TheChek News Flow

The famed US investor and hedge fund manager spoke to “Globes” during a recent visit to Israel about the US economy, AI, his Israel investments, love of Judaism, and anti-Semitism.

US investor Daniel Loeb, owner of the Third Point hedge fund, has earned a reputation as an activist investor who is not afraid to share his opinion on his investments, even when they’re unpleasant to hear. Whether in letters to investors or in media interviews, he doesn’t spare those companies that made mistakes and lost his trust, but at the same time, explains how mistakes or failures have turned into investment opportunities. He also hasn’t shied away from organizational changes, changing management and in extreme cases, even waging legal warfare.

Recently, however, it seems like something has changed for this New York shark. Loeb, who has been on a journey to rediscover his Judaism , engages in philanthropy and has set up a fund to invest in startups. Consequently, he’s spent less time on fighting, softened his tone toward portfolio companies, and devoted more time to spiritual activity. It seems, too, that the criticism he received for selling tech stocks too early has been replaced by a significant AI acquisition strategy with Nvidia in the starring role.

Above all, the 2025 model Loeb is a man of contrasts: He grew up in a secular home in California, but now leads Torah study groups in New York, invests in alumni of IDF elite intelligence units, and cultivates ties with rabbis, even those calling for draft exemption for haredi Jews. And now, as someone who was concerned about global antisemitism even before October 7, he intends to play a pivotal and influential role in the Jewish world, both in the US and in Israel. During a recent visit to Tel Aviv to attend institutional investor technology conference Jefferies TechTrek 2025, Loeb gave this exclusive interview to "Globes."

Nvidia back in the portfolio

The headlines generated by Loeb over the past two years dealt mainly with his investment strategy, particularly in AI and technology. He was criticized for selling all Nvidia holdings in 2023 and for believing that rising interest rates in the US in 2022-2023 would hurt tech companies.

Today, it appears that Loeb is a passionate believer in AI, especially Nvidia. After only holding shares last year, since the beginning of this year, Third Point has purchased $442 million in Nvidia shares, constituting 6% of the fund's total holdings - the third largest holding in its portfolio.

"We bought Nvidia in the past, and we sold it at a good return," Loeb tells "Globes." "This year, we bought again below our selling price, around the same time that the tariff effect lowered the price to between $90 and $130. I think it was a good deal."<

Two basic assumptions led to the repurchase, Loeb says. "The first is that the world is short on available computing power, and the second is the perception of how Nvidia operates in its market [which] needs to change from end to end. If we once thought it operated in a cyclical market where a company launches products, and demand for them is met until new competitors enter and profitability decreases, until the company launches more innovative products, then in the case of Nvidia, the disk has switched."

Please explain?

"The behavior of the AI market is different from anything we’ve known in the past. It now turns out that the increase in demand for computing exists not only for training models, but also for their next stage of operation, the "inference" stage. The use of tokens - the processing units of language models - is absolutely exploding while the demand for AI-powered data centers is growing. All this time, Nvidia has been releasing more and more chips and entering new fields and industries like robotics, placing heavy barriers to entry for future competitors. I'd say this could go on indefinitely, but even if it doesn't, it can continue for at least another extended period. So, instead of profits reaching a certain peak and then going down, they can only go up. Even if the market share falls slightly, growing demand will still overshadow all of this."

You're a big believer in the AI market, and you've also increased your holdings in Amazon shares, but you've actually sold shares in Microsoft, Meta and Google.

"We sold off a lot of things, all of which were sales held in advance of ‘Liberation Day' [the terms coined by US President Donald J. Trump upon presenting his tariff plan in April - A.G.]. Since then, we've repurchased a large position in Microsoft and a smaller one in Meta. These were actually trades we made because we thought the stocks were performing ahead of their time - so we avoided some of the volatility."

In the past, after holding Google shares, you became a harsh critic of the company. In 2022, you wrote, "The mistakes it made in launching Gemini further contributed to the narrative that it would eventually become the loser of the AI world." Do you still stand by those words?

"We're years after this event, and three years in the AI sector is like an eternity. In any case, we’re not currently investing in Google. I think they're doing a great job and that they have one of the best AI teams in the world, run by someone who is now a Nobel laureate, Demis Hassabis. Our decision to invest in the stock has more to do with risk analysis and valuations; we just saw good risk-reward prospects at a few other companies - it has nothing to do with Google."

"A weak dollar is not necessarily a negative phenomenon"

Loeb studied at University of California at Berkeley and then moved to New York and Columbia University, where he studied alongside Barack Obama. That same year, he made one of his largest and riskiest investments - $120,000 in a medical products company - and lost the entire investment. After completing his studies, he worked at several investment houses and funds: Warburg-Pincus, Jefferies and Citi.

Third Point, which now manages $21 billion - of which $9.4 billion is in its main hedge fund - was founded exactly 30 years ago, in 1995, with just $3.3 million in capital raised from family and friends. Since then, it has generated an average annual return of about 13%, beating the S&P 500, which posted an average annual return of 9.7% in the same period. Since the beginning of the year, the fund's return has been 6.8%, lower than the S&P 500 (10.8%), due to its large exposure to the credit market.

What is your outlook for the US economy after six months of turmoil due to Trump's tariffs, the decline in the dollar's status, and the weakening job market? This also hurt the performance of your hedge fund in the first half of 2025.

"In the first half of the year, our yield was lower than the S&P 500, but today we are rising almost as much. I think we're doing well overall. We don't measure ourselves in weeks or months or how many winning assets there are in this year's portfolio. It should be noted that the market as a whole is very strong this year, and that a weak dollar is not necessarily a negative phenomenon, in fact, for global companies, it’s an advantage."

And what will interest rate cuts do to the US market and your investments?

"It's a balancing act. The ljob market has been revised downwards, but less than expectations from the outset and in any case, we haven’t reached the low we were in the Biden days, so the economy is stronger than we thought. True, we may expect to see cuts and a residual effect of tariffs, but I think starting next year, and especially from March, we'll start to see fiscal stimulus, capital goods incentives , and tax cuts and the return we get from AI and its contribution to productivity will also be clear, so the economy will pick up at the end of that quarter. At the same time, we may see interest rates rise again at the long end of the curve."

"I saw how entrepreneurship was life-changing"

Loeb grew up in the coastal city of Santa Monica, in the 1960s and 1970s, to a secular Jewish family. His father was a lawyer - "not a great investor," he says - and his mother was a homemaker. "We were secular Jews in a secular environment. Maybe the food in the house was Jewish-European, with kreplach and kneidelach, - they spoke Yiddish - but we didn't go to synagogue and we went to school on Yom Kippur. Very few kids had a Bar Mitzvah and I don't think anyone around us observed Shabbat," he recalls. "I never encountered antisemitism there because everyone was so assimilated, there was nothing to be against anymore. It was never an issue, and at the time it felt like this was where Judaism went to die."

