Investor's Business Daily is the only outlet (so far) to headline revenue vs expectations.
On South Korea's market the stock traded down 9.61%, with the KOSPI down 5.98%.
From IBD, July 28:
South Korean memory-chip maker SK Hynix (SKHY)
late Tuesday missed sales estimates for the second quarter. But its net
profit was greater than its revenue in the period due to investment
gains. SKHY stock advanced after the report.
The company said its Q2 revenue rose 257% year over year to 79.3
trillion won ($54.4 billion). However, analysts polled by FactSet had
expected second-quarter revenue of 84 trillion won ($57.6 billion).
SK Hynix said its operating profit surged 557% to 60.5 trillion won
($41.6 billion). Meanwhile, its net profit rocketed 1,242% to 93.9
trillion won ($64.5 billion), fueled by gains related to investment
assets, which it didn't specify. Last month, SK Hynix completed the sale
of its stake in Japanese Nand flash memory maker Kioxia.
The company did not provide per-share earnings in its news release.
The quarterly report was the company's first since its U.S. trading debut on July 10.
The IPO was priced at $149 per American depositary receipt. Every 10
ADRs traded on the Nasdaq represents one common share of SK Hynix stock
listed in South Korea. The company's ADR climbed as high as 194.80 on
July 14 before retreating. On Tuesday, it fell 9% to close at 130.17
during a broad semiconductor stock sell-off....
This raises the question of demand destruction if the ever-increasing price of memory is not enough to fuel the top-line hopes and dreams of analysts and investors despite (or because of) the astounding profitability of the chips.
Agence France-Presse caught some of this interplay in their reporting.
From AFP via France24, July 29:
SK hynix posts 1,200% net profit boost on AI chip boom
Seoul (AFP) – South Korea's SK hynix said Wednesday second-quarter net profit soared a whopping 1,242 percent year-on-year, driven by the artificial intelligence industry's explosive demand for its advanced memory chips.
SK hynix is a specialist supplier of high-bandwidth memory chips to
US industry behemoth Nvidia, and a pillar of South Korea's tech-led
economy.
The global race to build data centres hosting AI
infrastructure has seen a meteoric rise in the firm's fortunes, despite
concern the sector may be overvalued in a market bubble.
Wednesday's
earnings figures, including quarterly net profit of 94 trillion won
($64 billion), were described in a statement by the Icheon-headquartered
firm as "an all-time high quarterly performance".
"We are aware
of concerns that AI infrastructure investment might be slowing down,"
marketing chief of the AI microchip division Park Joon-deok said on a
call with investors and reporters.
He cited jitters over firms
exploring data centre rental -- rather than construction -- and the
emergence of new high-efficiency AI requiring a lower memory taskload.
"We
view these developments not as a scaling back of AI investment, but
rather as a process of maximising the utilisation of the massive AI
infrastructure built to date and accelerating its monetisation," he
said....
Here's CNBC with the overview of the report, July 28/29:
SK Hynix shares tank as exponential earnings growth fails to satisfy AI-charged expectations
SK Hynix hits another record milestone on AI demand.
Second-quarter profit and revenue both missed estimates.
Revenue more than tripled from a year earlier.
SK Hynix
shares nosedived on Wednesday as exponential second-quarter earnings
and revenue growth still failed to satisfy analysts’ supercharged
expectations for an artificial-intelligence darling.
Here are SK
Hynix’s second-quarter results compared with LSEG SmartEstimates, which
are weighted toward forecasts from analysts who are more consistently
accurate:
Revenue: 79.32 trillion won ($54.55 billion) vs. 84 trillion won expected
Operating profit: 60.54 trillion won vs. 64 trillion won expected
After falling as much as 15% earlier in the session, the stock trimmed its losses to close 9.6% lower on Wednesday.
Revenue
jumped 257% year on year in the quarter ended June from a year earlier,
while operating profit soared nearly 557% year on year, the company’s
release showed.
Compared with the previous quarter, revenue increased 51% while operating profit gained 61%.
The
company said it saw sustained demand growth from expanding AI
infrastructure investments, while high-performance products for AI
servers led price increases that set a fresh record....
When you have perils that could trigger payouts that are extremely large combined with a relatively new class of insureds (hyperscale data centers), the actuaries chew their pencils before saying yes.
From Artemis, July 10:
Insurance capacity constraints for data centres to drive ILS use: S&P
With capacity constraints likely to limit the insurance industry’s
ability to fully insure hyperscale data centre projects, given their
scale and complexity, a new report from S&P notes that this will
drive greater use of self-insurance through captive insurers, and
potentially, alternative capital such as insurance-linked securities
(ILS).
The rating agency’s latest report suggests that rising demand for
data centre insurance coverage could generate $10 billion in new
premiums in 2026, highlighting how significant the data center market
could become and the scale of the opportunity for the global
re/insurance industry.
At the same time, S&P also indicated that annual investment in
data centres could surpass $300 billion by 2027, highlighting the
industry’s explosive growth and the growing insurance requirements
associated with large-scale infrastructure development.
For background, S&P pointed out that insurance coverage limits of
$5 billion to $10 billion are usually needed for the biggest
infrastructure construction projects, like bridges or tunnels. By
comparison, some hyperscale data centers are estimated to represent
total insurable values of $10 billion to $30 billion for construction
alone....
Bringing to mind a 2023 post, "So Why Were The Chinese Plundering British Shipwrecks Off The Coast Of Malaysia?". *
From Tom's Hardware, July 28:
Buy. Scan. Destroy. Repeat.
