Please see the below article from Tatton Investment Management discussing market sentiment, AI adoption and China’s economic outlook, received this morning – 26/10/2026.
The sentiment gap
Global stocks sold off last Thursday, as AI optimism waned. Despite strong earnings, Meta’s investment plan ramp up renewed fears about unsustainable debt-fuelled growth. The timing is strange, though; underneath the noise, equity fundamentals are still improving.
The media’s AI bubble-burst narrative feels to have reached fever pitch. The question is whether that’s more about commentators catching up with market stories than anything new? Tech firms’ credit default swaps (CDS) have spiked in recent weeks, with the highly leveraged datacentre firm CoreWeave at the centre.
CDS evokes memories of the 2008/2009 Global Financial Crisis, but we’re a long way from the days of interwoven and opaque bank debts. Barclays’ Ajay Rajadhyaksha points out that even for CoreWeave itself, the distressed unsecured bonds are only a small part of its funding pile. And CoreWeave is only a small part of the AI investment boom. If the company went bust, it could knock some companies with stakes in it (Nvidia and OpenAI) but there would be no immediate domino effect.
Our usual rebuke of the AI bubble narrative is that equity valuations have fallen this year (as share prices have outpaced earnings) so it’s not a classic bubble. Now, bond yields have fallen, credit spreads (in the last week) have stayed stable and share prices have dropped. So, even on a relative basis, stocks have cheapened.
Most yields have fallen anyway; French yields broke above 5% on fiscal anxiety, and gilt traders seem to be waiting for a tough UK autumn budget to be over with, before making any moves.
Falling yields and ever-stronger earnings (Q3 is on course to beat Q2) should support shares, but investors don’t feel supported. Perhaps the scare stories – like the PHP website code vulnerability or oil above $110pb – are sapping sentiment. They heighten the sense that double-digit earnings growth can’t last. But as long as earnings keep pace – and especially if yields keep falling – we should expect sentiment to rebound eventually.
Who’s using AI?
The AI sell-off earlier this month was primarily about perceived weakness in demand growth for AI computing power. The metrics for this are opaque, so it’s worth looking at a broader picture of who’s actually using AI.
According to GenAI Adoption Tracker, 60.8% of US working-age adults used AI tools in August, up from 44.6% a year earlier but slightly down from May’s 61.8%. Reported work usage fell even more: 40.5% in August from May’s 45.2%. Short-term data is inevitably noisy, but it’s interesting how closely the adoption rate follows the so-called S-curve: a quick increase, followed by plateau or decline.
Meanwhile, 19.8% of US businesses reported using AI in May 2026, while 20% of EU companies did so in late 2025. The crude survey figures don’t distinguish between the intensities of AI usage they suggest that adoption hasn’t been as rapid as fervent optimists predicted.
What of slower demand for AI ‘compute’? The story that knocked AI stocks was a drop in the LLM Token Expenditure Index, but industry insiders pointed out that individual tokens getting cheaper doesn’t mean people are using less. The latest models use vastly more tokens, and Ramp reported that aggregate business AI spending rose 60% in the nine months to July. Still, demand looks concentrated: just 1% of firms drive 80% of OpenAI and Anthropic’s demand.
There’s some evidence that businesses are switching from the latest token-intensive models to older but sufficient models. Perhaps that’s why AI leaders are advocating a government-enforced slowdown: they cannot unilaterally slow investment for fear of falling behind, but there’s a feeling that existing models already meet many businesses’ AI needs.
Such an interpretation might be cynical but, with so much opaque data, it’s no surprise narratives are moving AI stocks more than cold hard facts.
China’s Japanification is a monetary problem
Is China undergoing ‘Japanification’? The world’s second-largest economy had decades of remarkable growth, but a busted property market and aging population have created deflation. That description of 2020s China applies word-for-word to 1990s Japan.
The comparison might sound strange given China reported 4.5% year-on-year growth in Q3 2026. But, even if you take China’s GDP numbers at face value, the economy is heavily concentrated on industrial production. Persistently weak retail sales data shows the dearth of consumer demand and price deflation hurts company profits. Beijing keeps announcing stimulus measures but, two years on from the initial stimulus pivot and with little to show for it, these measures no longer excite investors.
The struggles of Chinese companies are reflected in share prices. Price wars have kept producer prices low and profits weak. Profits are actually growing by more than stock market falls would suggest – but investors doubt the outlook. Given the government’s history of interventions (such as a recent tax clawback), that’s reasonable. There’s two layers to the problem: Does Xi Jinping’s government actually want to boost company profits and, even if it does, can Beijing’s production-led centralised economic policy model really stimulate demand?
Japan-style ‘lost decades’ aren’t inevitable but, just like Japan in the 1990s, China’s government is either unwilling or unable to boost short-term consumption and investment. The biggest spur would be getting people to borrow again, but the PBoC’s tight monetary policy disincentivises debt-led investment. Chinese bond yields are low compared to the rest of the world, but are still so much higher than expected economic returns that there’s little incentive to invest anywhere other than bonds. If the PBoC injected more liquidity, that could change. Japan learnt the same lesson much too late, when ‘Abenomics’ finally changed the monetary and fiscal landscape. There’s been suggestions that the PBoC might loosen, but we won’t hold our breath.
Please continue to check our blog content for the latest advice and planning issues from leading investment management firms.
Marcus Blenkinsop
26th October 2026