One person who influenced Loeb as a child was his aunt Ruth Handler, one of the founders of the Mattel toy empire and the inventor of the iconic Barbie doll. "Ruth was my father's younger sister. She and her husband Elliot founded Mattel with his partner Dave Mattson, and the company's name is actually a combination of their names. Mattel was founded as a jewelry company based on a crazy idea they had, using a technology they developed to melt plastic scrap into molded jewelry. After World War II, they made the decision to focus on making toys using the same technology, and Mattson said, 'That's a stupid idea, buy my share and I’m out' - and the rest is history."

Handler, noticing that in the US, dolls were considered only for babies, understood from her daughter Barbara that older girls wanted dolls that looked more like real people. She persuaded the company to create a production line for a more mature-looking doll, which she named after Barbara. Beginning in the late 1950s, Mattel became a huge success, with Handler serving as the company's president until the mid-1970s, most of Loeb's childhood and adulthood.

"I can't say that I necessarily learned lessons from her that I apply to my business today, but she gave me something to aspire to," he says. "They lived in a beautiful penthouse. They had a beach house that we would go to. They drove in a nice car - a Rolls Royce with a phone in the back seat - in the 1960s one call would have cost a fortune. I thought it was really cool. She was the axis that connected us all, the family and the community. She was kind of our matriarchal mother. I saw how entrepreneurship was life-changing and a way to achieve great wealth."

"We’d make 100 investments in Israel if we could"

In recent months, troubling signs have emerged that future funding for Israeli startups is shrinking, with investments in Israeli funds down 80% and a decline in the share of foreign investors contributing to local investment leverage.

In the past four years, the part of foreigner investors in the Israeli investment pie has dropped from 61% to 51%. Does this worry you?

"It doesn't really matter where it comes from, but we see more money chasing deals here than when we started investing here. We made our first investment in Israel in 2015 with SentinelOne, when there were a relatively small number of foreign investors in Israel. I think these statistics are misleading, and you can see this increase in the capital market as well. We started to see a trend like this already in 2021 with capital that overheated the market, and then we took a step back from deals. Ultimately, we saw the tech bubble burst at the end of that year.

"I visited Israel a year ago. People were deeply depressed at the time, most of the hostages were still being held in Gaza and the war was at its peak. Even then I said that if Israel were a share, I would buy it. And since then, we’ve seen a massive influx of investors coming here, certainly on the venture capital side. There is a community of investors and entrepreneurs here who are doing well, despite the worst public relations campaign ever against Israel, and despite the fact that many of them serve in the reserves. Somehow, they manage to create some of the most innovative companies that attract the most early-stage capital we've ever seen."

Still, in all of 2025, you made only one investment in Israel, a small number compared to the investments made here by foreign venture capital funds such as Lightspeed, Bessemer or Sequoia.

We’re not cutting back on investments. In fact, we’re more active than ever in evaluating Israeli companies for investment. We submitted three or four proposals even before talking about price expectations, expanded our team in Israel, and we have dedicated capital earmarked for local firms. The challenge is that so much money is chasing Israeli startups that prices have skyrocketed, and we’ve remained disciplined investors. We won’t make an offer to entrepreneurs who expect a $250 million valuation when their revenues are close to zero. But that just shows you how strong the Israeli market is right now. We’d be happy to make 100 investments a year here if we could but we’re disciplined."

It sounds like we're back to the merry bubble days of 2021-2022.

"The pricing is even higher, but the quality of the companies has improved and the concentration of talent here has also increased. The expertise here in AI and cybersecurity has reached a peak at an important point in the development of technology, and for us, as investors, it’s an advantage, so we’re optimistic that we’ll be able to allocate more capital here down the road."

The industry is red hot and the companies are at record valuations because the big funds are ready to pay exorbitant prices and gamble on companies with no revenue.

"Their strategy is different from ours. Some of the large funds have a very, very early entry strategy for companies at their earliest [seed] stage, and they are less sensitive to the company valuation in the next funding round - the one where we would be able to enter - because they already have a significant stake in the company. That's why I wouldn't call them gamblers. We invest in later stages."

Childhood friend becomes VC fund strongman

In addition to the hedge fund, Loeb also invests in startups through his venture capital fund, Third Point Ventures. He raised funds from outside investors, separate from the hedge fund, and its investments are managed by Robert Schwartz , a friend from high school. Their friendship began in karate class, and then Schwartz, who was a few heads taller than Loeb, became his unofficial schoolyard bodyguard. Schwartz was born in the US to a Jewish family with Israeli connections: his great-grandfather was one of the first settlers of Ra'anana. Today, Rob Schwartz is the strongman at Third Point Ventures, which manages about 130 investments in companies, most in the US.

His investments in Israel are described as "Zionist activity in part, but not philanthropy." Yet, despite his activist image, Schwartz is not seen as someone who takes a leading role on the boards of Israeli companies. One of the investments was in Verbit, where, despite the need for changes due to company management and the development of competing technologies, the restructuring process moved at a sluggish pace - some say a little too late.

"I think things could have been worse, and the fact that we acted decisively-changing management, shifting strategy, and cutting costs helped the company," Loeb says. "Today, Verbit has a strong renewed chance of success and is close to breaking even. Every investment portfolio has a 'problem child,' but do you always know exactly how to handle it? That’s part of the job of a venture capital fund: how you work with companies that don’t necessarily go your way. Where we can preserve our investment, even if the company’s strategy has shifted from what we originally backed, we will."

Despite his regard for Israel, Loeb was not historically known as an investor in Israeli companies or stocks. It was an early investment that Schwartz led in the cybersecurity company SentinelOne, developed in the US by Israeli entrepreneur Tomer Weingarten, and which became one of the great successes that eventually led Loeb to open an office in Israel.

Along with other investors, Schwartz led a $25 million investment round in SentinelOne in 2015 at a $100 million valuation, according to PitchBook, and benefitted greatly from the IPO when it raised $1.2 billion at nearly $9 billion company value. The company is currently trading at of $6.1 billion market cap, which is lower than the IPO, but still represents an increase in value relative to the initial investment. Third Point exercised hundreds of millions of dollars worth of stocks, and gained a more than 100x return on its investment. The fund has also recorded handsome exits on the Palantir, Lyft and SoFi IPOs.

Not all of Third Point Ventures investments have been successful, and unlike the hedge fund that Loeb manages directly, it can't enter and exit its investments at a moment’s notice. Capital investments in private companies require years of sitting on the board of directors, and selling shares in such companies is a long and lengthy process, especially if the company encounters difficulties. For example, it invested in DiDi, the Chinese ride-hailing, taxi service and delivery company which clashed with the Chinese government and declined in value, and in crypto company FTX, which collapsed after widely reported revelations of massive fraud.