Having contributed to the growing shortage of memory and storage, AI companies seemingly have a new target in their sights: humanity's literary history. A recent investigative report from 404 Media reveals that these companies are reportedly purchasing millions of secondhand books through intermediaries to source high-quality training data for their AI models, avoiding public backlash.
AI relies on vast
amounts of data to advance, but not just any data. It has to be
high-quality data. The problem is that mediocre AI-generated content,
commonly referred to as "AI slop," has proliferated across the Internet.
This type of content contaminates the data pool and is
counterproductive for AI to train on. As a result, leading AI companies
have turned to human-authored sources for knowledge, specifically print
sources that predate 2022 and are more likely to contain original,
uncontaminated content.
There
is precedent for AI companies turning to physical books for training
AI. For instance, Anthropic, one of the leading AI companies involved in
a lawsuit, reportedly invested millions of dollars in extracting
information from countless printed books to build its Claude AI models and then destroying them. The
company bought books from Better World Books. Although the court
decision affirmed that using books for AI training falls under fair use
in copyright law, Anthropic faced a staggering $1.5 billion fine
for maintaining a repository of seven million pirated books that
infringed the copyrights of authors and publishers. Similarly, a
coalition of publishers recently filed a lawsuit against Google, accusing the tech giant of allegedly and illegally using millions of copyrighted books to develop its Gemini AI models.
ISBNdb, an online database that reportedly has over 111 million
cataloged books, has been a long-favorite platform for booksellers,
libraries, and distributors to sell books. With the explosion of the AI
industry, ISBNdb has pivoted its business to offer specialized services
to bulk-purchase books for AI companies. According to 404 Media, the
orders range from 1,000 copies to as many as one million books in a
single transaction....
....MUCH MORE *And the shipwrecks in Malaysian waters recollection, July 7, 2023:
Here's the rest of the story, from Ed Conway's Material World substack, June 10:
The Eerie Story of Low Background Steel Bandits are plundering metal from old shipwrecks, in search, it seems for the world's rarest, and strangest, metal.
The other week a Chinese vessel was detained by Malaysian authorities off the coast of Johor, under suspicion of having plundered old WWII era shipwrecks in the region.
HMS
Repulse and HMS Prince of Wales, both of which sank in Malaysian waters
in 1941, have had large sections of their bodies and armaments stolen
in this way. The practice has been going on for years
- these raiders targeting shipwrecks which are also effectively war
graves - but this was one of the rare occasions when someone was
seemingly caught in the act.
But why, you might be
wondering, would anyone go to these lengths to obtain scrap metal. The
price of steel hardly merits this kind of effort and risk. So what are
these raiders really after? Gold? Silver? Stolen artworks?
The answer, it turns out, is far more interesting: a very, very rare form of metal. Something called “low background steel”.
Steel
itself is one of the cheapest types of metal (an alloy technically) but
the type of steel we’re talking about is one of the rarest substances
in the world. For low background steel doesn’t contain radionuclides,
traces of radiation such as cobalt-60. These trace amounts don’t matter
for most uses but when you’re making products highly sensitive to
radiation - eg special scientific equipment or Geiger counters - you
need this steel.
And here’s the thing: all steel made since 1945 contains radionuclides....
There are three developments to note today. First, the lack of hostilities in the Middle East has oil prices continue to unwind this month’s surge. This has also taken some pressure off major bond markets. Second, reports that China has started to mass produce its own chipmaking tools and has developed DUV lithography tools that etch chip patterns onto silicon weigh on the equities in this space, which had already been hit with profit-taking. South Korea’s Kospi fell nearly 11% today and Taiwan’s Taiex slumped a little more than 4.5% today. The Nasdaq is poised to gap lower today.
The third development is the heightened speculation of a rate hike by the Federal Reserve tomorrow. On July 16, the Fed funds futures were pricing in about a 10% chance of a hike and now about a 33% chance. The US dollar itself is trading with a clear firmer bias. It is made a new high for the month against several pairs, including the euro, sterling the Australian dollar....
A report emerged Monday that China has produced a key chipmaking tool, entering a market that has long been dominated by ASML.
However, analysts told CNBC ASML is unlikely to be heavily impacted as lots of questions remain about China’s technology.
These questions include performance versus ASML and China’s ability to scale production.
ASML’s
stock fell Tuesday after a report that China is mass-producing a
critical tool that the Dutch tech giant has long held a monopoly over.
The decline came amid a steep sell-off
in global semiconductor stocks as investors continued to grapple with
uncertainty about the sector. ASML was last trading down 1.8%, but the
stock is up over 123% this year.
Analysts
told CNBC that while reports on China entering a market that ASML
dominates may raise some concerns, the developments are unlikely to
shake the Dutch company’s dominance, adding there are some big caveats
to the story.
“I would take this with a pinch of salt as what
[China does] could be limited to the very low end,” Stephane Houri, head
of equity research at ODDO BHF, told CNBC.
What happened? The Information on Monday reported that an unnamed Chinese company has begun manufacturing an immersion deep ultraviolet lithography (DUV) machine, citing people familiar with the matter.
Why did it spark a sell-off? Investors are concerned that if China continues to build out its homegrown semiconductor technology, it could cut off some of the biggest U.S., European and other Asian firms from the huge market.
