The Week Ahead:

Mon: G7 leaders summit begins

Tue:
* BOJ meeting: Bank of Japan set to hike key interest rate to 1% [31-year high]; BOJ will also review its bond taper plan, running through March next year, and lay out a new plan for fiscal 2027 and beyond. BOJ is also considering pausing the tapering of its government bond purchasing program, starting in April 2027.
* A Unknown-Known to some Coinbase announcement

Wed: FOMC meeting: “The FOMC will almost certainly leave rates unchanged, but the biggest source of intrigue may be the new chair’s debut press conference. Which Warsh will show up: the recent advocate of lower rates, or the inflation hawk we saw during and after the Global Financial Crisis?”. Will he submit a forecast for rate projections he disdains? How much will he say at a press conference he’d rather not hold?. Warsh wants the Fed to say less and let markets do more of the work…. So far, insiders say his promised “regime change” has been kinder and gentler than feared—more courtship than chain saw.

Thu: BOE meeting: will leave rate unchanged

Fri: US-IRAN Deal official signing ceremony in Switzerland.

Whole Week: Central banks week, rate decisions and Iran MOU details

FOMC RATE DECISION IMMINENT

FED LEAVES RATES UNCHANGED

*FED: NINE OF 18 FOMC PARTICIPANTS PENCIL IN 2026 RATE HIKE

*FED REMOVES STATEMENT REFERENCE TO ADDITIONAL RATE ADJUSTMENTS

FOMC STATEMENT

FED SAYS COMMITTEE WILL DELIVER PRICE STABILITY

TRADERS PRICE IN ADDITIONAL FED TIGHTENING AFTER DOTS REVISED

*April FOMC statement: ~ 350 words

*June FOMC statement: ~130 words

FED'S CHAIR WARSH DID NOT SUBMIT A DOT ON THE SEP DOT PLOT

WELCOME TO FED WARSH'S REGIME CHANGE

Fed Press Conference Takeaways:

Tldr: THE GOOD NEWS FOR YOU IS WE HAVE A TASK FORCE FOR THAT

1. Short statements.
2. No forward guidance.
3. Warsh said financial markets perform best when they react to incoming data not on "how will the Fed react to that incoming data"
4. FOMC ‘unambiguously and unanimously’ going to deliver 2% inflation. We'll fix five years of misses on inflation, warsh said
5. Warsh announced task forces that will look into several areas and make recommendations for improvements by yod: communications, the balance sheet, use of data sources, productivity and jobs, and the Fed’s inflation framework
6. On Balance Sheet: will review the benefits and risks of the current ample reserves regime and the composition of the balance sheet.
7. On Data Sources: will evaluate new info sources & consider methodological changes to improve data-gathering.

*WARSH EXPERIENCES WORST ‘FED DAY’ S&P 500 PERFORMANCE FOR A NEW CHAIR SINCE 1994

*TRADERS FULLY PRICE IN TWO FED RATE HIKES BY THE END OF 1Q 2027

MIRCOSTRATEGY'S STRC PREFERRED STOCK LOSES PEG, DROPS TO RECORD LOW $85: INFINITYHEDGE

TWITTER (X), A COMPANY OWNED BY WORLD'S FIRST TRILLIONAIRE DOWN AGAIN: INFINITYHEDGE

MICROSTRATEGY DROPS BELOW $100, FIRST TIME SINCE MAR 2024: INFINITYHEDGE

STRATEGY'S STRC PREFERRED STOCK DROPS TO RECORD LOW $75

STRATEGY DROPS BELOW $90: INFINITYHEDGE

MICROSTRATEGY'S ENTERPRISE MNAV FALLS BELOW 1

[INFINITYHEDGE]

JAPANESE YEN DROPS TO NEAR 40-YEAR LOW

YEN REACHES 161.97 AGAINST US DOLLAR, WEAKEST SINCE 1986

MEITUAN DEBUTS CHINA’S BIGGEST AI MODEL TRAINED ON LOCAL CHIPS

- 1.6T parameters · MoE with ~48B active · 1M context
- Both the full training run and the large-scale deployment are built entirely on AI ASIC superpods.
- LongCat-2.0 is pre-trained on over 50K AI ASICs (No GPUs).
- it used the Huawei Collective Communication Library (HCCL) to improve training stability. HCCL is a chip-to-chip communication system similar to the Nvidia Collective Communication Library (NCCL).
- Competing models like DeepSeek-V4-pro relied on domestic chips only for inference, Meituan utilized domestic hardware for both inference and the highly intensive pre-training phase.
- Because the domestic accelerators possess less per-device memory than an Nvidia H800 (80 GB), Meituan address this challenge along two dimensions: parallelism strategy and memory management.

More: https://longcat.chat/blog/longcat-2.0

OPENAI ENGINEERS DEVELOPED NEW OPTIMIZATIONS CUTTING MODEL INFERENCE COSTS BY MORE THAN 50%, REQUIRING LESS NVIDIA CHIPS: THE INFORMATION

OPENAI MAY USE THESE SAVINGS TO LOWER API PRICING OR INCREASE USER QUERY LIMITS

US JUNE NONFARM PAYROLLS RISE 57K M/M; EST. +110K
US JUNE UNEMPLOYMENT RATE 4.2%; EST. 4.3%
US JUNE AVERAGE HOURLY EARNINGS RISE 0.3% M/M; EST. +0.3%
US JUNE AVERAGE HOURLY EARNINGS RISE 3.5% Y/Y; EST. +3.5%

US JUNE TWO-MONTH PAYROLL NET REVISION SUBTRACTS 74k

TRUMP: COUPLE OF GUYS WENT SHORT MARKET, POOR BASTARDS ARE IN BIG TROUBLE, THEY'RE GETTING WIPED OUT

CHINA MULLS RESTRICTING OVERSEAS ACCESS TO TOP AI MODELS

ALIBABA, BYTEDANCE AND Z.AI ATTENDED MEETINGS WITH AUTHORITIES, SOURCES SAY

PARTICIPANTS DISCUSSED LIMITS ON BOTH CLOSED-SOURCE AND OPEN-WEIGHT AI MODELS

OFFICIALS RAISED POSSIBILITY OF NEW MEASURES TO RESTRICT WHO CAN FUND DOMESTIC AI STARTUPS

DEEPSEEK IS DEVELOPING ITS OWN AI CHIP DESIGNED FOR INFERENCE, A PUSH TO REDUCE RELIANCE ON NVIDIA AND HUAWEI CHIPS, HIRING CHIP-DESIGN ENGINEERS FOR THE PROJECT - SOURCES

DEEPSEEK'S EFFORT TO DEVELOP ITS OWN AI CHIP FOLLOWS A BROADER INDUSTRY TREND: OPENAI LAST MONTH UNVEILED JALAPENO, ITS FIRST CUSTOM INFERENCE CHIP, DEVELOPED WITH BROADCOM, WHILE ANTHROPIC HAS BEEN WEIGHING BUILDING ITS OWN AI CHIPS WITH SAMSUNG

CHINA’S AI LAB ZIPHU WEIGHS CUSTOM CHIP AS DEMAND FOR ITS GLM MODEL SOARS. ZHIPU IS ALSO LIKELY TO MANUFACTURE THE CHIP AT CHINESE FOUNDRIES - THE INFORMATION