In early 2022, Loeb and Schwartz recruited Sapir Harosh , then an investment manager at Pitango and active in the 8200 alumni organization, where Loeb is a donor, to establish Third Point’s first representative office in Israel. The Israeli VC investments include NextSilicon, an AI chip startup founded by Elad Raz that competes with Nvidia. "If we’re lucky enough to get a slice of Nvidia's share in the AI market, we’ll create tremendous value," says Schwartz; Trulion, an AI-powered accounting platform that automates financial workflows; Grip Security, a cybersecurity company that became Harosh's first investment for the fund; and Zenity, a cybersecurity company for AI agents that was ranked on this year’s Globes ten most promising startups list. This year, Third Point’s only investment so far is in Unframe, cofounded by Shay Levi , of Noname Security, developer of a user-friendly API security platform that was sold to Akamai for $500 million.

Everyone claims to do AI, but few companies truly deal in it. As an investor, how do you separate the wheat from the chaff?

"That’s a great question, because many times, behind companies in this field lie nothing more than matters related to data analysis or business processes, and not machine learning, for example. If you're processing data using OpenAI or Anthropic and reselling it, you're probably in a market that doesn't have significant barriers to entry and you won't be able to protect your business from competition that will hurt profitability and the unique value you offer, unless you're first and working fast to capture a significant market share. We'll need more than that to invest in your company, and for you to have a more substantial engagement in the field - whether it's through vertical specialization, for example, with proprietary information, or a horizontal core process."

Loeb's response to antisemitism in academia

Loeb was active against antisemitism in the US even before October 7. He had even planned to host a symposium on the subject at his home with Disney CEO Bob Iger, actor Chris Pine, and journalist Bari Weiss a few days before the massacre; the event was postponed to late October.

On October 7, Loeb says, "I was in a state of shock and pain, numbness. I was abroad but immediately returned home to be with my family and slowly began to connect with friends and acquaintances. This was the immediate effect of October 7: a spontaneous reaching out between New York Jews."

Like many other Jews who had studied at the prestigious institution, Loeb was horrified to see how the Columbia University campus had become one of the largest and most vocal centers of anti-Israel protests in the US. After the massacre, in parallel with the huge demonstrations that broke into the university's offices, Loeb told the "New York Post" that the anti-Israel riots might change his criteria for accepting candidates for jobs at his hedge fund. The decision led him to give preference to graduates of pro-Jewish campuses, such as Yeshiva University, University of Florida and Emory University, over long-standing institutions that did not fight antisemitism, such as Harvard, Yale and MIT.

You were one of the first to come out against the demonstrations on campuses about two years ago, the most prominent of which was in Columbia, where you studied.

"I think it's clear from the way the protests were conducted on campuses that it was organized in advance. You saw it on the signs, the leaflets that came out, the advance preparations. The other side has been preparing for this narrative shift for 20 or 30 years by infiltrating our universities, our politics, our curriculum from kindergarten to high school. It may be counterintuitive to think that such a barbaric act could be used to your advantage but they have managed to do it. And they were able to do it out of decades of ideology that they perpetuated, its link to Marxism, and the recruitment of lecturers who in turn influenced these students. Later on, they’ve excelled at using social media to recruit people who are vulnerable to these messages and flooded them with false information."

The demonstrations are down, but we have lost the young Americans. A recent poll shows that 60% of Gen Z support Hamas.

"Americans are broadly supportive of Israel and there are also populations with very high support rates such as conservative Christians and Republicans, but you're right that there's a problem. One problem is that even among Jews there are many opponents of Israel - the Jewish community in the US is very divided among itself.

What can Israel do to improve its public diplomacy profile in the world?

"I think it's incredibly complicated. Everyone has opinions about these things and they think they can do more, but it's hard. Look at the numbers: 16 million Jews in the world of whom perhaps only eight million are actually involved compared with a few billion, not just Muslims but people in general who are anti-Israel. The odds against us are just piling up."

Loeb's pro-Israel activism is not new, but it gained public prominence last year after an article in the "Washington Post" described his participation in a WhatsApp group about Israeli hasbara, alongside other Jewish businessmen such as Len Blavatnik, Bill Ackman and Joshua Kushner, brother of Donald Trump's son-in-law, Jared Kushner.

Loeb, who even before the war had been deepening his knowledge and faith in Judaism , sees the massacre and the subsequent war as a spiritual challenge. "Exactly a year earlier, on Simchat Torah in 2022, I began to study Torah. I was really looking forward to this date to mark the occasion with an exciting ceremony of a new reading."

Before the war, Loeb, together with Chabad, launched the "Simchat Torah Challenge," a US-wide Torah study group that currently numbers more than 20,000 people. "After October 7, Torah study groups spontaneously grew from a handful to a few dozen and then hundreds. People understand who they are and what they are working against - antisemitism and Jew-hatred, what happened on that terrible Saturday, what’s happening on campuses. Only a few have ever read Torah before, if at all, and yet, this is our way of honoring those who were murdered on Simchat Torah."

The spiritual awakening: "It began with the Shema"

Loeb's spiritual awakening begins with the illustrious roots he discovered about his own family that connect him to some of the most prominent Hassidic rabbis of late European Jewry. Loeb learned he is a descendent of prominent figures such as Rabbi Levi Yitzchak of Berdichev, Rabbi Baruch Teomim-Frankel, and "Haham Zvi" Hirsch-Ashkenazi, through whom he is also connected distantly to former Chief Rabbi of Israel, Meir Lau, whom he also met on his last trip to Israel earlier this month.

Loeb began taking Judaism more seriously during the Covid pandemic, when he heard about the Aleph Institute. Initiated by the Lubavitcher Rebbe, Rabbi Menachem Mendel Schneerson, Aleph works for humane punishment and incarceration in the US justice system and serves the religious needs of Jewish men and women in prison. The founder and executive director, Rabbi Sholom Ber Lipskar, contracted Covid in March 2020. "I asked him, is there anything I can do for you? Would you like me to donate to your synagogue?" says Loeb, "and he asked me to recite the Shema for him. I explained to him that I didn't really know how, so he had his son call me the next day and pray with me. It started with reciting the Shema together, continued to the siddur and then tefillin. Today I put on tefillin every day."

The friendship between the two and intensive Torah studies have become a Torah studies project, and since then Loeb has been an enthusiastic supporter of Beit Chabad and Hillel activities on US campuses, to create after-school Jewish frameworks for students. He also supports the N7 Initiative, which encourages joint investments by Israeli and Arab companies as part of the Abraham Accords. He is also a donor to Startup Nation Central which promotes policies to advance Israeli high-tech, co-founded by lead donor Paul E. Singer and headed by former Israel Innovation Authority chairman Avi Hasson. In addition, Loeb is a significant donor to the National Black Empowerment Council, an organization that fosters Black-Jewish alliances in the US.

