The week Wall Street finally asked what it's paying for

TL;DR

• US markets closed lower on Friday, June 26, as a sharp selloff in AI-linked semiconductor stocks dragged the S&P 500 down 0.6% and the Nasdaq Composite dropped as much as 1%, putting the benchmark on course for only its second weekly decline in 13 weeks [4][5][8].

• A reported delay in OpenAI's IPO to 2027 served as the immediate catalyst, rattling chip stocks and triggering a broader rotation out of technology and into defensive sectors like healthcare and consumer staples [3][4][9].

• US equity funds bled US$8.5 billion (about A$13 billion) — the first weekly outflow since March — with technology funds alone shedding a record US$9.3 billion (about A$14.2 billion) [1][2].

• South Korea's KOSPI was caught in the crossfire, with five circuit breakers triggered in June amid volatility in Samsung and SK Hynix [12].

• Healthcare names including AbbVie, Eli Lilly, and Johnson & Johnson hit all-time highs as investors ditched tech, and Goldman Sachs floated the idea that big tech hyperscalers could become more attractive amid chip volatility [1][6].

What happened

Friday, June 26 opened with all three major US indexes in the red after investors continued selling technology and AI-related stocks, a pattern that had been building through what multiple outlets described as one of the worst weeks for tech stocks of the year [1][2][8]. The selling was not uniform across the market — defensive sectors, including healthcare and consumer staples, actually pushed the main indexes into positive territory by lunchtime — but momentum faded into the close, and the major averages ended slightly lower [1][4][5].

The tech-heavy Nasdaq was the biggest loser among the three major indexes, falling as much as 1% during the session [4][8]. The S&P 500 declined 0.6%, heading toward only its second weekly decline in the last 13 weeks — a striking statistic given how relentlessly the index had climbed through the first half of the year [4][5]. The Dow, while also slipping, was cushioned by the rotation into defensive names [6][7].

The proximate trigger was a report that OpenAI had delayed its initial public offering to 2027 [3][4][9]. That single piece of news sent shockwaves through the semiconductor complex, which has been the most direct beneficiary of the AI spending boom. Chip stocks — the picks-and-shovels plays of the AI gold rush — extended losses throughout the session [1][2]. SanDisk was among the notable individual decliners, tumbling as tech weakness weighed on the broader market [6][7].

Beneath the surface, the fund-flow data told an even starker story. US equity funds recorded US$8.5 billion (about A$13 billion) in outflows — the first weekly redemption since March [1][2]. Technology funds alone accounted for US$9.3 billion (about A$14.2 billion) in outflows, a record [1][2]. The timing, as one report noted, was hard to ignore: the hose of capital that had been flooding into AI-adjacent equities just kinked [1].

The selloff was not confined to US shores. South Korea's KOSPI crashed hard enough to trigger five circuit breakers in June, driven largely by volatility in Samsung and SK Hynix — two of the world's most important memory-chip manufacturers and key links in the global AI supply chain [12]. SoftBank, whose portfolio is heavily exposed to AI bets, was also caught in the downdraft [9].

What it actually means

The real story here is not a single day's price action. It is the collision of two narratives that had been running on parallel tracks: the first, that AI spending would compound indefinitely; the second, that the market could keep paying any multiple for anything touching that theme. Friday was the day those tracks converged — and the market did not like what it saw.

Multiple sources framed the week as the moment Wall Street finally confronted a question it had been too euphoric to ask: what exactly are we getting for all this AI spending? [8] That is not a trivial question. Capital expenditure at the hyperscalers has run into the hundreds of billions of dollars globally, and the revenue models justifying that spend remain, at best, provisional. When OpenAI — the company whose valuation has become a proxy for the entire generative-AI thesis — reportedly pushed its IPO to 2027, it forced investors to price in a longer wait for liquidity and validation [3][4][9]. The chip stocks, which had been bid up as the purest expression of AI demand, were the most exposed to that repricing.

Yet the framing matters. Several sources quoted experts describing the move as a "healthy correction" driven by profit-taking [1][2]. That language is deliberate. A healthy correction implies the underlying thesis remains intact and that excess speculative premium is being shaved off. An unhealthy correction implies the thesis itself is breaking. The evidence so far — rotation into healthcare and consumer staples rather than a flight to cash or bonds — is more consistent with the former [1][4][5]. Investors are not abandoning equities; they are repositioning within them.

The fund-flow data reinforces this read. Yes, US$8.5 billion (about A$13 billion) left equity funds, and yes, the US$9.3 billion (about A$14.2 billion) tech outflow was a record [1][2]. But the broader market held up. Defensive sectors rose. AbbVie, Eli Lilly, and Johnson & Johnson hit all-time highs [1][6]. That is not the behaviour of a market in panic. It is the behaviour of a market reallocating — trimming the winners, adding to the unloved, and waiting for the next data point.

