Friday's numbers told the story of the week. The Nasdaq fell 2.1%, taking its weekly loss to 4.5%, and the S&P 500 dropped 1.9% over the week, both their worst since June.
Chip stocks led the way down again. The neoclouds CoreWeave and Nebius each lost almost a quarter of their value over the week and fuel-cell maker Bloom Energy, another darling of the AI power trade, a fifth. Wall Street's fear gauge, the VIX, climbed back above 20.
Over here, the FTSE 100 gained 3.1% to close near the 10,900 mark, its best week in six months.
So, is this the moment the AI momentum trade finally breaks, or just another wobble that ends with investors piling back in?
The important distinction is between the AI boom and the AI trade. AI investment can keep growing, chip demand can stay strong and the technology can transform entire industries, while the shares bought in anticipation of all that still fall. The question is no longer whether AI demand is real. It's whether the returns will justify the cost of building it.
What happened in London
The FTSE's rally was about what it isn't, as much as what it is. The UK's blue-chip index has very little direct AI exposure. It is heavy in pharma, consumer staples, energy, miners and defence, exactly what investors reach for when growth stocks look stretched. Reuters reports that shares in industries such as beverages rose 7.8% over the week and medical equipment 6.9%, while house builders gained more than 6% after Healey reaffirmed his commitment to fiscal discipline.
Defence names like BAE and Babcock kept climbing, with energy also helping. Brent crude closed on Friday near $110 a barrel, up about 5% on the week and roughly 80% this year, which flatters an index full of oil majors. Morningstar data shows UK large-cap ETFs have attracted around $3.4bn of net inflows this year, the most since 2020.
What makes this more interesting is that it happened even as energy prices climbed again and traders priced in further Bank of England rate rises. Investors weren't buying the UK because the macro looks rosy; it looks more like they were buying shelter. That's consistent with money rotating out of AI, though not proof of it: oil, defence and pharma all had their own reasons to rally.
We've been here before
The AI trade has a long record of scaring people and then recovering.
In the summer of 2024, Nvidia fell roughly a quarter from its peak amid a rotation out of big tech and the yen carry-trade unwind. It hit new highs within months. In January 2025, China's DeepSeek wiped almost $600bn off Nvidia in a single day, at the time the largest one-day loss of market value in history.
Within months that was forgotten too. In April 2025, Trump's "Liberation Day" tariffs pushed the Nasdaq into a bear market and chip stocks fell around 40%; by the summer they were back at records high. Each time, the playbook was the same: a scare, a rotation into safety, then a piece of good news, usually blowout results from Nvidia or one of the hyperscalers raising their spending plans again, and the money came rushing back. Of course, those are precisely the dips that recovered, which is why past form proves less than it seems.
The bulls will say this week fits the pattern. TSMC and ASML both said demand is outstripping supply into 2028. Fed chair Kevin Warsh used Friday morning to praise a "massive investment surge" in AI and the next round of potential good news is close: Microsoft, Alphabet, Meta and Amazon report over the next fortnight, Nvidia in late November and rumours suggest Anthropic may look to IPO in November.
Why this time might be different
Several things don't fit the old pattern.
First, this sell-off has lasted longer. The Philadelphia chip index peaked in late June after more than doubling, and is now around 20% below that high. Previous dips were sharp but brief; this one has been grinding on for four months, through a summer slump and now this week's rout.
Second, the worry has changed. In 2024 and 2025 the fear was about demand: would anyone pay for all this? This time the demand looks fine; the problem is financing and cost of capital. CoreWeave's debt trading at distressed levels and record prices for default insurance across AI borrowers are not the kind of problem a good set of results from Nvidia fixes overnight. Neither are persistent inflation and a Fed that may still be raising rates.
Third, investors are becoming more selective. As the FT's Unhedged pointed out on Saturday, the selling hasn't been indiscriminate. Nvidia and TSMC have held up relatively well, while memory-chip makers have been hammered: SK Hynix's US-listed shares fell 17% this week and Micron's 12%, after China's CXMT claimed a breakthrough in cutting-edge memory chips, raising fears that memory prices have peaked. Some perspective is needed, though: even after that fall, Micron's shares are still worth more than four times their low of the past year, which shows just how far memory had run. This looks like a market starting to separate the companies with pricing power and cash from those with competition and debt.