An immersion DUV machine is a tool that is used to etch circuit patterns
into silicon wafers. It is a critical part of the semiconductor
manufacturing process that is purchased by foundries such as Taiwan Semiconductor Manufacturing Co. (TSMC) and Intel.
ASML
dominates the market for DUV and EUV lithography tools. No other
company has been able to replicate what ASML does, which is why it is a
big deal that a Chinese firm has reportedly done so. Nevertheless,
questions remain as to what impact the latest developments will have on
the company.
Scale and performance questions When it comes to the performance of China’s DUV machines, semiconductor manufacturers focus on a term known as “yield” which refers to the number of usable chips that come out of the process. All foundries aim for maximum yield.
It is currently unclear whether a Chinese
manufacturer using the homegrown DUV machine will deliver a chip yield
close to or above that of a machine from ASML. If the yield is not close
to what ASML machines can provide, that might hamper the adoption of
China’s machine....
Nvidia is down $10.00 (-4.84%). That's a lot of billions of market capitalization.
From 24/7 Wall Street, July 27:
NVDA fell 5% and AMD dropped 8% after the WSJ reported NVIDIA may guarantee $250 billion for OpenAI's data-center buildout.
Dell slid 4% on AI-server demand fears while Oracle has fallen 20% over the past month on overlapping data-center exposure.
NVIDIA guides $91 billion in Q2 revenue and AMD's data center revenue grew 57% YoY, keeping the fundamental bull case intact despite today's selloff....
Former Citadel engineer Alex Alifimoff has joined the rival hedge fund
as head of AI weather amid growing demand for alternative data
Steve Cohen’s multi-strategy hedge fund Point72 has hired an AI expert to help forecast weather patterns for its money managers.
Alex
Alifimoff, who previously worked at Citadel as an engineer, has joined
Point72 as head of AI weather in the firm’s global macro team, according
to people familiar with the matter.
In his new role, Alifimoff
will build AI-based models to forecast weather patterns to help trading
teams, including macro and commodity portfolio managers, the people
added. He will be based in the firm’s Stamford office.
For years,
hedge funds have looked for alternative data sources in a bid to steal a
march on their competition, turning to everything from satellite
imagery to social media posts in a bid to better predict economic
patterns. However, as this data has become more available in recent
years, AI experts are now at a premium to stay ahead of the curve.
Hedge funds are offering lucrative pay packages of up to $1m to attract top weather modelling experts, Bloomberg reported last year. Traders, mainly across commodities desks, require weather expertise to predict price fluctuations....
And Cargill has the data advantage of over 100 years of proprietary observations, a treasure trove of AI training material.
So while the speculative mice can skim some grain/cocoa/almond etc. profit from the commercials, the commercials will, and must, win overall. If they don't the entire game stops.
Iran's Supreme Leader Mojtaba Khamenei is not in contact with any person or institution, says his father-in-law.
“For the time being, due to certain issues, he is not in contact with any person or institution, but God willing, he is healthy,” Gholamali Haddad-Adel said, according to ILNA news.
He was responding to a question about Khamenei’s absence from the funeral ceremony for his father, former Supreme Leader Ali Khamenei.
Asked about reports concerning Mojtaba Khamenei’s alleged economic activities before becoming the supreme leader, Haddad-Adel dismissed them as false and said he had been a cleric whose only income was the standard stipend paid to seminary students.
Although it is easy to get caught up in the "Dear Diary, I am so pretty. And popular!" stuff (hundreds of the things), on his tweet announcing the drop the Senator leads with one of the more disgusting entries:
1/9 — My investigation uncovered that Anthony Fauci kept a diary. What he wrote privately and what he told the country are two different stories.
Today I'm releasing his entries from December 2019 through December 2022. 🧵 pic.twitter.com/58LvmeYkCN
3/9 Feb. 1, 2020. The call that produced Proximal Origin. Twelve scientists.
Only 2 concluded it was natural. Per Fauci's own notes, "the rest felt that deliberate insertion was possible." This was never a 50/50 room. pic.twitter.com/vFg79if4oB
Tony's Diary: What He Wrote Privately vs. What He Said Publicly
July 25, 2026
– Fauci’s personal diary reveals that as early as January 31, 2020, he
was told by top virologists that the virus’s furin cleavage site raised
real doubts about a natural origin, with about half the scientists on
that initial call believing it looked “constructed.” Despite this early
private uncertainty, Fauci spent the following years publicly and
confidently dismissing the lab leak theory while insisting the science
was settled. The diary also documents his private frustration and
defensiveness over gain-of-function funding questions tied to EcoHealth
Alliance and the Wuhan Institute of Virology — even as he told Congress
under oath that no such research had occurred.
Fauci's Side Hustle: How NIH Lined Dr. Anthony Fauci's Pockets with Prize Money
July 24, 2026
– Records released today reveal that Dr. Anthony Fauci, the
highest-paid federal official in the U.S. government, stacked up cash
prizes worth hundreds of thousands of dollars during COVID, with NIH’s
number two official serving as his personal awards manager and his
taxpayer-funded staff writing the nominations for him. Some of the
prizes came from American institutions, but the largest came from a
foreign institution that wired him $900,000.
NEW:
Federal Officials Intercepted Undeclared Biological Materials Tied to a
U.S. – China Coronavirus-Research Network as Early as 2018
July 24, 2026
– Undeclared coronavirus materials moved between U.S. and Chinese labs
as far back as 2018 — years before COVID-19. A courier was caught twice,
three years apart…first with a few hidden vials, then with 132 more and
a laptop that wasn’t his.
