026 markets are not rewarding beats — they're punishing guidance uncertainty.

MU TONIGHT  Micron at $1,074 after −11.4% — the pre-earnings selldown may have reset the bar
MU entered Tuesday at approximately $1,225 and closed at $1,074.60 — a $150 single-day loss. This is the most aggressive pre-earnings selldown since the stock started its AI-driven rally. The dynamic: markets are pricing in a Micron disappointment (driven by SK Hynix's slowdown news) before the print even lands. But this cuts both ways: if Micron reports blowout HBM demand and raises guidance, $1,074 becomes a gift. If Micron confirms AI memory demand is plateauing, $1,074 becomes a temporary stop on the way lower. UBS: "DRAM bit demand is still set to vastly outpace supply growth in the coming years driven by more memory-intensive AI workloads." Needham target: $1,550 (+44% from $1,074). The bar has been meaningfully lowered in one session. Micron reports tonight at approximately 22:30 CET.

SKH CONTEXT  What "SK Hynix slowing AI memory expansion" actually means for Micron's HBM thesis
SK Hynix announced plans to moderate the pace of its HBM capacity expansion — news that markets interpreted as AI memory demand peaking. But there are two readings: (1) BEARISH: demand for HBM is slowing because AI capex is being more carefully managed; Micron's guidance tonight will confirm or deny. (2) BULLISH: SK Hynix is exercising supply discipline, keeping HBM scarce and pricing strong — which is good for Micron's margins. If Micron reports strong Q4 HBM guidance with rising ASPs (average selling prices) while SK Hynix holds back capacity: Micron gains market share at high prices. The nuance matters enormously. Watch for both demand guidance (units) AND pricing guidance (ASPs) in tonight's Micron call.

ORACLE JOBS  Oracle cut 21,000 jobs over the past year for AI — the automation signal hiding in plain sight
CNBC reported Tuesday: Oracle cut approximately 21,000 jobs over the past year as artificial intelligence adoption reshapes its workforce. Oracle's headcount reduction is not a demand collapse — it's a company restructuring its cost base for the AI era (fewer humans required for tasks now automated). Oracle shares fell 2.6% on the news. But the signal for markets is broader: the AI investment cycle is reshaping employment composition across enterprise technology. Companies that invest in AI infrastructure (Micron, Nvidia, AVGO, Vertiv) benefit from this trend. Companies that provide human labour to do what AI now does (traditional IT services like ACN, Infosys, Wipro) face structural headwinds.

IBM UPGRADE  IBM +$13 on JPMorgan upgrade — the "old tech gets AI religion" trade
JPMorgan upgraded IBM to Overweight with a $291 target (from $270), citing "greater confidence in the potential for acceleration in IBM's software business in H2 2026." President Trump also signed two executive orders on quantum computing development, targeting commercially relevant quantum computers by 2028 — directly benefiting IBM (which builds superconducting quantum computers). IBM +5% on the day while the Nasdaq fell 2.21%. This is the "old tech pivoting to AI" trade: IBM has been the beneficiary of enterprise AI platform adoption (WatsonX), and the quantum computing EO provides an additional government tailwind. Contrast IBM's government-backed positioning with Google's AI talent drain — one has a moat being built, one has a moat being eroded.

DXY: global risk-off from KOSPI circuit breaker = short-term USD safe-haven bid · PCE tomorrow is the real test
The KOSPI −9.99% circuit breaker triggered a global risk-off move that typically benefits the USD as a safe-haven currency. With the Nasdaq falling 2.21% and global tech stocks under pressure, DXY likely held or extended its one-year high on Tuesday. Today: same dynamic — Micron earnings uncertainty + PCE anticipation = cautious USD positioning. Thursday PCE at 14:30 CET is the macro event that can break the DXY one-year high: a cool May PCE (headline below 3.4%) begins the "stale dots" reversal → DXY from 99-100 toward 97-98. But the "split" result (headline down, core up per FactSet consensus) keeps DXY supported through this week.

EUR/USD — tech risk-off benefits USD, not EUR · 1.14 floor must hold through Micron + PCE
Global chip selloffs drive risk-off capital flows into USD, JPY, and CHF — not EUR. EUR/USD faces dual headwinds today: tech risk-off (USD safe-haven bid) and PCE uncertainty (will core PCE tick up?). The ECB hike floor keeps EUR/USD supported structurally at 1.13-1.14. Today is a holding pattern day; Thursday PCE is the directional catalyst. If PCE headline falls while core rises slightly: EUR/USD stays near 1.14-1.15, markets wait for July 10 June CPI. If PCE surprises dovishly: EUR/USD recovers toward 1.17.

XAUUSD — chip risk-off is not gold-positive · real catalyst remains PCE + July CPI
Chip selloffs don't reliably drive gold higher — they're tech-specific events, not macro inflation or geopolitical catalysts. Gold at ~$4,150 stays in its post-FOMC range. Oil at $74.82 (Iran license) remains the structural gold bull driver: falling oil → lower CPI → lower real rates → gold recovery. But that data lag means Thursday PCE only shows partial deflation (May energy prices, not June). The clean gold bull signal is July 10 June CPI at $70-$72 WTI. Today and tomorrow are consolidation; Thursday's split PCE likely keeps gold near $4,100-$4,200.