His visit to Israel at the beginning of the month included a meeting with former Chief Rabbi Meir Lau as well as Sephardic Chief Rabbi Yitzhak Yosef. Loeb is familiar with the spirited debate surrounding conscription, and Rabbi Yosef’s involvement. "I try to stay out of the political discourse in Israel, but I think it's an internal issue that Israel needs to solve," he says. "It is clear that the situation you’re in right now is unsustainable and unfair. I think we need to find a better solution than the existing one, but I'll leave that to the politicians and rabbis."

Minister says Hong Kong will take a two-pronged approach by strengthening its advantages and diversifying financial market

Hong Kong is accelerating reforms of its financial market by making the sector’s ecosystem more resilient and comprehensive, a minister has said, as the city grapples with geopolitics and a rapidly changing technology-led global landscape.

Financial Secretary Paul Chan Mo-po also said on Sunday that Hong Kong would take a two-pronged approach by strengthening its advantages and diversifying the financial market to contribute to the nation’s development into a financial powerhouse.

The city had a key role to play in China’s financial development and was in a favourable position to reform the market, he said.

“Artificial intelligence and blockchain are not only revolutionising technical tools but also reshaping capital flows and the boundaries of financial services, creating new opportunities for financial market development,” he wrote in his weekly blog.

“We are accelerating institutional reform and product innovation in the financial market, striving to more effectively transform the tremendous disruptive power of technological change into a driver of economic growth.”

Chan said favourable conditions for reform included the city’s strong performance in the stock market in the past year and its worldwide ranking as an international finance centre.

He noted that the HK$150 billion (US$19.3 billion) in proceeds raised from initial public offerings (IPOs) so far this year had put the city in the No 1 spot globally, with a marked increase in participation by international, long-term funds.

The stock market was buoyant, as reflected in an accumulated gain of more than 30 per cent in the benchmark Hang Seng Index in the year to date, in addition to a record-breaking daily turnover averaging HK$250 billion, he added.

Chan said that even mainland Chinese stocks listed locally had performed well amid the buoyant market.

For example, since 14 H-shares were listed on the Hong Kong stock market in the past year, the average daily turnover in value of these stocks’ A-shares listed on the mainland has jumped 28 per cent, he said.

Their A-share stock prices on average were 15 per cent higher since September last year, Chan noted.

“These encouraging market performances clearly demonstrate the unique value of Hong Kong’s financial system in connecting mainland China with global capital,” he said.

“In the face of profound changes in the global political and economic landscape, we must seize the current momentum and, through more forward-looking reforms and initiatives, transform these developments into lasting momentum driving higher-quality, more diversified development of Hong Kong’s financial industry.”

Chan added that the reform would be crucial in consolidating and enhancing Hong Kong’s status as a global finance centre.

“It is also a crucial path to unlocking growth potential and contributing to China’s development into a financial powerhouse,” he said.

Discussing Hong Kong’s crucial role in the country’s financial development, he noted the combined market capitalisation of the city, Shanghai and Shenzhen’s stock markets was equivalent to China’s gross domestic product (GDP).

In comparison, the combined market capitalisation of the US’ two largest stock markets was twice as big as the United States’ GDP, he said.

“This comparison shows that the country’s financial market still has huge room for development and can serve the needs of the real economy more broadly and deeply,” Chan said.

To press ahead with reform, Hong Kong would refine the existing stock market, asset and management services and green finance, and accelerate expansion into new areas of fixed income, currency, commodities and digital assets, he said.

Last week, the city’s de facto central bank, the Hong Kong Monetary Authority, and the Securities and Futures Commission rolled out a development blueprint for fixed income and currency markets.

The blueprint sets out plans to broaden offshore yuan usage and expand cross-border Connect schemes to enhance offshore liquidity and yuan-linked product offerings.

It also proposes widening the investor base for fixed income products to include family offices, investment funds and corporate treasury centres.

“The goal is to build Hong Kong into an important global hub for fixed income and currency products,” Chan said.

Billy Mak Sui-choi, an associate professor in Baptist University’s Department of Accountancy, Economics and Finance, said the Financial Secretary was providing a clear roadmap for the city’s financial development.

“This is a clear signal to the business sector about what they can do in response to the government’s directional roadmap,” Mak said. “This is important.”

He added that it was also crucial to make it easier for investors to access bond trading, for example, through the HKMA’s subsidiary, CMU OmniClear, which operates a central money markets unit to clear bond products.

The financial secretary aid CMU OmniClear would join forces with the Hong Kong stock exchange to allow investors to use the same clearing platform to manage different stocks and bonds, including assets under the Connect schemes.

Chan also said Hong Kong authorities were establishing connections with Switzerland and the United Arab Emirates to advance offshore renminbi-denominated sovereign bonds as qualified collateral for clearing institutions.

“This will help take Hong Kong’s offshore RMB ecosystem to a new level,” Chan said.

Hong Kong manages about 80 per cent of global offshore yuan payments and hosts the largest offshore yuan liquidity pool, with deposits reaching 938.2 billion yuan (US$120.7 billion) as of July.

PLA scientists trial cutting-edge military technology aimed at completely destroying enemy targets in multi-pronged attack

The US military launched a covert strike on Iranian nuclear facilities in June, dropping several massive bunker-busting bombs on a single target. While US officials claimed the sites were completely destroyed, some reports suggested the core of these facilities might have survived.

This scenario underscores a persistent challenge in modern warfare: even the most powerful conventional weapons may fail to eliminate deeply buried, hardened underground targets.

But what if nuclear weapons were used instead – not just one, but a coordinated sequence of multiple warheads striking the same location in rapid succession?

A new study by Chinese scientists, published on September 10, in the peer-reviewed journal Explosion and Shock Waves, suggests the Chinese military has taken a major step towards answering the question with cutting-edge experimental technology.

The research, led by Xu Xiaohui, an associate professor with the Army Engineering University of the People’s Liberation Army (PLA) in Nanjing, details the development of the world’s first laboratory system capable of simulating the effects of multi-point, high-yield nuclear explosions deep underground – specifically, the impact of three nuclear warheads detonating in close succession at the same target.

Historically, studies on nuclear earth-penetration effects have focused on single-warhead detonations, as analysts have long assumed that a single high-yield bunker buster could collapse or destroy deeply buried facilities.

But modern defence engineering advances rapidly. Xu’s team argues that new low-yield, precision-guided, earth-penetrating nuclear weapons – now reportedly operational in US and Russian arsenals – are being designed with multiple independently targetable re-entry vehicles.

