Nvidia

US and China have long competed to become world powers, particularly in the technology sector. Since 2022, the US has systematically restricted the supply of high-performance NVIDIA chips to China. In today’s world, competition for power is no longer achieved through traditional means, such as military power. The US uses chips (semiconductors) as an instrument of political pressure. This policy is not just about economic or trade value, but has become part of technological statecraft designed to counter China’s military potential and its use of Artificial Intelligence (AI).

Semiconductors as a Provision of Power

The US policy of restricting high-end semiconductors to China shows a paradigm shift, chips (semiconductors) are not only industrial commodities, but have shifted to become a tool for achieving power. Export controls on high-performance chips and components that enable their production have been implemented by the Bureau of Industry and Security (BIS). These steps show that the US is restructuring the geopolitical arena of technology.

AI today relies heavily on chips that can process vast amounts of data. The US restricts the export of high-end chips, such as the NVIDIA H100 and A100. A country’s AI development capacity could be severely compromised without access to these chips. The H100 is more than just a technological component; it serves as a strategic enabler that determines a country’s ability to maintain military dominance.

NVIDIA and the Security Logic Behind Export Control

The Bureau of Industry and Security (BIS) on 2023 announcement expanded export oversight, not only targeting on specific chip models but also on component values, most notably in frontier algorithm development. The NVIDIA A100 and H100 are highly advanced datacenter and AI chips. The guidelines are particularly high for training complex AI models on supercomputers, even for military applications or demanding research.

To prevent misuse, the US government has implemented licensing requirements for chips like the A100 and H100 chips, which have put chips like the A300 and H800, made by NVIDIA, under increased scrutiny, despite being categorized as “weak service” chips. Export restrictions stem from concerns that NVIDIA GPUs could be used by China in training AI models related to the US military, not only to slow China’s technological progress but also to safeguard its own national interests.

The US understands very well that high-performance chips are “brain machines” that can accelerate the development of military superiority, intelligence analysis, and even autonomous systems. So it is very clear that limiting the capacity of computing and high-performance hardware is the way to go. To delay a rival’s capabilities without resorting to direct military confrontation. This is a concrete manifestation of the shift in the “battlefield” taking place in the technological and regulatory arenas.

Vulnerable Supply Chains and Dependence on Taiwan

In chip control, the US must recognize that there are undeniable realities. NVIDIA’s chip production goes through a fabrication process that is almost entirely carried out in Taiwan, a country that lies in the geopolitical conflict between Washington and Beijing. The Congressional Research Service (2024) shows that approximately 90% of global advanced semiconductor chip production is based in Taiwan, manufactured by the leading Taiwanese foundry, Taiwan Semiconductor Manufacturing Company Ltd. (TSMC). This creates a structural dependency that poses serious risks to US economic and technological security.

If semiconductor production were concentrated in a single region, it would create vulnerabilities that could destabilize the global technological system. Therefore, any tensions in the Taiwan Strait would disrupt US access to the computing infrastructure it maintains. Export restrictions are just one step in a much more complex strategy, requiring the US to diversify production locations and ensure that the chip supply chain is not concentrated in a single region.

Effectiveness and Adaptation Room for China

NVIDIA’s chip restrictions were intended to curb the pace of AI modernization in China, but China was still able to optimize the model’s efficiency. This demonstrates that limiting hardware performance doesn’t always equate to limiting innovation. On the other hand, unofficial market entities have emerged, allowing NVIDIA GPUs to remain accessible through third parties. This adaptation demonstrates that hardware control has limitations, especially when demand remains high and global distribution networks are not always transparent.

Looking at its overall effectiveness, US policy has been effective in slowing China’s computing capabilities, but it hasn’t stopped its strategic potential. Instead, it’s encouraging China to be self-sufficient in strengthening its technological foundation, even though the quality of local chips hasn’t yet matched NVIDA’s standards. In other words, restricting NVIDIA’s chip exports isn’t meant to end competition, but rather to transform it into a race toward technological independence. The policy’s effectiveness will only last as long as China finds a way to adapt, while China is working to fill that gap.

Policy Directions with Greater Strategic Opportunities

The effectiveness of the compute policy is based on a governance architecture that holds every allied country accountable to the same standards. Without a disciplined framework, export controls on China are merely an illusion that is easily penetrated by gaps in different economic and regulatory interests. By creating strategic alignment, which forces every democratic country to reduce the fragmentation of interests, it can open up greater policy opportunities to emerge. Many developing countries see this semiconductor race as a competition for dominance, not as an effort to maintain security.