Goldman Sachs's suggestion that big tech hyperscalers could become more attractive amid chip volatility is the most interesting forward-looking signal in the bundle [1]. The logic is straightforward: if chip prices fall, the companies buying those chips — the hyperscalers building data centres — see their input costs decline. The suppliers take the margin hit; the buyers get a break. That is a sector-level rotation within tech itself, not a rejection of the AI thesis.

Hype deconstruction

Let's be clear about what this is not.

This is not the dot-com crash replayed. The companies at the centre of the AI build-out — the hyperscalers, the foundries, the memory-chip makers — are generating real, substantial revenue. They are not Pets.com. The selloff was a single-week correction that put the S&P 500 on course for only its second weekly decline in 13 weeks [4][5]. That is a pullback within a powerful uptrend, not a trend reversal.

This is also not a referendum on AI itself. The OpenAI IPO delay to 2027 is a liquidity and timing event, not a technological failure [3][9]. Generative AI models continue to improve. Enterprise adoption continues, however unevenly. What changed on Friday was the market's estimate of when it would get paid for that adoption — not whether it would get paid at all.

The "record" tech outflow of US$9.3 billion (about A$14.2 billion) sounds alarming in isolation, but it needs context [1][2]. Record outflows are almost a statistical inevitability in a sector that has seen record inflows for months. The absolute number is less informative than the direction — and the direction was out of tech, into defensives, not out of equities entirely.

Finally, this is not a signal to sell everything AI-related. The KOSPI's five circuit breakers in June are dramatic [12], but they reflect a market structure — Korea's heavy concentration in memory-chip stocks — that amplifies volatility rather than a fundamental collapse in semiconductor demand. Samsung and SK Hynix are not going to zero. They are, however, being repriced.

Stakeholder landscape

Chip and semiconductor companies are the most immediate losers. Their stocks absorbed the brunt of the selling because they had captured the brunt of the euphoria. Any company whose valuation is predicated on infinite AI demand growth is now facing a market that wants to see evidence before extending the multiple further [1][2][8].

Hyperscalers — the big tech companies building AI infrastructure — occupy an ambiguous position. Goldman Sachs's note suggests they could benefit from cheaper chips [1], but they are also the ones spending the capital that the market is now questioning. If the scrutiny intensifies, their capex commitments become a liability, not a strength.

Defensive-sector investors are the clear winners this week. Healthcare names hitting all-time highs is not a coincidence — it is a deliberate signal that institutional money is seeking earnings visibility and dividend reliability over speculative growth [1][6]. Consumer staples played the same role [1][4][5].

Retail investors who piled into AI-themed ETFs and single-name chip stocks at peak euphoria are the most exposed. They arrived late, paid top dollar, and are now sitting on the losses. The "healthy correction" framing from experts [1][2] is cold comfort to anyone who bought at the top.

Korean and Asian markets are collateral damage. The KOSPI's circuit-breaker triggers [12] show how concentrated, export-driven markets amplify US-led selloffs in specific sectors. SoftBank's exposure [9] makes it a barometer for the entire AI venture-capital complex — when OpenAI's timeline slips, SoftBank's portfolio mark-to-market gets hit.

OpenAI itself is the elephant in the room. The IPO delay to 2027 [3][4][9] removes a key liquidity event from the near-term calendar and forces private-market investors to hold longer. Whether that delay reflects strategic patience or unresolved structural questions about the company's governance and path to profitability is the open question.

Cross-layer implications

The most non-obvious connection here is between chip-stock volatility and the hyperscaler capex cycle — and it runs in the opposite direction to what the headlines suggest.

When chip stocks fall, the immediate read is AI demand is weakening. But Goldman Sachs's observation that hyperscalers could become more attractive [1] points to a second-order effect: cheaper input costs for the buyers. If semiconductor prices ease — even marginally — the economics of building AI data centres improve for the likes of Microsoft, Amazon, Google, and Meta. The market may be setting up a rotation from suppliers to buyers within the tech complex, even as it rotates from tech to defensives at the sector level.

That has implications beyond equities. If hyperscaler capex holds steady while chip prices fall, the AI infrastructure build-out continues — it just becomes less profitable for the suppliers and more profitable for the deployers. That could reshape the competitive landscape over the next 12–18 months, concentrating power among the handful of companies that can afford to buy chips at scale.

There is also a geopolitical layer. The KOSPI's volatility [12] is a reminder that the AI supply chain runs through countries — Korea, Taiwan, the Netherlands — whose markets are far more concentrated than the US's. A chip-stock selloff in New York translates into systemic stress in Seoul. Policy-makers in those jurisdictions are watching not just prices but circuit-breaker counts as a gauge of financial stability.