These warheads can be programmed to strike the same location in a “cluster” or “focused” pattern, potentially creating a synergistic destructive effect far greater than the sum of individual blasts.

Until now, however, such multi-point nuclear attacks have existed only in theory.

No nation had a reliable way to physically simulate or test the combined cratering and shock effects of multiple simultaneous or near-simultaneous underground nuclear explosions – especially at realistic scales, according to Xu and his colleagues.

The PLA team’s breakthrough lies in a novel vacuum chamber-based test system, which allows researchers to model large-scale nuclear cratering effects at a fraction of the cost and risk of full-scale testing.

Using principles of similitude theory, the system scales down massive nuclear blasts into small, controlled experiments.

The core of the set-up is a miniature explosive source system based on a two-stage high-pressure gas gun that fires small projectiles to rupture pressurised glass spheres containing simulated blast gas.

This mimics the rapid energy release of a nuclear detonation in a highly controllable and repeatable way.

Crucially, the system enables multiple explosive sources to be triggered with near-perfect synchronisation. The study reports a timing discrepancy of just 0.8 milliseconds between detonations – negligible on the timescale of underground shock wave propagation.

The experimental results are striking.

In simulations based on the US “Palanquin” test, an underground test in 1965 with a single 4.3-kiloton warhead at 85 metres (279 feet) depth, the three-point explosion more than doubled the crater radius from 46 to 114 metres and increased the crater depth from 28 to over 35 metres.

It also boosted the crater volume by an order of magnitude compared to a single detonation of the same total yield.

Most significantly, the projected surface damage area – a proxy for the zone of destruction – expanded from 6,600 to over 80,000 square metres, about three-fifths as big as the Pentagon.

Even in a shallower scenario – 5 kilotons at 20 metres depth – the triple strike increased the surface damage area four times, showing that multi-point attacks were consistently more effective at maximising destruction.

“This study, for the first time, shows through a comparison of simulation results with prototype test data that deeply buried multi-point explosive sources exhibit significantly higher cratering efficiency than single-point sources, providing data support for coordinated multi-warhead penetration strategies using nuclear earth-penetrating weapons,” Xu’s team wrote.

These findings would also “directly serve the nation’s security and protection needs for deep underground engineering”, they added.

But to achieve such precise, synchronised multi-warhead strikes, a military would need extremely advanced capabilities, including hypersonic delivery systems with high terminal accuracy, advanced flight control and timing synchronisation, and robust command, control and communication systems to coordinate warhead detonation sequences.

These were not capabilities associated with traditional nuclear arsenals, but some advanced nuclear weapons deployed recently could have solved these problems, according to the study.

Moreover, by using low-yield warheads (around 5 kilotons), such strikes could be framed as “limited” or “tactical”, reducing the political threshold for nuclear use – a concept that has drawn increasing interest from policymakers in Washington.

What’s Nvidia buying by pumping money into the AI economy through its $100 billion OpenAI investment and other deals? Influence, threat deterrence and a gusher of chip sales, that’s what.

Even by the standards of one of the most prodigious dealmakers in tech, the past month or so has been a head-spinning one for Nvidia’s Jensen Huang.

Just in September, Nvidia agreed to buy any unused Nvidia graphics processing units that cloud provider CoreWeave doesn’t sell to other customers over the next seven years, a deal potentially costing Nvidia around $6.3 billion. It also invested $700 million in U.K. data center startup Nscale and paid more than $900 million to hire the CEO and engineers of networking startup Enfabrica and to license its technology. It bought a 4% stake in struggling chipmaker Intel for $5 billion as part of an agreement to co-develop chips with the company.

And this week, Huang topped all these deals with Nvidia’s letter of intent to invest up to $100 billion in OpenAI, the leading artificial intelligence model maker, over the coming years. As part of a partnership between the companies, the AI startup will use the investment to build AI data centers that consume at least 10 gigawatts of power, the equivalent of between 4 million and 5 million Nvidia GPUs—about the number of total GPUs Nvidia plans to ship in all of 2025. “This is a giant project,” Huang said on CNBC this week.

Nvidia’s investments have become so large and broad—often consisting of multiple companies in data centers, model making and other categories—that it’s starting to turn into something like AI’s government.

The amount of cash it’s pumping into AI companies like OpenAI, for example, approaches the level of a stimulus program. More than any other company, Nvidia benefits from the rising tide of the AI economy, since its GPUs are the brains used to train and run the most advanced large language models. The money Nvidia pours into AI startups helps them afford their GPU bills, effectively boomeranging some of the cash back to the company in the form of chip sales (Nvidia says it doesn’t predicate its investment deals on companies buying its chips).

At other times, Nvidia’s money acts as a form of soft power that could help it buy goodwill from actual governments. Huang has grown close to President Donald Trump, for example, who has taken a keen interest in the turnaround of Intel and American dominance in AI. And last week, when Nvidia announced a recent string of U.K. investments, Huang appeared onstage in London with U.K. Prime Minister Keir Starmer, whose competition authority previously blocked Nvidia’s acquisition of chip designer Arm.

Nvidia even assumes the role of a central bank. For example, in recent weeks, it has become harder for some data center builders to arrange the financing for large-scale projects, according to two people familiar with their discussions. Nvidia’s willingness to step in and backstop financing for data center companies has become helpful in getting other funders of those projects on board, the people said.

Throwing so much cash around has other possible benefits for Nvidia, which already has the largest market capitalization in the world, at $4.33 trillion. The money could lessen one nagging fear some Nvidia executives have: the long-term threat of customers switching to alternatives to Nvidia’s chips, whether they’re OpenAI’s in-house design or chips from Amazon, Google and Advanced Micro Devices.

Still, Nvidia’s investment in OpenAI is so large that skeptics worry it could make demand for Nvidia’s products look more robust than it actually is.

“They’re trying to stimulate demand for their product, which is a perfectly valid strategic direction,” said David Yoffie, a professor at Harvard Business School. “On the other hand, the scale of this investment and the fact that it’s being made into only one customer raises the question of—are they stimulating artificial demand, which won’t be sustained over time and could produce some severely negative returns down the road?”

Yoffie, who served on the board of Intel for nearly three decades until 2018, believes Nvidia’s recent investment in OpenAI could give big cloud computing companies an incentive to put more into developing their own chip designs. Doing so could give them more control over the cost, performance and supply of chips, the most important ingredient in their data centers.

OpenAI executives have also talked publicly about their plans to one day rent out their extra cloud computing capacity—which would put the AI startup in direct competition with today’s major cloud providers. Nvidia’s prominent role in making that vision possible could give cloud providers another reason to lessen their dependence on the chipmaker.

“If I were in Google or Amazon’s shoes, I would try to accelerate—more than I am already—my efforts to build an alternative architecture to Nvidia,” Yoffie said.