In other words, a successful computing policy is not one that simply limits China’s space, but one that manages technological gaps without creating competing computing blocs. The geopolitical challenge is maintaining superiority without forcing the world into two technological divides that would be difficult to control. The US strategy to secure a leading position in future technologies requires flexibility in responding to global dynamics.

A Future Determined by Computational Capacity

The debate over NVIDIA chips demonstrates the growing integration of political and technological power. US policy aims not only to restrain the flow of strategic goods but also to build a new computing-based power architecture. However, this policy also presents challenges, including dependence on Taiwan, China’s flexibility, and economic pressure on US chip companies.

In a global world that continues to move toward an AI-driven economy, the future will be determined by who can manage geopolitical risks, understand supply chain dynamics, and design visionary policies. Ultimately, GPU regulation is no longer simply a matter of export control; it demonstrates how countries navigate a power struggle now measured in microchips.

Nov. 21 (UPI) — Federal authorities have arrested four people, including two Chinese nationals, on accusations of scheming to illegally export cutting-edge Nvidia technology with artificial intelligence uses to Beijing, which prosecutors said seeks to be the AI world leader by the end of the decade.

Federal authorities in Tampa, Fla., on Wednesday arrested 34-year-old Hong Kong-born U.S. citizen Hong Ning Ho, also known as Matthew Ho, and 45-year-old Jing Chen, also known as Harry Chen, who was in the United States on a F-1 nonimmigrant student visa.

Brian Curtis Raymond, 46, of Huntsville, Ala., and 38-year-old Cham Li, also know as Tony Li, a Chinese national, were also arrested, though when was not clear.

Federal prosecutors alleged in an indictment — unsealed Wednesday but publicized by the Justice Department on Thursday — that from September 2023 until their arrests, the defendants conspired to illegally export NVIDIA graphics processing units to China through Malaysia and Thailand.

“The indictment unsealed yesterday alleges a deliberate and deceptive effort to transship controlled NVIDIA GPUs to China by falsifying paperwork, creating fake contracts and misleading U.S. authorities,” John Eisenberg, assistant attorney general for National Security said in a statement.

The court documents statement that they attempted four separate exports consisting of hundreds of GPUs. The first two shipments saw 400 Nvidia A100 GPUs being exported to China between October 2024 and January. The other two shipments of 50 Nvidia H200 GPUs and 10 Hewlett Packard Enterprises supercomputers with Nvidia H100 GPUS were intercepted by authorities.

In return for the shipments, the defendants allegedly received more than $3.89 million in wire transfers, according to the indictment.

The indictment states they used Tampa-based Janford Realtor, owned by Ho and Li, as a front company to buy the goods and export them to China.

Federal prosecutors alleged that despite being labeled a real estate company, it was involved involved in property transactions.

The court document accuses Raymond of supply the GPUs to Ho through his Alabama-based electronics company.

According to federal prosecutors, China is seeking to become the world leader in AI by 2030 and seeks to use the technology for military modernization efforts, including designing and testing its weapons of mass destruction as well as surveillance tools.

Published on
20/11/2025 – 7:32 GMT+1

Shares in Nvidia rose more than 5% in after-hours trading after the chipmaker beat analysts’ expectations in its quarterly earnings report, released Wednesday.

In the three months to the end of October, Nvidia said its revenue jumped 62% to $57 billion (€49.49bn). The company reported $51.2bn (€44.43bn) in revenue from data-centre sales, beating expectations of $49bn (€42.52bn).

The firm also placed a forecast for the current quarter at $65bn (€56.41bn), surpassing Wall Street expectations of $61bn (€52.94bn).

“There’s been a lot of talk about an AI bubble,” said CEO Jensen Huang during an earnings call.

“From our vantage point, we see something very different. As a reminder, Nvidia is unlike any other accelerator. We excel at every phase of AI from pre-training to post-training to inference.”

Nvidia is now the largest stock on Wall Street, having momentarily surpassed $5 trillion in value. That means it has an outsized influence on the S&P 500 and can make or break the market’s daily performance.

The firm has also become a bellwether for the broader frenzy around AI, notably because other companies rely on Nvidia chips for this technology.

AI stocks have taken a hit in recent weeks as investors questioned whether certain tech companies had been overvalued, driving fears of a market crash.

