Barclays declares AI boom has 'much further to run' despite brutal stock sell-off
Despite the market bleeding billions, Barclays’ global research chief says AI doom talk is overdone and it will a key driver for years to come.
After a brutal week for stocks, investors will be wondering if the AI boom has peaked.
Australia’s S&P/ASX 200 index is already down 5.2 per cent this month after suffering its worst week since April. The S&P 500 in the US has fallen 3.5 per cent. Both are having their worst November since the global financial crisis of 2008.
But after a rally on Wall Street overnight on Friday, Australian shares are expected to claw back some of their losses, with the futures market indicating a more than 1 per cent bounce. US investors took heart from New York Fed president John Williams’ comments that the central bank could still lower interest rates in the “near term,” despite the latest employment figures fanning expectations that no more reductions were imminent.
AMP chief economist Diana Mousina said investors had been uneasy about the “overvalued” tech sector and signs of softer earnings growth.
“In our view, there is still more upside for tech earnings as AI adoption is still in its early stages,” Ms Mousina said. “The same pace of growth for US tech stocks is unsustainable, but this does not mean that prices have to ‘pop’. The bubble could just deflate a bit.”
Similarly, Ajay Rajadhyaksha, Barclays’ global chairman of research, says doom talk about the collapse of AI is “overdone” and the boom has further to run.
“We expect AI to be the most important macro factor in 2026, as traditional drivers such as monetary policy and trade policy fade,” Rajadhyaksha says in his latest global outlook.
Investors are certainly in need of some reassurance after a torrid few weeks. The ASX 200 has fallen for four straight weeks and is down 7.5 per cent from last month’s record high of 9115.2. It’s the worst month for Australian stocks since September 2022.
“Clear question marks (are) circling now around the extent of any Santa rally,” says IG market analyst Tony Sycamore.
November is usually one of the best months for stocks.
A 1.6 per cent fall in the ASX 200 on Friday came after US shares fell overnight on Thursday despite stronger than expected quarterly results and guidance from Nvidia, as well as upbeat comments from its chief executive, Jensen Huang. Nvidia fell 3.2 per cent after initially rising 5.1 per cent. The reversal sparked a broader rout in stocks that sent the Nasdaq down 2.2 per cent. It comes amid the biggest fall in Bitcoin since 2022. Amid fading hope of US rate cuts, the cryptocurrency has fallen as much as 36 per cent from a record high last month.
Stephen Innes from SPI Asset Management says the sell-off in stocks reveals deep concerns about whether tech companies can justify their huge AI spending. Nvidia’s accounts receivable rose about 45 per cent to $US33.39bn in the quarter.
Goldman Sachs traders said Nvidia’s results and outlook were “positive” but when shares fall after good results it’s “typically a bad sign”. Oracle’s credit default swaps had been blown out and were a “barometer for AI risk and draws sceptics looking for a hedge,” they added.
The big cloud providers have issued $US121bn in debt this year to fund the build-out, and their capital spending is expected to hit about $US500bn next year.
“Investors aren’t questioning chip demand – it remains explosive. They’re questioning whether the buyers can justify the capital outlay, monetise the build-out, and finance the next leg without the credit markets blinking,” said SPI Asset’s Innes.
But Barclays’ Rajadhyaksha dismisses comparisons to past tech bubbles. The largest seven companies in the dotcom bubble traded at valuations twice as high as today’s “Magnificent 7” tech giants. And unlike the 1990s giants, today’s leaders boast much wider profit margins.
“Adjusted for margins and earnings growth, US equities do not appear in a bubble at this point,” he argues. More importantly, he said demand for AI computing power and inference will “clearly outstrip supply” for at least the next several quarters.
“Perhaps there will come a point when the trillions of dollars of AI capex will be seen as folly. We don’t know if that will happen; no one does,” he says.
“What we do feel comfortable calling is that this is not a 2026 event; if there is a demand-supply mismatch, it is in the direction of not enough supply.”
The numbers backing the AI boom are staggering. Nvidia’s Jensen Huang expects between $US3 trillion and $US4 trillion of AI infrastructure spending by 2030. That translates to between 70 and 95 gigawatts of power demand – enough to power millions of homes.
The big cloud providers are forecast to spend about $US1.2 trillion over the next two years.
The Stargate venture backed by OpenAI and SoftBank is on track for 10 gigawatts and $US500bn of commitments by year’s end. Rajadhyaksha says it explains why chip demand remains so strong.
“AI infrastructure demand still exceeds supply, keeping chipmakers and component suppliers such as Nvidia in a favourable position,” he says.
He expects US tech to retain market leadership in 2026, powered by sustained AI-driven spending and resilient double-digit growth in cloud and digital advertising. “With trade and fiscal headlines fading in importance, AI capex will now be the main factor driving US GDP growth,” he adds.
The AI revolution is also creating opportunities beyond tech stocks. Rajadhyaksha notes that the massive demand for metals and critical minerals needed for AI infrastructure is triggering the early stages of a commodity investment cycle.
Chile, Peru and Australia are likely to be key beneficiaries.
Energy constraints pose the biggest risk to AI ambitions, he warns.
For sharemarket investors nursing losses through November’s rout, Barclays says don’t bet against AI.
To join the conversation, please log in. Don't have an account? Register
Join the conversation, you are commenting as Logout