A spokesperson for OpenAI said the firm is focused on expanding its data center and GPU capacity to meet its own computing needs, rather than on building a cloud service. In an interview with CNBC, Greg Brockman, OpenAI’s president, said this week he believes eventually every person on the planet will need their own dedicated GPU to handle demand for future AI services.

“You’re talking on the order of 10 billion GPUs we’re going to need,” he said.

There’s another important factor propelling Nvidia’s spending: the sheer amount of money burning a hole in its pockets. Nvidia’s free cash flow will hit $97 billion in fiscal 2026 and $148 billion in fiscal 2027, which ends in January, according to estimates from S&P Global Market Intelligence. Just two years ago, in fiscal 2024, Nvidia had free cash flow of only $26.9 billion.

The company could find it difficult to spend the money on major acquisitions due to the regulatory environment, leaving investments as a more practical option for spending its growing pile of cash.

Nvidia’s cash has also become more essential as data center projects grow ever larger and more expensive. The total cost of OpenAI’s 10 GW of data centers, for example, will add up to an estimated $500 billion to $600 billion, including land acquisition, power and cooling systems, and construction.

At the center of it all is Huang, to whom other leaders in the industry show unusual forms of deference. For example, when Amazon and Google have news to announce about their in-house AI chip efforts—which they’re developing to lessen their dependence on Nvidia— they’ve learned it’s best to first give a heads-up to Huang, say several people involved in these communications.

The cloud computing businesses of Amazon and Google can’t survive without access to Nvidia’s GPUs. And while Huang tolerates their chip efforts, his partners know he doesn’t like surprises about what they’re up to, said those people.

One possible factor behind Nvidia’s spending spree is fear.

Inside the company, a scary scenario sometimes haunts its leaders: What if big breakthroughs in AI occur that don’t rely on Nvidia’s chips? Some of them believe that could be the beginning of the end for the company, according to two Nvidia executives.

Executives have been especially concerned by the prospect that OpenAI could be the one that achieves a game-changing, Nvidia-free milestone, the two said. For instance, OpenAI is developing its own AI server chip with the aid of Nvidia rival Broadcom, which is due out next year.

Nvidia has similar concerns about Chinese tech giants and AI startups, which is why it has been so keen to remain an active supplier to AI companies in the country, despite U.S. national security concerns, the two executives said. They fear the company’s absence in the market could accelerate the advancement of homegrown alternatives to Nvidia’s GPUs.

Still, Nvidia executives and many others in the AI industry have a hard time taking current efforts to compete with Nvidia’s chips very seriously. For example, when news broke last summer that OpenAI planned to develop its own chips with Broadcom, many Nvidia executives brushed off the threat.

At the moment, Nvidia’s hold on the AI chip market seems nearly unshakable. The company has spent years developing its technologies for data centers into mature products—from its CUDA software to the networking that connects chips to each other—whose foothold makes it challenging for competitors to catch up.

And rivals have found it devilishly hard to produce AI chips that outperform Nvidia’s while providing meaningful cost advantages.

OpenAI and Broadcom, for example, have hit snags with the development of OpenAI’s chip, according to two people involved in the chip production. OpenAI wanted something more powerful and sooner that Broadcom could deliver on its initial effort. OpenAI had been pushing to roll out the chip in the second quarter of next year, but Broadcom has said the third quarter is the earliest it can be ready, one of the people said.

Even when that chip is done, it won’t compete with Nvidia’s general-purpose GPUs, which can handle both the training of AI models and inference—the stage when the models are fully operational. OpenAI’s chip will only handle inference chores.

OpenAI’s initial plans to deploy the chips appear modest, a strategic approach given that this is its first shot at designing server chips for AI workloads. But if the rollout proves successful over subsequent generations, these chips could have greater potential to replace Nvidia’s silicon. On a recent call with investors, Broadcom CEO Hock Tan said an unnamed customer, widely believed to be OpenAI, plans to spend around $10 billion with Broadcom next year, including on chips and servers.

Still, many disruptive technologies start out as the butt of jokes. In 2008, when Apple acquired a boutique microprocessor startup called PA Semi, almost no one in Silicon Valley took Apple seriously as a designer of its own chips. Over nearly two decades, though, Apple began to design more and more of the chips in its devices, including in its Mac computers, ending its relationship with Intel, its previous supplier of microprocessors for the Mac. That event contributed to Intel’s decline.

That lesson has been seared into the minds of Nvidia’s leaders.

“If Jensen doesn’t follow the Andy Grove philosophy of only the paranoid survive, I’d be shocked,” Yoffie said, referring to the former Intel CEO, who died in 2016. “They have to be worried about it.”

While Nvidia is an active startup investor, it has been boosting startups in a far more direct way: by becoming their customer and helping them fund more GPU purchases.

When it comes to upstart cloud providers that aim to compete with Amazon Web Services and other major cloud providers, Nvidia has directly helped them by becoming a customer. Nvidia has signed deals to rent compute capacity from these firms, which has helped them grow their revenue.

It all started in early 2023, when Nvidia was experiencing intense demand for its GPUs, following the launch of ChatGPT a few months earlier. All of the major cloud providers were rationing the chips and couldn’t get enough of them from Nvidia.

Then Huang made a curious move: He decided to lend an assist to a new wave of cloud providers, many of which used to focus on cryptocurrency before shifting to the AI frenzy. In a secretive deal called Project Osprey, Nvidia agreed to spend $1.3 billion over four years to rent its advanced chips back from CoreWeave. The deal made Nvidia CoreWeave’s second-largest customer, behind Microsoft. Nvidia also around the same time invested $100 million in the startup.

The deals helped grow CoreWeave’s revenue from around $25 million in 2022 to nearly $2 billion in 2024.

Nvidia has continued this practice. This summer, it signed landmark deals with cloud provider Lambda to rent back more than 10,000 of its chips. One of the deals was the largest in Lambda’s history—$1.3 billion over four years, similar to Nvidia’s CoreWeave deal. The arrangement with Nvidia, which is also a Lambda investor, could help the firm as it looks to go public.

Some have criticized these deals as a seemingly circular flow of money.

As to why Nvidia needs to rent GPUs in the first place, it has teams of AI researchers who need access to compute—just like all of the other firms working on developing AI, according to several people with knowledge of the compute deals.

But Nvidia also signed these cloud deals so it could build out its own cloud service. Nvidia planned to use some of the compute capacity to forge more direct relationships with enterprise firms. However, earlier this year Nvidia backed off on its efforts to attract businesses to the cloud service, DGX Cloud, and today it primarily uses the chips internally. Many cloud firms see Nvidia’s cloud service as competitive with their own offerings, and in some cases they have felt pressured to participate in it.

