The AI boom suffered a major market test this week as investors rushed to exit technology stocks, wiping nearly $3 trillion off the market value of S&P 500 companies in a sharp selloff that spread from Wall Street to Asian markets.

The Nasdaq Composite plunged 5.5% on Thursday, its steepest single-day decline in 19 months, pushing the index into correction territory and putting it on course for a fourth consecutive weekly loss. The Nasdaq 100 fell 6.4%, erasing approximately $2.5 trillion in market value.

The panic quickly travelled across global markets. Japan’s Nikkei slumped 6% on Friday, marking its steepest one-day decline since April 2025, while South Korea and Taiwan’s benchmark indices fell more than 8% during early trading.

Taiwan’s index eventually closed 4.5% lower, recording a 12.2% weekly decline, its biggest since 1991. Semiconductor heavyweights Samsung Electronics, SK Hynix and TSMC were among the major casualties, reflecting the extent to which the AI trade had become embedded in global technology valuations.

The selloff also exposed the extent of investor anxiety beyond equities. Wall Street’s CBOE Volatility Index (VIX), often referred to as the market’s fear gauge, recorded its biggest one-day jump in approximately 19 months. Hedge funds and retail investors rushed to reduce exposure, while leveraged positions and margin calls added to the pressure.

As confidence weakened, investors moved towards safer assets, pushing government bond prices higher and drawing demand towards the US dollar, Japanese yen and Swiss franc.

Although technology stocks recovered some ground in Friday’s trading, the sharp reversal has raised questions about whether the AI-led market rally has become too dependent on expectations of future growth.

The US Midterm Elections Become An Unexpected Trigger

The immediate trigger for the selloff was an unexpected reassessment of the political environment surrounding artificial intelligence in the United States.

Democratic victories in the midterm elections earlier this week, including in several important data-centre states, unsettled investors who fear that a shift in political control could bring tougher scrutiny of AI development, tighter regulations and restrictions on the rapid expansion of data-centre infrastructure.

The concern is significant because the AI investment boom is not limited to software developers and chipmakers. It rests on an enormous physical infrastructure expansion involving data centres, computing capacity, semiconductor manufacturing and energy supply.

Much of this investment depends on the ability of technology companies and their infrastructure partners to secure land, power, financing and regulatory approvals. Any political resistance to that expansion could affect projected growth, delay projects and weaken the expected returns on billions of dollars already committed.

The growing opposition to data-centre development at the local level is adding to these concerns. Texas Governor-elect Gina Hinojosa said on Thursday that the state would introduce laws to protect residents in connection with data centres, citing their impact on rural communities.

The comments have reinforced investor concerns that resistance to large-scale infrastructure projects could become a more significant obstacle to the AI industry’s expansion.

However, the political developments have arrived at a time when investor confidence was already weakening over the enormous cost of AI infrastructure and uncertainty about when those investments would become profitable.

The election results may have accelerated the selloff, but they have also brought an existing concern into sharper focus: how much more money can companies pour into the AI race before investors begin demanding tangible financial returns?

Beneath The Crash, Questions About The AI Boom

While the political developments provided an immediate trigger for the selloff, the deeper concern is centred on the economics of the AI boom itself.

Technology companies have committed enormous sums to building the infrastructure needed to support artificial intelligence, from advanced semiconductors and cloud computing capacity to hyperscale data centres.

However, as spending continues to climb, investors are increasingly questioning whether the revenue and profits generated by these investments will be sufficient to justify the scale of capital being deployed.

The concern is particularly pronounced among companies building and operating AI infrastructure. Data centres require substantial upfront investment, long development timelines and access to reliable power, while their commercial viability depends on securing customers willing to pay for computing capacity over extended periods.

For smaller infrastructure providers, the challenge is compounded by higher borrowing costs and tighter financial conditions, making it increasingly difficult to raise capital and sustain expansion plans.

CoreWeave, an AI cloud computing provider that has become something of a barometer for investor sentiment towards the sector, illustrates these pressures. Its shares fell 10% during the selloff, while the cost of insuring its debt against default through five-year credit default swaps widened to a record level.

The movement suggests that investors are increasingly concerned not just about the company’s share price, but also about the financial risks associated with heavily debt-funded AI infrastructure expansion.

Concerns are also emerging around the relationship between major technology companies and the smaller infrastructure providers supporting their AI ambitions. Microsoft has reportedly tightened oversight of certain data-centre projects involving neocloud providers such as Nebius, IREN and Nscale, adding another layer of uncertainty for companies dependent on large technology customers and financing commitments.

The scale of the correction suggests that investors are beginning to distinguish between the long-term potential of artificial intelligence and the financial viability of the companies spending heavily to build it. AI may continue to transform industries, but that does not automatically guarantee that every infrastructure provider, technology company or investor participating in the boom will generate adequate returns.

SoftBank Emerges As A Major Casualty Of The AI Selloff

Among the companies caught in the market turmoil, Japan’s SoftBank Group has emerged as one of the most closely watched casualties. Its shares plunged nearly 11% on Friday, while the cost of insuring its debt against default surged to its highest level in four years, reflecting growing investor unease about its exposure to the AI investment cycle.