BTC $62,174 — holding 200-week MA through global chip carnage · Deribit expiry Friday
BTC fell 2.81% to $62,174 on Tuesday — dragged down by the broad risk-off from the chip sector rout, but notably more resilient than the Nasdaq (−2.21%) or chip stocks (SMH −6.5%). BTC is holding at and above the 200-week MA (~$62,000). The relative outperformance vs chips is meaningful: BTC has decoupled from its prior high correlation with semiconductor stocks, likely because the crypto market is pricing the oil/inflation/PCE narrative more than the AI chip cycle narrative. Thursday's PCE is BTC's near-term directional catalyst more than tonight's Micron earnings. If PCE shows disinflation → DXY reverses → BTC recovers above $65K. If PCE is hawkish → BTC retests the 200-week MA floor.

Deribit quarterly BTC + ETH expiry Friday June 26 at 10:00 CET — positioning note
With BTC at $62,174 and Deribit quarterly expiry two days away, the max pain dynamic is increasingly important. Quarterly expiries are the year's largest crypto options events — billions in BTC and ETH positions settle. At $62,174, BTC is near the likely max pain zone for the June quarterly (most open interest concentrated near $60,000-$65,000 range). This creates "pinning" pressure toward the max pain price through Friday, with post-expiry volatility releasing the artificial support/resistance. Monday June 27 is historically more directional than the pre-expiry week. Plan accordingly: Micron (tonight) + PCE (Thursday) + Deribit expiry (Friday) = three overlapping catalysts creating maximum volatility density this week.

Clarity Act: 10 days to July 4 · Senate still dark on floor vote · last realistic window this week
10 days to July 4. No Senate floor vote announcement. Senate Majority Leader Thune has not indicated when — or whether — the Clarity Act gets floor time before the July 4 recess. The bipartisan Banking Committee passage (15-9) and the 200+ company joint letter were not enough to force scheduling. Galaxy's 60% and JPMorgan's sub-50% passage odds reflect this uncertainty. If Thune doesn't schedule the vote by end of this week (June 27), the August recess almost certainly pushes this to September. Each day without a scheduling announcement is a marginal negative for BTC regulatory optimism. Watch for any Senate leadership statement today or Thursday.

Fed bank stress tests today 4 PM ET — 33 major banks. Capital returns in focus
The Federal Reserve releases its 2026 annual stress test results at 4:00 PM ET — the same day as Micron earnings AH. Tests evaluate major US banks' resilience under a hypothetical severe recession scenario. With the 30-year mortgage at 6.52%, housing starts at 6-year lows, and October rate hike odds at 60.7%, this year's stress test scenario is more realistic than most. A clean pass (all 33 banks pass) → banks announce dividend increases and buyback authorizations Friday → financials rally Friday → broadens the market beyond tech. Any bank failure → capital return programmes frozen. Historical precedent: all major banks have passed for 4 consecutive years. Expect unanimous pass again, with dividend/buyback tailwinds for Friday.

Carnival EPS beat, guidance miss — the consumer spending resilience story with an asterisk
Carnival Q2: EPS $0.41 (beat $0.35, +17%), Revenue $6.663B (missed $6.692B by 0.4%), Q3 EPS guidance $1.35 (below $1.42 consensus, −5%). FY2026 EPS guidance $2.22 (just below $2.23 consensus). Stock fell ~6% to $28.40. The beat on EPS suggests cruise demand is real and margins have held. The guidance miss reflects caution about the second half — possibly from Iran-related geopolitical uncertainty (Hormuz transit risk for Caribbean-to-Europe routes), or from the "cumulative inflation shock" of the past year leaving consumers more cautious. Oil at $74.82 (WTI) should improve CCL's fuel cost outlook for Q3, but the guidance didn't capture this tailwind yet.

NVIDIA SHAREHOLDER MEETING — Today's Nvidia shareholder meeting is the most watched corporate governance event in tech. Jensen Huang's formal comments on AI demand, Blackwell GPU ramp, and data centre capex trends will be parsed word-by-word. Any positive AI demand commentary from Huang — especially on HBM consumption (which is Micron's product) — would pre-position markets positively for tonight's Micron print. Conversely, any caution about AI capex timing would amplify the SK Hynix slowdown narrative. Watch live stream or commentary in the hours before Micron's AH results.

SPCX BOND ISSUANCE — SpaceX confirmed its first-ever bond issuance this week — tapping debt markets to fund operations. At $147-150, SPCX stock is down from the $175 post-IPO range. The bond issuance signals confidence in cash flow generation (you don't issue bonds if you can't service them). SPCX NDX fast-track inclusion (Reuters: within 1 month) + MSCI passive buying still ongoing + first bond = SpaceX maturing from startup to institutional-grade public company. The IPO correction (from $176.52 debut high to $147-150) is normal post-honeymoon re-pricing; the structural story is intact.

PCE TOMORROW — Tomorrow's May PCE at 14:30 CET will be the most parsed number this week. Dallas Fed: Iran war pushed headline PCE up 1.7pp annualised in Q1 with "effects expected to remain elevated through Q3." WTI at $74.82 removes most of this premium for May. But core PCE (ex-food, energy) was 3.3% in April and is expected to tick HIGHER for May per FactSet. This creates the "split": headline lower, core higher. A split PCE won't give the market the clean dovish signal it wants but won't confirm the hawkish case either. It keeps the September dot-plot revision in play without accelerating it. Net for markets: slight relief on headline, limited impact on hike odds. The real dovish test is July 10 June CPI.