Pressure to curb pharmaceutical companies’ pricing power is weighing on the sector

It’s becoming increasingly clear that pharmaceutical companies can live with President Trump’s tariffs. What the industry can’t live with is uncertainty on drug prices.

The sector has long traded at a discount to the broader market, but that gap has widened to its largest in decades as investors fret over Trump’s policies—ranging from tariffs and price controls to the unpredictable influence of Health and Human Services Secretary Robert F. Kennedy Jr.

Some risks now look less pressing. Drug approvals are still moving through the Food and Drug Administration without major disruption, and this past week brought relief on tariffs. Trump announced a 100% levy on pharmaceutical imports from companies not building U.S. plants. In practice, that was close to a best-case outcome, notes Bernstein analyst Courtney Breen. Most big drugmakers are already expanding U.S. manufacturing, which should shield them from steep duties. Altogether, they have pledged more than $350 billion in domestic investments by the end of the decade.

Wall Street wasn’t yet ready to pop the Champagne on Friday: The NYSE Arca Pharmaceutical Index, which consists mainly of large-cap drugmakers, closed up a modest 1%.

That is because the bigger risk is drug pricing, and investors fear it could come to a head as soon as this coming week. Proposals being contemplated to tie U.S. prices to those in other wealthy nations would cut into the industry’s most reliable profit source.

Drugmakers have tried to appease the administration by offering direct-to-patient prices that circumvent middlemen while launching new products overseas at higher prices. But on the issue that matters most—the level of discounts to U.S. prices—they have made little headway.

In July, Trump set a Sept. 29 deadline for companies to lower U.S. prices to the lowest charged in other developed markets, warning that if they didn’t, the government would “deploy every tool in our arsenal to protect American families from continued abusive drug pricing practices.” Reports this week suggest the White House is weighing pilot programs under the Centers for Medicare and Medicaid Services Innovation Center to benchmark some Medicare and Medicaid prices closer to global levels. This resembles ideas from Trump’s first term, efforts that sparked industry lawsuits and are likely to do so again.

Investors are also worried about how Trump might use the drug-pricing powers created under the Biden administration. The Inflation Reduction Act allows Medicare to negotiate prices on top-selling medicines. Negotiated prices for the next list of 15 drugs subject to discounts in 2027 are due to be released in November. Analysts fear the Trump administration could apply the law’s framework more aggressively than the Biden administration, seeking discounts of 30% to 50% on products such as Novo Nordisk’s Ozempic and Teva’s Austedo, according to J.P. Morgan’s Chris Schott.

For Wall Street, the problem isn’t simply lower prices on some drugs. Targeted cuts on a handful of drugs can be absorbed by an industry where the biggest players still run net profit margins in the midteens, well above the corporate average. What unnerves investors is the lack of clarity. They don’t know exactly which drugs will be targeted, how deep the discounts will be and how quickly they will take effect.

Some hit to margins now looks inevitable, but much of that risk is already priced in. The NYSE Arca Pharmaceutical Index trades at just 13.6 times forward earnings, versus 23 for the S&P 500—a discount of more than 40%. That gap suggests the sector could rally once investors know the size of the pain.

Big Pharma is learning to live with the Trump administration. But Wall Street won’t celebrate until investors can see the full price tag.

A growing body of research is finding that some of the most commonly held beliefs about sex are wrong

A young newlywed recently confided to me that he’s worried about his sex life: “I know we’re supposed to do it at least every other day,” he said. “But we can’t keep up.”

Who wants to tell him?

Some of the most commonly held beliefs about sex are incorrect, according to a growing body of research. Recent studies reveal new insights about desire—how to spark and maintain it. Others are overturning long-held ideas about the ideal frequency for physical intimacy.

Misconceptions about sex are often passed along between friends, parroted in chat rooms or spread by uninformed social-media influencers. Even well-meaning doctors and therapists who lack specific training in sexual issues sometimes pass along misleading advice.

“We have a sexual literacy problem in our culture,” says Justin Garcia, executive director of the Kinsey Institute at Indiana University. “And in the absence of good information, we create a mythology and treat it as gospel.”

We shouldn’t be surprised. Sex education and research are perpetually underfunded and often under attack from politicians and religious groups, Garcia says.

Our own reticence plays a role, too. Many people are uncomfortable talking about sex.

Yet good relationships benefit from good information. With that in mind, here are several commonly held beliefs about sex that recent research has debunked.

You should be having more sex

Many people believe that happy couples have a lot of sex—and that those who don’t are headed for relationship trouble.

“No matter how much sex we are having, we think we should be having more,” says Gurit Birnbaum, a professor of psychology at Reichman University in Israel, who studies sex and relationships.

That notion is false!

Different couples have different sexual needs. Some prefer a lot; others, less. A couple’s ideal frequency also changes over time, with age, length of relationship, illness and stress all having an impact.

Still, there’s a sweet spot: People who have sex an average of once a week report greater relationship and life satisfaction than those who have it less frequently, according to research included in a review of 279 studies on sexuality, published earlier this year in Nature Reviews Psychology and co-written by Birnbaum. And—surprisingly—those who have sex more than that don’t report being any happier.

This might be because once a week is enough to maintain connection, the researchers say. More might become routine—and exhausting. Also, the idea of a sexual afterglow is real: People remain more satisfied with their relationship for days after they have sex, research shows.

“You don’t have to have sex every day to feel the intimacy and the closeness,” Birnbaum says.

Keep fantasies to yourself

Almost everyone—97% of the population—reports having sexual fantasies, according to Justin Lehmiller, a senior research fellow at the Kinsey Institute who, since 2014, has been conducting one of the most comprehensive studies ever done on fantasies. Yet therapists, he says, sometimes caution people against sharing them with a partner, warning that it could create unnecessary conflict or even doom the relationship.

Lehmiller’s research shows that the opposite is true: Most people report positive experiences when they share their fantasies—even when the fantasy is about opening up a monogamous relationship to other people. (After all, fantasies don’t have to be acted upon!) Sharing can bring partners closer, create more excitement in the bedroom, and lead to greater relationship and sexual satisfaction, he says.

Deciding whether to share depends on the partner, the relationship and the fantasy, Lehmiller says. To determine what’s right for you, he suggests asking why you want to share. Is it to jazz things up in the bedroom, help your partner understand you better, or act out your fantasy? “It will help you have clarity and present it to your partner in the right way,” he says.

And remember: You don’t have to share every fantasy you have, especially if you think it might be a turn off or make your partner feel insecure.

“Let’s say you have a fantasy about your partner’s best friend,” Lehmiller says. “You may want to keep that to yourself.”

The best sex is spontaneous

Watch any movie with a sex scene and you’ll likely see two people, overcome with lust, ripping each other’s clothes off. No wonder most people believe that the most passionate, satisfying sex just happens.