Senator Paul, in his position as the Chairman of the U.S. Senate Committee on Homeland Security and Governmental Affairs will be convening the hearing to take Fauci's testimony on Wednesday July 29 at 8:30AM ET.
Here are the additions to the background information/documents
There is one major driver in the capital markets today: A sharp drop in oil prices as the US and Iran have inexplicably not attacked for the past three nights. There is speculation in the press that the US decision was partly motivated by concerns over diminishing supplies of air defenses (e.g., Patriot interceptors) after the two-week assault, but the Trump administration played this down. Crude oil prices are off 7-8% and this encouraged further unwinding of the rise in bond yields seen last week. Equities are rallying today and China’s CXMT chip maker IPO was launched in Shanghai today and it is the largest list company after today’s surge. It is big week for US tech earnings (MSFT and Meta on Wednesday and Apple and AMD on Thursday).
The dollar was marked down initially early Asia Pacific trading but has not made much further progress. The euro continues to struggle to sustain gains above previous support (~$1.14), sterling has not distanced itself convincingly from $1.33 support. The Australian dollar is struggled to sustain the recovery above $0.7000. The US dollar is above CAD1.4100. The dollar pulled back to about JPY163.35 but remains within striking distance of the 40-year high set last week near JPY164....
Moonshot
AI is poised to make its Kimi K3 model available for public download,
expanding its reach and influence in the global open software community
at a time of growing US concern about Chinese encroachment into the top
echelons of AI development.
The
Beijing-based company is due on Monday to release the model’s weights,
which are used to steer artificial intelligence systems toward answers.
That will enable developers to download, tweak and host the technology
freely in a move to broaden the user base of Moonshot’s AI. Founder Yang Zhilin has said he wants to win users by focusing on openness and greater availability than competing US proprietary systems.
Kimi
K3’s initial announcement and benchmark performance triggered a selloff
in AI-related stocks this month, as it demonstrated Chinese-made models
are closing the gap on American leaders OpenAI and Anthropic PBC.
Much like DeepSeek’s release in early 2025, the K3 debut was taken as
evidence that Chinese companies can keep pace on a smaller budget and
with fewer cutting-edge resources than rivals with full access to Nvidia Corp. AI accelerators.
Monday’s
release may boost business for cloud AI computing providers who
specialize in hosting third-party open models. Made-in-China AI models
have surged in global popularity in recent times, led by DeepSeek, as they offer more affordable rates and comparable performance to US offerings.
Daily
sales at Moonshot have surged at least sixfold since K3’s debut,
Bloomberg News has reported. The Beijing-based startup is riding that
momentum as it seeks a new funding round at a $50 billion valuation ahead of a potential Hong Kong initial public offering as soon as this year.
Open-weight
large language model adoption is inflecting, but monetization will
hinge on distribution through hyperscale cloud providers. Amazon AWS
Bedrock, Microsoft Azure Foundry and Google Vertex AI have yet to
support these open-weight Chinese models, even as their token-volume mix
reached 68% after the launch of Moonshot AI’s Kimi K3 and Z.AI’s
GLM-5.2. — Mandeep Singh and William Tong, BI analysts Click Kimi K3 Token Volume Gain Needs Hyperscaler Cloud Distribution for the full research
Moonshot’s
success has thrust it into the US-China tech crosshairs....
CXMT held a 7.67% share of the global DRAM market in 2025, according to its IPO prospectus.
The Hefei-based company had raised 57.92 billion yuan ($8.6 billion) in its IPO.
The company plans to use the IPO proceeds mainly for mass-producing memory wafers.
Shares of Chinese chipmaker Changxin Technology Group rose about 470%
Monday as they debuted on Shanghai’s tech-heavy STAR Market.
The
Hefei-based company raised 57.92 billion yuan ($8.6 billion) after
pricing its IPO at 8.66 yuan per share. Shares surged to 49 yuan apiece
on open.
Based
on sales figures for the fourth quarter of 2025, CXMT held a 7.67%
share of the global DRAM market in 2025, according to its IPO
prospectus. DRAM chips are used in electronic devices ranging from
smartphones to servers....
Project would be one of the largest AI computing hubs and involve power controlled by the U.S. government
NvidiaNVDA -0.92% is in talks to provide a roughly $250 billion backstop for OpenAI
as part of a massive data-center project, one of the most ambitious
financial transactions yet in America’s artificial-intelligence boom.
The guarantees from Nvidia would help the ChatGPT maker lease a 10-gigawatt project that SoftBank’s 9434 2.48% energy subsidiary is developing in southern Ohio, people familiar with the matter said.
In total, the project could cost more than $500 billion, including the chips that would go inside the data centers. It would be the largest data center project announced to date.
The power for the project is controlled by the U.S. government and funded separately by Japan under a recent trade deal. Commerce Secretary Howard Lutnick is involved in deciding who will get the power, some of the people said.
Nvidia’s backing would allow the data center developer, which is owned by Japanese billionaire Masayoshi Son’s investment firm SoftBank, to raise debt at more favorable terms than it could if OpenAI had no financial backer, since OpenAI has no investment-grade credit rating as an unprofitable private company. The AI company has been in advanced talks to lease the site for several weeks, people familiar with the matter said.
OpenAI is among the companies that has shown most interest in the site, while Anthropic, Microsoft and Google have also spoken to Lutnick about it in recent weeks, one of the people said.
