Swissquote: Fragile nerves ahead of tech earnings
Swissquote: Fragile nerves ahead of tech earnings
By Ipek Ozkardeskaya, Senior Analyst, Swissquote
Last week was a good one to be away, as things got worryingly ugly on the economic data, earnings and AI fronts.
Mixed bank earnings were overshadowed by a hotter-than-expected CPI report that fuelled hawkish Federal Reserve (Fed) expectations, pushing October rate hike bets to roughly 50% before Warsh’s words pulled them backed below 13%, phew...), while US crude rose above $100 per barrel despite the controversial US decision to ease sanctions on Russian diesel. Meanwhile, a sharp selloff hit Nvidia-backed neocloud provider CoreWeave’s shares and bonds.
Funnily — or disquietingly — the thunder struck as co-founder Brannin McBee discussed securing investment-grade ratings to access what he described as insurers’ 'unlimited' capital. That’s not necessarily what investors want to hear from a company pouring cash into infrastructure — and reporting negative free cash flow in the latest quarter.
These famous last words certainly didn’t help: the cost of insuring CoreWeave’s debt against default surged to nearly 1’200 basis points, with pricing implying a more than 60% probability of default within five years, under the model’s assumptions. Rough.
The stock price tanked nearly 24% last week. And more importantly, the nerves spread well beyond CoreWeave, rattling other neocloud providers and the broader AI trade. The Philadelphia Semiconductor Index dropped more than 10% over the week.
Why?
First, CoreWeave may sound like a little drop in the ocean, but its ties to companies including Nvidia and Microsoft make it part of a much bigger financial ecosystem. Higher financing costs raise the hurdle for future projects and refinancing. Existing fixed-rate debt doesn’t suddenly become more expensive, but the next round of borrowing certainly could. And given the massive buildout and deeply negative free cash flow, external financing will remain essential to CoreWeave’s (and similar businesses’ expansion).
In this configuration, if the economics stop adding up — if returns can’t sufficiently exceed financing costs — spending could slow, and the consequences could reach suppliers, customers, shareholders and lenders. Yes, that’s where circular financing could go wrong: one company’s spending supports another’s revenue, which helps justify another round of investment. If financing becomes harder to secure, that support could start travelling in reverse.
Second, there is a broader concern about timing. Big Tech is spending heavily on infrastructure today, while the revenue generated by businesses adopting AI take longer to catch up. That gap needs financing. And as we saw last week, investors are becoming less enthusiastic about funding the wait — and funding endlessly...
As such, tech is coming under growing pressure while investors seek shelter. US Treasuries offered that shelter last week, pulling the US 10-year yield towards 5% from nearly 5.40% earlier this month. Escalating tensions in France and broader risks surrounding euro-area bonds could also support demand for longer-term US Treasuries, putting downward pressure on their yields. Meanwhile, gold is finding support around $4’000 per ounce as long-term yields ease.
This flight to safety could continue, alongside further rotation towards non-technology pockets of the market. Indeed, the S&P 500’s equal-weighted index performed better than its market-cap-weighted, technology-heavy counterpart. The Nasdaq 100 was hit harder than its major US peers, while the Dow Jones Industrial Average ended a chaotic week with a 0.6% gain and Stoxx 600 rose more than 2% (with Cac 40 up by similar amount, well in contrast with what’s going on in the streets...)
But rotating away from tech may not solve the broader problems.
- Technology has been a major engine of earnings growth and investment, helping sustain market optimism despite higher energy prices and borrowing costs. If that engine slows, the rest of the market may struggle to compensate. On the contrary, a financing slowdown could delay projects, weaken suppliers’ order books and cool investment and hiring. On top, falling share prices could dent household wealth and spending, while losses among lenders could make credit harder to obtain.
- Persistently above-target inflation and the prospect of further interest rate hikes would further dent growth. Without the precious support of AI spending, economic growth would look much less shiny. Remember, a broad measure of technology investment accounted for around half of year-on-year US GDP growth in Q2 2026. That isn’t all AI, but it shows how important the technology investment boom has become.
Now, note that none of this requires AI demand to disappear: strong results from TSMC and ASML last week confirmed that AI-related demand remains robust, with supply expected to lag behind demand for years. But returns on AI investment simply arriving later than expected could be enough to cause trouble when financing gets expensive.
Anyway, having said all this, the week starts in a slightly better mood. Chinese growth numbers came in better than expected, though growth remains heavily supported by exports rather than domestic consumption, at a time when trade tensions with the EU are rising too...
South Korea’s Kospi, which dropped more than 14% during last week’s tech turmoil, is up around 0.85% at the time of writing, hinting that the selloff may be easing. But SoftBank, a bellwether of the AI trade, kicks off the week down more than 3.5%, suggesting that the dark clouds haven’t cleared just yet.
This week, tech earnings will put that fragile mood to the test. Meta takes centre stage, alongside Tesla, IBM, Western Digital and Texas Instruments. Investors will want to hear more than another ambitious spending plan: they will want evidence that AI is helping generate enough revenue and profit to justify the bill.
As for spending, it remains a delicate balance. On one hand, strong AI spending plans reassure investors that demand for AI infrastructure remains intact and sustain appetite for AI enablers, starting with chipmakers and data centre providers. But too much spending fuels concerns about financing — including leverage and circular financing — and returns.
On the other hand, more cautious spending could ease those financing and return concerns, while pulling the rug from under the companies counting on that money. So what investors really need is stronger monetisation — proof that the revenues generated with AI can catch up with the spending on AI tools. Because after last week, another promise to spend billions may not sit well in investors’ stomachs...