And perhaps it does if you’re newly in love or you don’t have a job, mortgage or kids. For the rest of us, though, it takes a little planning.

Here’s the good news: People who plan for sex—think date nights—find it just as good as spontaneous sex, research shows. And one soon-to-be published study finds that it’s sometimes even better: When participants were told about the benefits of scheduling sex, they had it more often and enjoyed it more.

“Planning can build anticipation,” says Amy Muise, a professor of psychology at York University in Toronto and director of the Sexual Health and Relationships Lab, who co-wrote the research. “And it shows that it’s a priority.”

After all, most of the important things in our life are planned.

To make scheduling sex seem less mundane, try thinking of your sessions as trysts, says Laurie Mintz, a sex therapist and emeritus professor of psychology at the University of Florida. She often advises clients to picture their early dates with their partner—the way they put on perfume and sexy underwear, flirted shamelessly all night and had sex at the end of the evening.

“That was never unplanned,” she says. “It was so well-orchestrated that you just tricked yourself into believing that it was spontaneous.”

The rivalry could be one for the ages—up there with the feud between Bill Gates and the late Steve Jobs

When Mark Zuckerberg walked on stage the other day with those chunky black AI glasses, some viewed a possible future rival for the Apple iPhone.

Others, however, saw the potential for a new front in a battle with Elon Musk over robots.

In Silicon Valley these days, it can be hard to keep track of how the various tech lords are trying to get a leg up on each other, and what they’re competing over at any given moment. The Musk-Zuckerberg rivalry—once headed to the Colosseum for a cage fight—could be one for the ages, up there with the feud between Bill Gates and the late Steve Jobs.

So far, however, the stakes between them have been comparably small. Mostly, just bragging rights about the sizes of their manliness or fortunes. The 2023 introduction by Meta Platforms META -0.69%decrease; red down pointing triangle of a text-based social-media platform, Threads, to compete with Musk’s X added some meat to things.

Now, things could get really interesting if Zuckerberg’s Meta and Musk’s Tesla TSLA 4.02%increase; green up pointing triangle are headed toward a collision involving humanoid robots.

Thanks in part to Musk’s showmanship, the prospect of humanoid robots is one of the hottest fields in tech.

Long the stuff of sci-fi fantasy, the technology holds the promise of becoming more real with recent advances in artificial intelligence. Whereas robots previously could be programmed to perform specific tasks, such as stacking boxes, it was a rigid and narrow set of abilities. Now, roboticists say they can use AI that learns from watching humans perform tasks, greatly opening up the use cases for robots.

“The promise of that is basically now we can very quickly add a whole host of new tasks where anything becomes fair game,” Jeff Cardenas, co-founder of a humanoid robot company called Apptronik, told me for Friday’s “Bold Names” podcast. “And that kind of changes everything about what robots are capable of and where they can be applied in the future.”

Part of the bull case for why vehicle maker Tesla is so well positioned in robotics owes to its work to develop computer vision to help robot cars see and navigate the world. It already has eight million vehicles on the roadways, collecting video data that helps improve its systems.

A natural evolution, Musk has argued, is humanoid robots.

In May, Milan Kovac, then a Tesla vice president, provided some insight into how the company was working to develop its robot, dubbed Optimus. “One of our goals is to have Optimus learn straight from internet videos of humans doing tasks. Those are often 3rd person views captured by random cameras etc.,” he posted on X. “We recently had a significant breakthrough along that journey, and can now transfer a big chunk of the learning directly from human videos to the bots (1st person views for now).”

With its own first-person views, Meta’s glasses, too, could be well positioned to collect reams and reams of video data crucial for robot development.

The newest version of those glasses goes on sale Tuesday and includes a video display within the lens and a built-in camera able to capture what the user sees. Meta tells users that certain video and audio data can be used for improving its products.

One of the demonstrations during the Meta event on Sept. 17 was supposed to include a chef cooking food for a party with the help of instructions from the AI watching him through the glasses. A technical hiccup, however, ended that effort.

Still, Adam Jonas, the well-known Morgan Stanley analyst, argues that within two years Meta could have 20 million pairs of its glasses in use around the world—almost twice the number of Tesla vehicles expected to be on the roadways by then.

“Every Meta glasses user may be training a humanoid avatar iterated in simulation across billions of scenarios in a digital omniverse,” Jonas wrote to investors in a note this past week.

In other words, Meta could be sitting on a gold mine of the world’s most boring home videos that could be useful to robots trying to learn how to navigate domestic life.

That’s when things could get really meta. Imagine: The Meta AI instructing you step-by-step how to make dinner while in turn gathering video to later train a robot how those instructions manifest in the physical world.

While Meta hasn’t said if Live AI, what it calls its feature in the consumer glasses, will be used to gather data for robot training, it has other wearables aimed at just that. The new generation of its Project Aria research glasses, announced earlier this year, were designed, according to the company, to gather data for AI and robotics.

By 2040, Musk predicts, there will be at least 10 billion humanoid robots in the world, remaking the idea of work and life—even if some roboticists say such technology is still a work in progress.

So far, Zuckerberg’s ambitions for robots are less clear.

Meta’s efforts are still in the early stages, having emerged from its broader work in AI and extended reality devices. “If you’re thinking about having an AI that is always on, using cameras and microphones to assess the situation that it’s faced with, to be an assistant to you through a wearable device—that’s actually pretty similar to a robot who’s got the same kind of packages in terms of how it should understand the world,” Andrew Bosworth, Meta chief technology officer, said in a social-media post in June.

Earlier this year, Meta hired Marc Whitten, the former CEO of General Motors’ autonomous-car subsidiary Cruise, to oversee a new robotics effort. Among other recent recruits is Sangbae Kim, a Massachusetts Institute of Technology robotics professor, to lead research and development as Meta’s robotics architect. He had attracted attention for his MIT lab’s development of a small four-legged “cheetah” robot that could run quickly.

Those hires have garnered far less attention than the more-recent spate of big AI names—and a recruitment drive that has included hundred-million-dollar pay packages. That new effort is aimed at helping Meta catch up with the likes of OpenAI and others in developing advanced AI—“superintelligence” as Zuckerberg calls it. He has said such technology is at the heart of the glasses and his effort to enable a new computing paradigm.

Amid worries of an AI bubble, Zuckerberg says he’s trying to make the right bet.

“If we end up misspending a couple of hundred billion dollars, I think that is going to be very unfortunate obviously, but what I’d say is I actually think the risk is higher on the other side,” Zuckerberg said on the “ACCESS” podcast earlier this month. “If you build too slowly…then you’re just out of position on what I think is going to be the most important technology that enables the most new products and innovation and value creation in history.”

Whatever the cost, a robot cage match could be fun.