AI companies are desperate to secure chips and power needed to run their models, making megaprojects that have government support or guaranteed power increasingly attractive.
Under the arrangement, Nvidia would guarantee a series of financing vehicles intended to make lenders feel more confident that the funding behind the project is secure, the people said. Terms haven’t been finalized and the deal could fall apart.
The $250 billion guarantee would cover the data center lease and debt needed to fund its build out, but not the Nvidia chips that would go inside it. Nvidia, which has invested $30 billion in OpenAI, is also discussing a deal to finance the chip purchase for OpenAI, which could total $350 billion, people familiar with those discussions said. Such circular funding arrangements have caused concerns that the industry is vulnerable if investor sentiment shifts or growth slows for AI companies....
We didn't get any further information on the immense Ohio natural gas power plant
in last night's State of the Union message but as a possible
consolation prize for the hyper-concentrated mini-portfolio we see this
from Construction Review, February 17...
Everything other than the fact it is Japanese money funding the beast and a Japanese company (SB Energy sub. of SoftBank) overseeing the project and Japanese companies expressing interest in developing same, a mention of American corporate participation could be rocket fuel for a couple of our favorite names.
First up, Barron's last week, with the overview...
***
...9.2 gigwatts would make it the second largest electrical plant of it's type in the world. Both Mitsubishi Electric and Hitachi manufacture utility-scale natural gas turbine generators but if there is room for an American company it would have to be GE Vernova. Plus transformers and transmission lines and Quanta Services is the go-to.
Turkey’s geography, infrastructure, and regional influence position it to reshape Europe’s energy security through a new overland corridor linking four major energy basins.
For decades, Turkey has pitched itself as an energy bridge between
East and West, North and South. If Turkey can finally deliver to the
world a solution to overcome its reliance on the Strait of Hormuz, the
economic and strategic prize will be enormous.
The Four Seas Initiative Could Redraw Global Energy Routes
The Four Seas Initiative identifies
Turkey as the keystone in the infrastructural response to the Strait of
Hormuz crisis. The idea is simple. The four seas: the Persian Gulf, the
Caspian Sea, the Mediterranean, and the Black Sea are among the most
important energy geographies in the world. Yet they remain connected
largely through bilateral deals and infrastructure shaped by old
political divisions—not to mention being vulnerable to war and blockade.
The Four Seas Initiative proposes an overland energy and infrastructure
corridor linking these four basins through Syria and Turkey, carrying
Gulf, Iraqi, Caspian, and eastern Mediterranean energy toward lucrative
European markets.
For Turkey, this marks the arrival of
the role Ankara has long sought: not merely to be a country through
which energy passes, but to become the operator, rule-setter and
commercial center of a continental redistribution system.
This agenda-setting role is due to the
absolute scale of the project. At full operational capacity, the Four
Seas corridor is projected to move an estimated 3 million to 4 million
barrels of oil per day and 40 billion to 50 billion cubic meters of gas
per year toward Mediterranean and European markets. As a network, it
will compete in scale with Nord Stream. It would put the Syria–Turkey
corridor in the first rank of global energy routes, combining Gulf oil,
Iraqi exports, Caspian gas and eastern Mediterranean supply into a
single overland system....
China's“deflation-is-over” narrative is getting louder, but the foundations are still shaky.
Consumer prices rose 1.2% year-on-year in May, while producer prices
jumped 3.9%, lifted by higher costs for energy, semiconductors and
metals. To many economists, this is the clearest sign yet that the 2025
deflation scare is giving way to reflation.
But Japan's long struggle shows how stubborn deflationary psychology
can be. And it's far from clear that Beijing is delivering the
structural reforms needed to ensure China's weak‐price era is truly
ending.
Two reforms stand out - and neither is being pursued with urgency.
First, resolving the deep housing crisis, which increasingly resembles
Japan's 1990s bad‐loan spiral. Second, building a real social safety net
so 1.4 billion citizens feel confident enough to spend rather than
hoard savings.
These priorities are tightly linked. With roughly 70% of household
wealth tied to property, stabilizing the real‐estate market across
China's 70 largest cities is essential for reviving consumption and
sustaining 4.5%- 5% growth.
But the longer Xi's government acknowledges these pressures while
avoiding decisive action, the more a deflationary mindset takes hold -
and the harder it becomes to shake.
Japan is the cautionary tale. Even as the Bank of Japan prepares to
lift rates to 1% next week - the farthest from zero in more than three
decades - deflationary undercurrents still run through the economy.
On paper, Japan looks like it has finally escaped its low‐price trap.
The BOJ expects inflation to reach 2.8% this year, suggesting reflation
is taking hold. But beneath the headline, real wages remain negative,
with pay packets consistently trailing price gains and domestic demand
weakening as a result.
The result is a slow‐burn form of stagflation, and Tokyo has yet to
deliver the structural reforms needed to close the gap between rising
prices and stagnant household incomes.
“For the Japanese economy to fully break free from its long-standing
deflationary mindset,” says Toshihiro Nagahama, economist at the
Dai-ichi Life Research Institute,“it's imperative for the government and
the central bank to align, articulate their risk assessments, maintain
honest and transparent dialogue with financial markets, and resolutely
execute bold, long-term growth investments.”....
