August 9 -15, 2025
Big news. This is certainly the statement of the US Treasury Secretary Scott Bessent: “We will consider the assets of allies (Japan, Korea, Europe) as an American sovereign investment fund, which, at the discretion of the President of the United States, will be used to revive American industry.”
One can interpret these words as much as one wants, including explaining that it was a slip of the tongue (like Trump, in his phrase that he was “going to Russia”). But this no longer makes any sense, Bessent, in fact, confirmed that the old financial model, which can be conditionally called “Reagan’s”, is dead. And the guarantees that it gave to the owners (do we need quotation marks here now?) of assets, died along with it.
And any discussions about medium-term (not to mention long-term) economic or business plans must take this sad circumstance into account. Nothing can be done about it.
Макроэкономика. Подошли очередные китайские данные за июль. Они разочаровали: импульс казённых стимулов слабеет, все ждут новых вливаний.
Macroeconomics. The latest Chinese data for July has arrived. They are disappointing: the momentum of government stimulus is weakening, everyone is waiting for new injections.
Investments in fixed assets +1.6% per year, excluding Covid, by far the worst dynamics in history; the nominal value of investments (in yuan) is at the level of 10 years ago:
Рис. 1
Рис. 2
Annual growth in industrial production (+5.7%) is the weakest in 8 months:
Рис. 3
The same picture is with retail (+3.7% per year), and there is a collapse in oil products (-8.3% per year):
Рис. 4
Unemployment (5.2%) is at its highest level in 5 months:
Рис. 5
Only the decline in prices for new buildings (-2.8% per year) continues to slow down. But even there, the monthly dynamics remain negative, indicating weak demand:
Рис. 6
New bank loans in China are -50 billion yuan per month. A historical anti-record; this figure has actually only gone into the negative once before (20 years ago):
Рис. 7
Рис. 8
Overall, loans in China are +6.9% per year, the worst dynamics in all 27.5 years of observations:
Рис. 9
UK goods trade balance close to January 2022 low:
Рис. 10
India’s CPI (consumer inflation index) (+1.55% per year) is only 0.01% off its record low of July 2017:
Рис. 11
And wholesale food prices in India (-6.3% per year) have already hit a new record low in 20 years of data collection:
Рис. 12
Deflation. A clear sign of industrial decline in the situation of refusal to issue.
US PPI (consumer inflation index) +0.9% per month. This has not happened since March 2022:
Рис. 13
Even more noticeable is the same size jump on the price chart without food and fuel:
Рис. 14
Even prices for the total volume of industrial goods have increased:
U.S. consumer sentiment from the University of Michigan has fallen slightly:
Рис. 16
The current situation assessment has deteriorated particularly sharply, returning to record lows:
Рис. 17
Inflation expectations are rising again:
Рис. 18
Especially 5 year olds:
Рис. 19
This is about the dispute between Trump and Powell. The head of the Fed has reasons not to lower the rate, which we have written about many times.
The yield chart for 30-year German government bonds continues to update 14-year highs against the backdrop of the end of the ECB easing cycle and the likely impact of Trump’s tariffs on prices around the world.
The number of workers on the payroll in Britain has been declining monthly for 6 months in a row:
Рис. 20
And the unemployment rate has been at its highest since January 2017 (excluding Covid):
Рис. 21
The Bank of Australia cut its rate by 0.25% to 3.60%. The Bank of Norway made no changes to its policy.
Main conclusions. The structural crisis continues. Those who print money have inflation, those who do not print have deflation. And the standard of living of the population is falling, as is the real (not nominal) GDP in both places.
In addition (this is not macroeconomics), Pavel Ryabov’s assessment of the state of affairs with artificial intelligence:
“A turning point in AI
Exponentially growing investments in AI for continuous expansion of computing power with limited growth in model performance (the principle of linear scaling no longer works) clearly indicate that the industry has reached a dead end, without even having time to accelerate.
With revenues in the tens of billions, the required investments reach hundreds of billions (inadequately high salary requirements for AI specialists, reaching millions per person, electricity costs in gigawatts, super-expensive equipment, marketing).
The most interesting thing is that after a year, not a single AI company has presented either a plan for monetizing investments or an AI development plan, remaining at incredible losses in the AI direction. There is a panicky tightening of functionality in all spectrums and experimentation in an attempt to find breakthrough niches.
Everyone seems to understand that AI is cool, can be useful and has great prospects, but no one understands in which direction everything is moving.
Currently, AI companies are investing hundreds of billions in an attempt to capture the audience, creating a loyal base from which, according to the idea, they will squeeze cash in the future, but this plan has many flaws:
🔘Everything that requires a continuous innovation race with an unpredictable result and a negative return (it is impossible to recoup such expenses) wins.
🔘The audience quickly migrates from one model to another as soon as competitors have a more effective solution.
🔘The pace of innovation is so high that any delay, planning error or decrease in the rate of investment immediately knocks you out of the race for leadership in AI – there is no room for error.
🔘The breadth of competitors presented leads to a split focus, dispersing the audience, reducing the concentration of cash flow.
The highest speed of innovation and intra-industry transformation, extreme competition force businesses to continuously increase investments in order to “stay in the saddle without falling to the side of the AI race.”
On the one hand, this increases the entry threshold, filtering out random participants, and on the other hand, it makes the business unstable, extremely risky and guaranteed to be unprofitable.
The current business scheme is unviable against the background of fundamental architectural limitations of modern LLMs – there is no way to recoup investments. Especially against the background of free and open Chinese analogues.
Now there is a leveling of competition, when it is difficult to become a leader when releasing the next model and obviously, the foreseeable development trend will be directed towards polishing models if there is no possibility of a qualitative increase in productivity (increasing stability, accuracy, expanding functionality, etc.).
In terms of return on capital, this is the craziest investment in the history of mankind. No technology has ever cost so much with such a stunning speed of development and with a limited space for monetization, which implies a high speed of structural transformation (current leaders can become outsiders tomorrow at the first mistake).
The collapse of the AI bubble is inevitable, since there is no scenario for monetizing the accumulated trillion-dollar investments, given the structure and specifics of the industry. The development and integration of AI will now continue forever, but with significantly muted hype.”
Or, a brief summary: AI as a basis for unlimited market monetization (that is, for sterilization of emission liquidity) ceases to operate. With all the consequences.
And finally, we wish all our readers a good rest on the weekend and a sober look at reality during the work week.