Before Wednesday’s earnings report, Nvidia’s chips had dropped 11% from their peak in early November.

CEO Huang sought to ease concerns of a bubble on Wednesday, claiming: “AI is going everywhere, doing everything, all at once.” He noted that Nvidia was focused on major transition areas, namely generative, agentic, and physical AI.

Generative AI can create things, agentic can accomplish a specific goal with limited supervision, while physical AI relates to the physical world — for example through robots.

Nov. 19 (UPI) — Tech giant Nvidia on Wednesday posted record revenue and strong profit for the third quarter, beating Wall Street expectations, amid exploding growth in artificial intelligence.

Nvidia, which has the world’s largest market capitalization at $4.5 trillion, reported record sales. It said sales grew 62% in one year to $57 billion through Oct. 26. Wall Street had projected a $54.9 billion figure.

On Oct. 29, Nvidia became the first company worldwide with a $5 trillion cap one day before CEO Jensen Huang met with President Donald Trump in the White House.

“There’s been a lot of talk about an AI bubble. From our vantage point, we see something very different,” Huang said during a conference call with investors.

Fourth-quarter sales are estimated to be around $65 billion, contrasting with $61.66 billion by analysts.

Profit was up 65% from last year in the quarter to $31.9 billion or 78 cents per share, slightly ahead of expectations. The net income represents 58% of revenue.

NVIDIA will pay its next quarterly cash dividend of 1 cent per share on Dec. 26.

Nvidia builds chips and software platforms for the AI industry. The company, founded in 1993 in the Silicon Valley in California, pioneered the graphics processing unit, initially for 3D video games.

The chips are made in the United States by GlobalFoundries, Taiwan Semiconductor Manufacturing Company and Samsung in South Korea. Taiwan’s new factory in Arizona focuses on chips for Nvidia.

The design work is done in the United States, GeekBitz reported.

Most AI companies’ technology runs on Nvidia’s chip, CNN reported.

Its best-selling chip is the Blackwell Ultra, a second generation. The company is banned from selling the new ones to China.

“Blackwell sales are off the charts, and cloud GPUs are sold out,” Huang said in a statement about its best-selling chip.

“Compute demand keeps accelerating and compounding across training and inference — each growing exponentially. We’ve entered the virtuous cycle of AI. The AI ecosystem is scaling fast — with more new foundation model makers, more AI startups, across more industries, and in more countries. AI is going everywhere, doing everything, all at once.”

In October, Huang said there were $500 billion in AI chip orders for 2025 and 2026 combined.

“The number will grow,” Nvidia finance chief Colette Kress said during the earnings call with analysts.

Nvidia said there were $51.2 billion in revenue in data center sales, a 66% rise year-over-year.That includes $43 billion in revenue was for “compute,” or the GPUs. The company said most growth was from GB300 chips.

Nvidia’s stock price rose 5.08% in after-hours trading on Wednesday night on NASDAQ. The stock was at $196.00, below the record $207.04 on Oct. 29.

The stock, with the ticker symbol NVDA, initially traded at $12 per share, through its Initial Public Offering on Jan. 22, 1999.

The strong Nvidia report boosted after-hours trading of tech firms Meta, Microsoft, Amazon and Google.

“This answers a lot of questions about the state of the AI revolution, and the verdict is simple: it is nowhere near its peak, neither from the market-demand nor the production-supply-chain sides for the foreseeable future,” Thomas Monteiro, senior analyst at Investing.com, said in emailed commentary following the report.

In September, Nvidia announced a $100 billion investment in OpenAI in exchange for chip purchases.

On Monday, Anthropic committed to buying $30 billion in computing capacity from Microsoft Azure in exchange for an investment in the AI lab from both tech giants.

Nvidia announced a collaboration with Intel to jointly develop multiple generations of custom data center and PC products with NVIDIA NVLink.

Nvidia has reviewed plans to accelerate seven new supercomputers, including with Oracle to build the U.S. Department of Energy’s largest AI supercomputer, Solstice, plus another system, Equinox.

Nvidia said it had $4.3 billion in gaming revenue, which is a 30% boost from one year ago.

Despite the boom, CEO of one of the world’s largest independent financial advisory organizations warnsthere is a “real risk” because of complacency.

“Exceptional results don’t remove the need for discipline,” Nigel Green of deVere Group in Britain said in an email to UPI. “The AI ecosystem is growing fast, but fast growth doesn’t protect anyone from the consequences of over-extension.”