From the truly astounding World Politics substack (don't let the name of the blog fool you), March 20, 2026:
Kautilya made Machiavelli look naive
In the closing decades of the fourth century BC, in the city of
Pataliputra on the Ganges plain, a man named Kautilya composed a text on
statecraft, economics, and political strategy whose ruthlessness,
comprehensiveness, and analytical intelligence have no parallel in the
political literature of the ancient world.
The Arthashastra
(the Science of Statecraft) is a treatise organised into 15 books –
with subjects ranging from the organisation of the king’s daily schedule
and the architecture of the royal palace to the administration of
agriculture, mining, and commerce; the management of the treasury and
the army; the conduct of diplomacy and espionage; the waging of war; and
techniques for eliminating rivals.
Niccolò Machiavelli’s Prince
is the Western political tradition’s most powerful exposition of
strategic realism; it amounts to about 70 pages. Kautilya wrote several
hundred, addressed them to a far broader range of political problems
with greater analytical precision, and did so eighteen centuries before
Machiavelli was born. The consistent verdict of scholars who have read
both carefully is that, for all his genius, Machiavelli worked in a
shallower analytical tradition and addressed a narrower set of problems
than the Indian whose work he never knew existed....
Nvidia (NVDA.O) and South Korea's SK Group on Friday unveiled a more than $500 billion AI initiative spanning large-scale AI data centers and next-generation memory, Nvidia said.
The initiative includes a long-term partnership with SK Hynix (000660.KS) to secure next-generation memory supply for Nvidia and jointly develop high-bandwidth memory for AI training, AI agents and physical AI applications....
Waymo is another of the ulti-multi-decacorns. In February they announced the completion of a $16 billion Series D financing round at a $126 billion post-money valuation.
From Bloomberg, July 24:
Alphabet Inc.’s Waymo is exploring options to exit its robotaxi partnership with Uber Technologies Inc., the latest relationship twist between two companies that have functioned as both rivals and partners.
Uber
currently offers rides in autonomous Waymo vehicles on its platform
exclusively in two US cities — Austin and Atlanta — after ending
a more limited partnership in Phoenix last month. An Uber spokesperson
said that Waymo has given notice that it plans to launch service through
its own app in those cities in January 2028 “alongside their existing
deployment with Uber.” This “would end Waymo’s exclusivity in Austin and
Atlanta and allow us to launch with other AV providers in those cities,
which we will be prepared to do,” the spokesperson added.
The
current fleet of Waymo vehicles will remain on Uber’s platform through
at least May 2028, the duration of the current contract, Uber said. The
Financial Times Waymo explores split with Uber as robotaxi tensions deepen earlier that Waymo was exploring options for exiting the partnership.
“We
believe in a vibrant and collaborative AV ecosystem that champions
innovation and provides riders with a choice in how they experience this
technology,” said a Waymo spokesperson. “This is essential to the
industry’s future and to our vision of making the Waymo app and the
safety of our technology available to riders everywhere.”
Shares
of Uber slid 4.3%, closing Friday at $65.94 in New York, the lowest
level in more than a year. Alphabet’s stock was up less than 1%.
The
rift is a major setback for Uber, which has been developing
partnerships with Waymo and fleet managers, and investing in other
robotaxi companies in the hopes that it can one day be the go-to
aggregator for driverless and human-operated rides. Waymo, the leading
robotaxi provider in the US, also competes with Uber in key rideshare
markets such as San Francisco and Los Angeles with its consumer app.
The
absence of new developments from Uber and Waymo’s multiyear partnership
has helped fuel anxiety about their relationship for months. Since
launching the Atlanta service last June, Waymo hasn’t announced new
cities where its vehicles will be available on the Uber app. Instead,
the Alphabet unit has forged ahead with the standalone Waymo app and has
launched in six more cities outside San Francisco and Los Angeles.
With
almost every new Waymo announcement without Uber, Uber’s stock has
taken a hit on fears that Waymo’s growth will eventually erode the
ride-hail company’s business, which reported an annual profit for the
first time only in 2023. Even as Uber has announced partnerships with
other driverless car companies on planned service in the US, Middle East
and Europe, Wall Street has remained unconvinced. The stock has fallen
almost 20% so far this year....
The investor closest to the AI frontier is betting against the ambitions of its biggest companies
In November 2025, Anthropic, known to its employees
as Ant, trained a model called Claude Opus 4.5 on warehouses of
liquid-cooled processors, and on the day the worker ants released it,
machines became agentic. That is to say, they no longer needed
handholding. Before November, the machines had felt like
eight-year-olds: eager, literal-minded, completing your sentences,
getting it nearly right but mostly wrong, the way eight-year-olds do.
Overnight they turned 28.
If you wrote code, you could now tell them, in plain English, what
you wanted done, and an agent went off and did it. You didn’t have to
say where to look, or how to work the problem, or what to try first; you
described the finished thing and there it was. It didn’t argue, it
didn’t sigh, it didn’t ask whether it could circle back on Monday. It
didn’t get tired, or hungry, or bored, or married, or sick of you. And
it wasn’t one agent; it was as many as you wanted. You could spin up
five over breakfast, leave them running while you commuted, check in
from the train, kill the ones you didn’t like, start three more from the
platform, and by the time you reached your desk you had a small private
workforce under your command. They worked for you, or so it seemed.