There's a fourth, less visible reason: leverage. Goldman Sachs' prime brokerage data shows hedge funds dumping AI stocks for six straight sessions this week, and an unusual pattern: from Covid to the end of 2025, all three main hedge-fund strategies never fell more than 0.8% on the same day. It has now happened three times this year, most recently on Wednesday. One of the others was 28 July, the week Leopold Aschenbrenner's Situational Awareness, the poster child of AI momentum investing, reportedly had to sell its entire book to Citadel after margin calls. Many of its favourite names, CoreWeave, Nebius and Bloom Energy among them, were hit hard again this week. When crowded trades unwind, forced sellers can push prices well beyond what fundamentals alone would justify. Whether that, or genuinely worsening economics, is driving this sell-off is the key question: the first tends to reverse quickly once the selling stops; the second doesn't.
The counterargument
It would be easy to read too much into one week. The FTSE has had good weeks before without starting a new era and London's rally owes a lot to oil near $110 and defence spending, not just AI nerves. The AI companies at the centre of the trade are still growing at extraordinary rates and the money rotating into defensives can rotate back just as quickly.
The strongest bull case is that today's spending is building infrastructure for a much larger future market, and revenue always lags the capital spent: heavy spending and thin free cash flow early on are not, by themselves, proof that the cycle has failed. Momentum trades rarely end with a clean break; they tend to end when the good news stops working, which hasn’t been tested here yet.
My read
I don't think this is the end of the AI boom, but I think it could the end of the beginning phase of the AI trade. For two years, any AI-related share went up on good news and every dip was bought indiscriminately. What I suspect we're seeing now is the market splitting AI into winners and borrowers: the companies that generate cash from the build-out and the ones borrowing to fund it. With the US 10yr still north of 5%, after a very strong rally this week, the cost of capital may yet rise further. Impacting any business model that is rate sensitive.
The rotation into the FTSE looks to me like investors buying insurance rather than abandoning AI. The real test comes over the next fortnight, and it isn't just whether the hyperscalers spend more. It's whether they can show AI revenue and cash flow growing fast enough to justify it. If spending rises, but the revenue doesn't keep up, the shares could keep falling for weeks, not just a day. That would probably force a rethink as to whether the momentum trade is finally dead.
The Comms Read
For UK companies, this is a rare window. International money is taking a fresh look at London and the FTSE's appeal due to its old world economy nature: cash flow, dividends and low exposure to AI hype.
For UK IR teams my message would be: Carpe Diem! You should be putting your investment case in front of US and European investors, before the rotation reverses.
For the AI companies, the narrative has shifted from "how fast can you grow" to "how do you pay for it". Results over the next fortnight should lead with funding, cash generation and returns on capital, not just the size of the order book.
What to watch
Microsoft, Alphabet, Meta and Amazon results over the next fortnight; US inflation and the Fed's late-October meeting; UK CPI on Wednesday; credit spreads on CoreWeave and Oracle; and whether the FTSE can push through its 52-week high of around 10,990.
Time will tell. GLA & DYOR.
Disclaimer
This newsletter is published by James Lea / CorpCast for general information, educational and journalistic purposes only. Nothing in it constitutes financial, investment, legal or tax advice, or a personal recommendation, and it is not an offer, invitation or inducement to buy, sell or hold any security, cryptocurrency or other financial instrument.
Any views, opinions, forecasts or interpretations expressed are the author's own personal opinions at the time of publication, do not reflect the views of any business or entity the author is affiliated with, and may change without notice. No account is taken of any reader's personal circumstances, objectives or financial situation, and nothing here should be relied upon as the basis for any investment decision.
The author may from time to time hold long or short positions in securities, cryptocurrencies or other assets referenced in this newsletter, and may transact in them at any time without further notice. This is a standing disclosure covering all issues and is not repeated or itemised per asset.
Investing and trading involve substantial risk of loss, including the potential loss of your entire investment. Past performance is not a reliable indicator of future results. Figures, data and third-party information are believed reliable but are not guaranteed to be accurate, complete or up to date.
James Lea / CorpCast is not authorised or regulated by the Financial Conduct Authority. Always conduct your own independent research and consult a suitably qualified, FCA-regulated financial adviser before making any investment decision.