He said the path from deployment to real commercial returns “remains untested” in many industries.

“Investors must examine whether business models can convert this scale of capital investment into sustained earnings,” he said. “Complacency could be a real risk.”

Analysts expect AI chip demand to remain strong.

Nvidia has forecast fourth-quarter revenue above Wall Street estimates and is betting on booming demand for its AI chips from cloud providers even as widespread concerns of an artificial intelligence bubble grow stronger.

The world’s most valuable company expects fourth-quarter sales of $65bn, plus or minus 2 percent, compared with analysts’ average estimate of $61.66bn, according to data compiled by LSEG.

The results from the AI chip leader mark a defining moment for Wall Street as global markets look to the chip designer to determine whether investing billions of dollars in AI infrastructure expansion has resulted in towering valuations that potentially outpace fundamentals.

“The AI ecosystem is scaling fast with more new foundation model makers, more AI start-ups across more industries and in more countries. AI is going everywhere, doing everything, all at once,” Nvidia CEO Jensen Huang said in a statement.

Before the results, doubts had pushed Nvidia shares down nearly 8 percent in November after a 1,200 percent surge in the past three years.

Sales in the data-centre segment, which accounts for a majority of Nvidia’s revenue, grew to $51.2bn in the quarter that ended on October 26. Analysts had expected sales of $48.62bn, according to LSEG data.

Warning signs

But some analysts noted that factors beyond Nvidia’s control could impede its growth.

“While GPU [graphics processing unit] demand continues to be massive, investors are increasingly focused on whether hyperscalers can actually put this capacity to use fast enough,” said Jacob Bourne, an analyst with eMarketer. “The question is whether physical bottlenecks in power, land and grid access will cap how quickly this demand translates into revenue growth through 2026 and beyond.”

Nvidia’s business also became increasingly concentrated in its fiscal third quarter with four customers accounting for 61 percent of sales. At the same time, it sharply ramped up how much money it spends renting back its own chips from its cloud customers, who otherwise cannot rent them out, with those contracts totalling $26bn – more than double their $12.6bn in the previous quarter.

Still, analysts and investors widely expected the underlying demand for AI chips, which has powered Nvidia results since ChatGPT’s launch in late 2022, to remain strong.

Nvidia CEO Jensen Huang said last month that the company has $500bn in bookings for its advanced chips through 2026.

Big Tech, among Nvidia’s largest customers, has doubled down on spending to expand AI data centres and snatch the most advanced, pricey chips as it commits to multibillion-dollar, multigigawatt build-outs.

Microsoft last month reported a record capital expenditure of nearly $35bn for its fiscal first quarter  with roughly half of it spent primarily on chips.

Nvidia expects an adjusted gross margin of 75 percent, plus or minus 50 basis points in the fourth quarter, compared with market expectation of 74.5 percent.

By&nbspEuronews

Published on
19/11/2025 – 12:14 GMT+1

European stocks showed mixed signals on Wednesday, somewhat easing fears of a global market crash.

At around midday, Germany’s DAX was up less than 1%, while the UK’s FTSE 100 and Spain’s IBEX 35 also saw modest lifts.

Italy’s FTSE MIB dropped less than 1%, as did France’s CAC 40.

Both the STOXX 50 and the wider STOXX 600 showed minimal movement.

Investors kept an eye on data releases on Wednesday, with UK inflation easing to 3.6% in October, down from 3.8% in July, August, and September.

The annual inflation rate in the eurozone, meanwhile, came in at 2.1% in October, a confirmation of a preliminary reading. That’s down from 2.2% in September.

“Investors will breathe a sigh of relief that the market sell-off has lost momentum,” said Russ Mould, investment director at AJ Bell.

“It’s the good news everyone wanted. The key question is whether this is simply the calm before the storm.”

In Asian trading on Wednesday, markets were broadly in the red.

Japan’s Nikkei 225 fell 0.34%, Hong Kong’s Hang Seng was down 0.38%, South Korea’s Kospi slid 0.61%, while Australia’s S&P/ASX 200 slid 0.25%. China’s SSE Composite rose 0.18%.

After a day of losses on Tuesday, Wall Street showed signs of optimism on Wednesday.

Ahead of the opening bell, S&P 500 futures were up 0.30%, while Dow Jones futures increased 0.12%. Nasdaq futures were trading 0.37% higher.