What this means for you

If you hold AI-themed ETFs or individual chip stocks, do not panic — but do pay attention. A single-week correction is not a reason to sell at a loss. It is a reason to reassess your position sizing. If your portfolio's performance depends entirely on chip stocks continuing to rise at their recent pace, you are taking more risk than you may realise.

If you have been waiting for an entry point into quality tech names, this may be the beginning of one, not the end. The hyperscaler thesis — that the biggest tech companies will benefit from AI regardless of which chipmaker wins — is arguably strengthening as chip prices come under pressure [1].

If you are invested in defensive sectors — healthcare, consumer staples — enjoy the rally but do not overcommit. The rotation into these sectors is a tactical move, not a structural one. When the AI narrative regains its footing, money will flow back out of defensives just as quickly as it flowed in.

For Australian investors, the direct exposure is limited — the ASX is not a major venue for AI-chip stocks — but the indirect exposure is real. Your superannuation fund almost certainly holds US tech equities, and the Australian market tends to track Wall Street's risk appetite. A sustained tech correction would weigh on local sentiment even if the fundamental impact on Australian companies is minimal.

Uncertainty ledger

• The OpenAI IPO delay's rationale is unresolved. Is it strategic patience — waiting for better market conditions or higher valuations — or does it reflect governance, regulatory, or profitability concerns? The answer would dramatically change the read on the entire AI sector [3][4][9].

• The durability of the rotation is unknown. A one-week move into defensives could reverse quickly if next week's data — US jobs, Fed policy signals [10] — reinvigorates risk appetite. The S&P 500's 13-week run of near-uninterrupted gains suggests the underlying bid is strong.

• Whether this is a correction or a turning point depends on earnings. The next round of hyperscaler and chipmaker results will be the real test. If capex guidance holds and revenue beats expectations, the selloff will look like a blip. If guidance is cut, the "healthy correction" framing collapses.

• The KOSPI's circuit-breaker triggers [12] may reflect local market-structure issues as much as global AI sentiment. Disentangling Korea-specific factors from the global chip cycle would refine the analysis.

• Goldman Sachs's hyperscaler call [1] is a single-source suggestion, not a consensus view. It is plausible and logically sound, but it has not been corroborated by other analysts in this bundle.

Bottom line

This is the market's first genuine stress test of the AI thesis, and it passed — barely. The selloff was real, the outflows were record-setting, and the rotation into defensives was unmistakable, but the broader market held and the selling looked like profit-taking with a purpose, not capitulation. The question Wall Street finally asked — what are we getting for all this spending? — does not have an answer yet, and until it does, chip stocks will remain volatile and defensives will remain bid. The AI trade is not dead; it is just being asked to grow up.

Sources

1. THE LOCAL REPORT ARTICLES. (26 June 2026). Stock market close: Chip stocks extend losses, S&P 500 and Nasdaq end lower - THE LOCAL REPORT ARTICLES.
2. Durva More. (26 June 2026). Stock Market Close: S&P 500, Nasdaq end lower as chip stocks extend losses. Hindustan Times.
3. Durva More. (26 June 2026). US stock market opens lower as S&P 500 holds steady, Nasdaq slips on OpenAI IPO delay. Hindustan Times.
4. Amalya Dubrovsky. (26 June 2026). Stock market today: Dow, S&P 500, Nasdaq slip as report of OpenAI IPO delay rattles chip stocks. Yahoo! Finance.
5. Amalya Dubrovsky. (26 June 2026). Stock market today: Dow, S&P 500, Nasdaq waver as report of OpenAI IPO delay rattles chip stocks. Yahoo! Finance.
6. Emma Newbery. (26 June 2026). Stock Market Today, June 26: Sandisk Tumbles and Tech Weakness Weighs on Markets. The Motley Fool.
7. Motley Fool, Emma Newbery. (26 June 2026). Stock Market Today, June 26: Sandisk Tumbles and Tech Weakness Weighs on Markets. Yahoo! Finance.
8. The Times of India. (26 June 2026). US stock market today: AI stocks sink again, pulling Wall Street towards weekly loss.
9. A Ksheerasagar. (26 June 2026). Can an OpenAI IPO delay derail the AI chip rally? Here's why it matters | Stock Market News. mint.
10. Jeff Marks. (26 June 2026). The Club's top 10 things to watch in the stock market Friday. CNBC.
11. Colin Laidley. (26 June 2026). Stock Market Today: Nasdaq Futures Fall as Tech Stock Sell-Off Continues. Investopedia.
12. Chosun Daily. (26 June 2026). New York Stock Market Ends Lower on AI Semiconductor Slump. 조선일보.