The five-year credit default swap spread for SoftBank climbed to 466 basis points, its widest level since October 2022, before settling at 457.50 basis points, according to Markit data. The increase indicates that investors are demanding significantly higher premiums to protect themselves against the possibility of a default. The development is particularly significant because SoftBank has increasingly positioned itself as a major financier of the global AI expansion.

Its exposure is substantial. SoftBank has invested $64.6 billion in OpenAI, holds a 90% stake in chip designer Arm and is financing data-centre development through its subsidiary SB Energy.

These investments have made the Japanese conglomerate an important proxy for investor confidence in the future of artificial intelligence. However, the scale of its commitments has also brought greater scrutiny of its finances, particularly as it has taken on additional debt at record yields to fund its OpenAI investment.

The timing has become more complicated with OpenAI delaying its anticipated public offering until 2027, pushing back expectations of a potential liquidity event for investors. Meanwhile, SB Energy’s proposed public listing, which had reportedly been targeting a valuation of $50 billion, has also been delayed until December or later, according to media reports.

The concern is no longer simply whether SoftBank has backed the right companies in the AI race. Investors are increasingly examining how much debt the group can sustain, how long it will take for its investments to generate returns and whether the anticipated growth of AI infrastructure will justify the capital already committed.

Anthropic’s $1.5 Trillion IPO Valuation Faces Investor Resistance

The growing unease surrounding AI investments is also beginning to spill into the private markets, where some of the sector’s most valuable companies are preparing for potential public listings.

Anthropic, the artificial intelligence startup behind the Claude family of models, is facing questions from prospective investors about whether it can command the recently discussed $1.5 trillion valuation for its proposed initial public offering (IPO).

The concerns come at a particularly sensitive time, with the technology stock selloff and uncertainty over the future of AI infrastructure spending threatening to complicate its plans for a public debut.

According to people familiar with investor discussions, potential investors are pushing for a more attractive valuation before committing to the offering. The company’s proposed roadshow was initially expected to begin during the week of November 9, with pricing anticipated the following week. However, the sharp correction in technology stocks has raised doubts about whether the proposed timeline can be maintained. Anthropic has declined to comment on the investor concerns.

The valuation debate also reflects questions about the company’s financial requirements. Despite its position among the leading AI developers, Anthropic has a record of quarterly losses and substantial financial commitments to partners involved in building the infrastructure needed to train and operate its models. Investors are also waiting for the company’s October revenue figures, which were left blank in its updated S-1 filing, to assess whether it has managed to accelerate growth following a slowdown in recent months.

The potential scale of the offering adds another layer of complexity. Some estimates suggest the IPO could reach $100 billion, making it one of the largest public offerings ever contemplated.

Earlier in the year, investor expectations had placed Anthropic’s potential valuation as high as $2 trillion. The recently discussed $1.5 trillion figure itself represents a significant valuation, but the resistance emerging from prospective investors suggests that private-market enthusiasm may no longer translate automatically into public-market demand.

A delay would also carry implications beyond Anthropic. With large technology companies dominating major stock indices and AI-related investments accounting for a significant part of market enthusiasm, difficulties in taking one of the industry’s most prominent startups public could further unsettle confidence in the wider AI ecosystem.

The Last Bit, Is This A Temporary Correction Or A Deeper Reassessment?

Despite the severity of the selloff, there is no clear consensus on whether the latest market turmoil represents a temporary correction or the beginning of a more sustained reassessment of the AI investment boom.

Some analysts believe much of the decline has been driven by aggressive position unwinding, leveraged trading and investors rushing to protect profits after months of gains. The scale of selling has also been amplified by automated trading strategies and leveraged exchange-traded funds, which are required to adjust their positions as market volatility increases.

According to a Nomura estimate, leveraged ETFs were expected to sell approximately $47 billion worth of stocks into Thursday’s market close as part of their rebalancing activity.

The anticipated selling, largely concentrated in technology and AI-focused funds, was among the largest recorded by the brokerage since it began tracking the data in 2006. Margin calls and the retreat of retail investors, who had previously been willing to buy into market declines, added further pressure.

However, the more important question is whether investors are simply reducing exposure after an exceptionally strong rally or beginning to question the underlying economics of the AI industry.

The distinction matters because a temporary correction may leave the broader investment cycle intact, while a sustained loss of confidence could make it harder for companies to raise the capital required for their expansion plans.

For now, the long-term promise of artificial intelligence remains intact, and many investors continue to believe that the technology will become central to the global economy. What is changing is the willingness to accept enormous spending commitments and extraordinary valuations without clearer evidence of profitability.

The latest selloff has therefore brought the AI industry to an important inflection point. The race to build the infrastructure powering artificial intelligence is far from over, but investors are increasingly asking a question that the sector has largely been able to postpone until now: when will the money pouring into AI begin to deliver returns?