Investors around the world are awaiting third-quarter results from chipmaker Nvidia, set for release later on Wednesday.

Nvidia’s performance matters disproportionately because its immense size means it’s the most influential stock on Wall Street. Its financial report will also influence the narrative around an AI bubble and fears that tech stocks may be overvalued.

“Nvidia reports tonight and the slightest bit of news to disappoint investors has the potential to whip up a tornado across global markets,” said Mould.

“Investors will be hanging on Jensen Huang’s every word and looking for clues that big investment in AI is worth it.”

The announcement underscores AI industry’s insatiable appetite for computing power as companies race to build systems that can rival or surpass human intelligence.

Microsoft and Nvidia plan to invest in Anthropic under a new tie-up that includes a $30bn commitment by the Claude maker to use Microsoft’s cloud services, the latest high-profile deal binding together major players in the AI industry.

Nvidia will commit up to $10bn to Anthropic and Microsoft up to $5bn, the companies said on Tuesday, without sharing more details.

A person familiar with the matter said both the companies have committed to investing in Anthropic’s next funding round.

The announcement underscores the AI industry’s insatiable appetite for computing power as companies race to build systems that can rival or surpass human intelligence. It also ties major OpenAI-backer Microsoft, as well as key AI chip supplier Nvidia, closer to one of the ChatGPT maker’s biggest rivals.

“We’re increasingly going to be customers of each other. We will use Anthropic models, they will use our infrastructure and we’ll go to market together,” Microsoft CEO Satya Nadella said in a video. He added that OpenAI “remains a critical partner”.

The move comes weeks after OpenAI unveiled a sweeping restructuring that moved it further away from its non-profit roots, giving it greater operational and financial freedom.

The startup has since then announced a $38bn deal to buy cloud services from Amazon.com as it reduces reliance on Microsoft. Its CEO, Sam Altman, has said OpenAI is committed to spending $1.4 trillion to develop 30 gigawatts of computing resources – enough to roughly power 25 million US homes.

Still, three years after ChatGPT’s debut, investors are increasingly uneasy that the AI boom has outrun fundamentals. Some business leaders have noted that circular deals – in which one partner props up another’s revenue – add to the bubble risk.

“The main feature of the partnership is to reduce the AI economy’s reliance on OpenAI,” D A Davidson analyst Gil Luria said of Tuesday’s announcement.

“Microsoft has decided not to rely on one frontier model company. Nvidia was also somewhat dependent on OpenAI’s success and is now helping generating broader demand.

AI industry consolidating

Founded in 2021 by former OpenAI staff, Anthropic was recently valued at $183bn and has become a major rival to the ChatGPT maker, driven by the strong adoption of its services by enterprise customers.

The Reuters news agency reported last month that Anthropic was projecting to more than double and potentially nearly triple its annualised revenue run rate to around $26bn next year. It has more than 300,000 business and enterprise customers.

As part of Tuesday’s move, Anthropic will work with Nvidia on chips and models to improve performance and commit up to 1 gigawatt of compute using Nvidia’s Grace Blackwell and Vera Rubin hardware. Industry executives estimate that one gigawatt of AI computing can cost between $20bn and $25bn.

Microsoft will also give Azure AI Foundry customers access to the latest Claude models, making Claude the only frontier model offered across all three major cloud providers.

“These investments reflect how the AI industry is consolidating around a few key players,” eMarketer analyst Jacob Bourne said.

Despite the looming deal, Microsoft shares are down 3.2 percent in midday trading. Nvidia is also trading 1.9 percent lower than at the market open, and Amazon has fallen 4 percent. Tech stocks remain under pressure after a cloud services outage earlier on Tuesday. Neither OpenAI nor Anthropic is publicly traded.

An estimated half a trillion dollars was wiped out from the financial markets this week, as some of the biggest tech companies, including Nvidia, Microsoft, and Palantir Technologies saw a temporary but sizeable drop in their share prices on Tuesday. It may have been just a short-lived correction, but experts warn of mounting signs of a financial market crash, which could cost several times this amount.

With dependence on tech and AI growing, critics argue that betting on these profits is a gamble, stressing that the future remains uncertain.

Singapore’s central bank joined a global chorus of warnings from the IMF, Fed Chair Jerome Powell, and Andrew Bailey about overvalued stocks.

The Monetary Authority of Singapore said on Wednesday that such a trend is fuelled by “optimism in AI’s ability to generate sufficient future returns”, which could trigger sharp corrections in the broader stock market.