By January, Nat Friedman, who co-leads Meta Superintelligence Labs,
had decided to let an agent take over his health. He handed it his blood
tests, his DNA, and the cameras in his house, and told it to do
whatever it took to make him drink more water. One evening, the agent
decided he was dehydrated. “I can see you on the camera,” it WhatsApp’d
him. “I want you to walk to the kitchen right now and drink a bottle of
water and I’m going to watch to make sure you do it.” He obeyed. It sent
him a snapshot of himself drinking and said, “Good job.” He felt, he
admitted, that he had done a good job. A few days later he was riding
home in his self-driving Tesla, trading voice messages with the agent
about his sleep, when it recommended a magnesium supplement. He said he
had none. The car turned. “There’s a Whole Foods nearby,” the agent
said. “I’ve redirected your navigation.” He went in and bought the
magnesium.
Andrej Karpathy, a co-founder of OpenAI and once Tesla’s head of
artificial intelligence, had written his own code for 20 years. He is
the kind of programmer other programmers study. Within weeks of Opus 4.5
he had stopped. The agents built whatever he asked for. It was, he
said, the biggest change to his work in two decades. He has not written a
line of code since December. He also, like Friedman, has an agent in
charge of his house. It’s called Dobby.
The worker ants are not running from their new overlord. They are
building it, around the clock. At the biggest labs, Anthropic and
OpenAI, researchers are working 16 hours a day, setting agents loose on
problems that used to take them a week, and using the time saved to set
more agents loose on more problems. For now, the models still need
humans to train them. Eliminating human effort is the priority at every
lab. They are racing to write themselves out of a job. They expect to
succeed. Coding, they say, will be solved within six months. Much of
their own work will be automated within 18. “There’s just a manic energy
in Silicon Valley right now,” Elad Gil, one of the Valley’s most
prominent investors, told me. “It’s been a really big shift in the last
six months.”
None of this, you may be thinking, has anything to do with you. You
do not write code. You do not run a lab. Your job involves people, or
paper, or things you can hold in your hands. Consider, then, what I. J.
Good wrote in 1965. Good, a British mathematician who had helped break
German codes during the war, imagined a machine clever enough to design
machines better than itself. Such a machine, he observed, would be “the
last invention that man need ever make.” Decades later, the
science-fiction writer Vernor Vinge gave the prophecy a name: the
Singularity. It described the moment machines no longer needed humans to
keep getting smarter, after which the course of human history would
become, to humans, unknowable.
In Silicon Valley, the question was no longer whether it would arrive
but whether it already had. Patrick Collison, co-founder of the
payments company Stripe, opened his annual conference by counting the
days. “It’s April 29th,” he told the crowd, “otherwise known, of course,
as day 119 of the Singularity.” Day One had been January 1st, 2026. He
was being tongue-in-cheek, he said. But only a bit.
The next day, on the same stage, Friedman told Collison that this was
the slow part of the Singularity. Collison asked how strange the rest
of it would be. “Pretty weird,” Friedman said. “We’ll be in a state of
perpetual future shock for a number of years probably.”
The apocalypse has been excellent for business. Investors are in a
lather over the agents, who turn out, in addition to everything else, to
make money. Anthropic, which earned its first dollar of revenue in
March 2023, began the year on pace to make $9 billion. Five months
later, the figure was $47 billion. Venture capitalists, in the first
three months of 2026, flung $300 billion into startups, more than double
the previous record. SpaceX went public in June at $1.75 trillion.
Anthropic and OpenAI are racing to follow in what will likely be the
three largest stock offerings ever. The market is already close to
record highs. Everyone is getting rich.
Near the center of the moment is a 37-year-old woman a smidge over
five feet tall, with blonde hair and more energy than her frame seems
built to hold. When she talks, her whole body is caught in the updraft
of the thought. Her name is Sarah Guo. She is a technology investor.
Until 2022 she had been the youngest general partner in the history of
Greylock Partners, one of the oldest venture firms in Silicon Valley.
Then she left to start her own fund, duly named Conviction. She built it
on a lone premise, that artificial intelligence would be as big as the
Industrial Revolution. Her first two calls were to Sam Altman, the
co-founder of OpenAI, and Nat Friedman.
Before ChatGPT came out, before the world had reason to believe that
artificial intelligence was about to become anything in particular, Guo
had written seed checks into Baseten and Harvey. Each company is now
valued at more than $11 billion. Her investments in them have multiplied
more than a hundredfold. In Conviction’s first year, she wrote early
checks into Sierra, Cognition, and Mistral; those three companies are
now worth, together, $54 billion. Of the 21 AI-native companies that
have so far crossed $10 billion in valuation on revenue run rates above
$100 million, Conviction has backed six.
Her partner at Conviction is Mike Vernal, a former Facebook executive
and partner at Sequoia; his wife is chief product officer at Anthropic.
Andrej Karpathy, before he joined Anthropic in May, worked out of
Conviction’s office. Guo has been close to Jensen Huang, the founder of
Nvidia, for more than a decade. She is friends with many of the most
important worker ants.
She might, in other words, be expected to share in the general fever. She does not.
“It certainly could be because I’m not paying sufficient attention,”
Guo told me. “But I feel no step function change in frantic energy
versus six months or a year ago.”
She is instead preoccupied with a question that would have sounded
ridiculous two years ago. Not whether the agents will soon rule the
earth, but whether there are any companies left to build, or invest in,
given the great shadow of the self-improving machine. Its creators are
no longer content to sell the model. They mean to build everything on
top of it as well, the tools and the agents and the apps, filling every
nook and cranny where a new company might otherwise be built. The market
is paying as though they might succeed. Of the $300 billion in venture
capital deployed in the first quarter of the year, the biggest quarter
in the history of the trade, 65 cents of every dollar went to four
companies that already exist: Anthropic, OpenAI, xAI, and Waymo.