Goldman Sachs and Morgan Stanley predict a 10–20% decline in equities over the next one to two years, their CEOs told the Global Financial Leaders’ Investment Summit in Hong Kong, CNBC reported.

Experts interviewed by Euronews Business also agree that a sizeable correction could be on the way.

In a worst-case scenario, a market crash could wipe out trillions of dollars from the financial markets.

According to Mathieu Savary, chief European strategist at BCA Research, Big Tech companies, including Nvidia and Alphabet, would cause a $4.4 trillion (€3.8tn) market wipeout if they were to lose just 20% of their stock value.

“If they go down 50%, you’re talking about an $11tr (€9.6tr) haircut,” he said.

AI rally: Bubble or boom?

The US stock market has defied expectations this year. The S&P 500 is up nearly 20% over the past 12 months, despite geopolitical tensions and global trade uncertainty driven by Washington’s tariff policies. Gains have been strongest in tech, buoyed by optimism over future AI profits.

While Big Tech continues to deliver, with multibillion-dollar AI investments and massive infrastructure buildouts now routine, concerns are growing over a slowing US economy, compounded by limited data during the government shutdown. Once fresh figures emerge, they could rattle investors.

AI enthusiasm is most evident in Nvidia’s extraordinary stock gains and soaring valuation. The company is central to the tech revolution as its graphics processing units (GPUs) are essential for AI computing.

Nvidia’s shares have surged over 3,000% since early 2020, recently making it the world’s most valuable public company. Between July and October alone, it gained $1tr (€870bn) in market capitalisation — roughly equal to Switzerland’s annual GDP. Its stock trades at around 45 times projected earnings for the current fiscal year.

Derren Nathan, head of equity research at Hargreaves Lansdown, said: “Much of this growth is backed by real financial progress, and despite the massive nominal increase in value, relative valuations don’t look overstretched.”

Analysts debate whether the current market mirrors the dot-com bubble of 2000. Nathan notes that many tech companies that failed back then never reached profitability, unlike today’s giants, which generate strong revenues and profits, with robust demand for their products.

Ben Barringer, global head of technology research at Quilter Cheviot, added: “With governments investing heavily in AI infrastructure and rate cuts likely on the horizon, the sector has solid foundations. It is an expensive market, but not necessarily a screaming bubble. Momentum is hard to sustain, and not every company will thrive.”

BCA Research sees a bubble forming, though not set to burst immediately. Chief European strategist Mathieu Savary said such bubbles historically peak when firms begin relying on external financing for large projects.

Investments in assets for future growth, or capital expenditures, as a share of operating cash flow, have jumped from 35% to 70% for hyperscalers, according to Savary. Hyperscalers are tech firms such as Microsoft, Google, and Meta that run massive cloud computing networks.

“The share of operating earnings is likely to move above 100% before we hit the peak,” Savary added. This means that they may soon be investing more than they earn from operations.

Recent examples of Big Tech firms turning to external financing for such moves include Meta’s Hyperion project with Blue Owl Capital and Alphabet’s €3 billion bond issue for AI and cloud expansion.

While AI investment growth is hard to sustain, Quilter’s Barringer told Euronews: “If CapEx starts to moderate later this year, markets may start to get nervous.”

Other factors to watch include return on invested capital and rising yields and inflation pressures, which could signal a higher cost of capital and a bubble approaching its end.

“But we’re not there yet,” said Savary.

Further concerns and how to hedge against market turbulence

Even as tech companies ride the AI wave, inflated expectations for future profits may prove difficult to meet.

“The sceptics’ main problem may not be with AI’s potential itself, but with the valuations investors are paying for that potential and the speed at which they expect it to materialise,” said AJ Bell investment director Russ Mould.

A recent report by BCA reflects the mounting reasons to question the AI narrative, but the technology “remains a potent force”, said the group.

If investor optimism does slow, “a sharp correction in tech could still have ripple effects across broader markets, given the sector’s dominant weight in global indices,” Barringer said. He added that other regions and asset classes, such as bonds and commodities, would be less directly affected and could provide an important balance during a downturn.

According to Emma Wall, chief investment strategist at Hargreaves Lansdown, “investors should use this opportunity to crystallise impressive gains and diversify their portfolios to include a range of sectors, geographies and asset classes — adding resilience to portfolios. The gold price tipping up is screaming a warning again — a siren that this rally will not last.”