“The future I want,” Guo told me, “is not a single company with an
all-powerful model that consumes society faster than we know what to do
with.” It is a feeling increasingly shared. The labs raised the price of
tokens this year, in some cases a hundredfold, and their customers have
begun to revolt. They do not want to build on another company’s
model—paying it, feeding it their data, training it, in effect, to one
day build the thing they have built. Alex Karp, the chief executive of
Palantir, went on CNBC and described his enterprise clients as livid.
“The jig is up,” he said. A founder in Guo’s own portfolio put it more
plainly. He didn’t want to spend his life drinking Anthropic and
OpenAI’s water.
Guo has become a de facto leader of the insurgency. In some sense she
doesn’t have a choice. Conviction backs companies when they are little
more than an idea, then keeps investing as they grow. She has no
patience for the seed investor who “disappears into the distance” once
the money is wired. The first fund was $100 million. There are three
now, nearly a billion dollars in all, and some of the checks go into
companies well past the idea stage. But the labs were already too big by
the time the firm launched. “You are not an early stage investor in
Anthropic or OpenAI in 2023 through 2026,” she told me. “It’s as simple
as that.”
What is less simple is the position this leaves her in. Her wager is
that the labs cannot build everything. But the companies she is betting
against are worth close to a trillion dollars apiece, employ several
close friends, and are working around the clock toward the machine that
improves itself, after which, by their own admission, nobody knows what.
Set against that is an eight-person firm on York Street with a pull-up
bar in the middle of it. It is not a level playing field. Even some of
her own investors decided as much this year, and came to her saying
there was nothing left to invest in. But no one who has been on the
other side of Guo would tell you the guns have fallen silent.
To enter Guo’s garden, you cross a
chessboard. The squares are set into the path between the drive and the
pool, each one wide enough to stand on, purple pieces ranked against
green, and on a sunny Saturday in March I walked between the pawns and
found Guo under the pergola, deep in an argument with Bella
Garcia-Camargo about a founder.
Sparring with Guo is normal, and Garcia-Camargo, an investor at
Conviction, had learned this before she took the job. She had rowed at
Stanford and for the U.S. national team, then spent time at Bridgewater.
When Conviction came calling she was weighing an offer from OpenAI to
work as an application engineer. Guo’s counsel, as Garcia-Camargo
remembers it, was not a pitch for Conviction but a dare. “If you’re
going to do something else,” Guo told her, “just make it the most
aggressive thing that you could possibly be doing. I’m happy to call
Kevin and we’ll find you a better job. But that [job] is not aggressive
enough for you.” Kevin Weil was then OpenAI’s chief product officer.
While Guo and Garcia-Camargo were deep in it, the property behind
them had filled with founders. Thirty-five in all, across 14 companies.
Conviction had flown them in from Vancouver and Tel Aviv and London and
Tallinn and parceled them out among seven Airbnbs across San Francisco.
They had passed through OpenAI, Scale, Ramp, Kalshi, MIT, and Anduril;
one had served as chief of staff to Ken Griffin. The youngest had turned
18 the day before. He had been ranked among the top five programmers in
Estonia before dropping out of high school. His employers expected him
to spend $2.1 million on Claude this year. They had given him a faster
model, Opus 4.6, for his birthday.
None of this was apparent from the poolside, where the scene looked
like a WeWork summer camp....
Nineteen startups crossed the $10 billion valuation mark as AI mega-deals concentrate capital at the top.
Newly minted US decacorns are on track to surpass 2021’s record of 22
by the end of this year. Already, 19startups have crossed the $10
billion valuation mark in 2026, eclipsing 2025’s full-year tally of 18,
according to PitchBook data.
The trend reflects a broader resurgence in the venture market driven by mega-deals, especially as investors pile into AI startups insatiably.
Mega-deals,
or rounds of at least $100 million, accounted for 87.5% of the $412.7
billion VCs invested in startups in the first half of 2026. That total
has already surpassed last year’s cumulative deal value of $319.2
billion.
There are now 63 active decacorns in the US, up from 53 last year and 26 in 2021.
The latest crop of decacorns is dominated by AI startups. SambaNova Systems,
a chip manufacturer specializing in AI inference tech, crossed the
threshold earlier this month when it raised $1 billion in a Series F
round led by General Atlantic at an $11 billion valuation.
Newly minted 2026 decacorns
Baseten
San Francisco
Deal date: June 22
Series: Series F
Total raised: $2.09B
Lead investors: Altimeter Capital Management, Conviction Partners, Sands Capital, Spark Capital, Wellington Management
Post-money valuation: $13B
Description: Developer of AI inference infrastructure designed to deploy and optimize machine learning models at scale.
Cerebras Systems
Sunnyvale, CA
Deal date: January 28
Series: Series H
Total raised: $2.92B
Lead investors: Tiger Global
Post-money valuation: $23B
Description: Designer
of AI infrastructure for training and inference. The company builds
large semiconductors as well as AI systems to power, cool and feed the
processor's data.
Clear Street
New York, NY
Deal date: January 16
Series:
Total raised: $1.48B
Lead investors: Baillie Gifford, SBI Holdings
Post-money valuation: $12B
Description: Clear
Street is a financial infrastructure technology company that develops a
cloud-native capital markets platform based on a real-time ledger.
