Apple adds to AI woes
For nearly two years, it has been an article of faith that the artificial intelligence (AI) revolution would fund its own momentum, and any enterprise failing to allocate exorbitant amounts of capital risks irrelevance.
More recently, however, that narrative has frayed under the weight of escalating capital expenditure (capex), declining free cash flows, and growing investor fatigue, if not nervousness, about the circular financing, which has escalated again with chip makers Nvidia, AMD, and Broadcom recently acting as US$5 billion cornerstone investors in Meta Platforms’ US$25 billion stock sale.
Because Meta didn’t strictly need supplier capital to raise funds, the deals suggest a potential shift toward “pay-to-play.” This is where suppliers buy equity to secure sales from the hyperscalers, and it now seems to reflect growing competition among chip makers.
Of course, it also compounds stock volatility: by investing directly in key customers, chip suppliers double down on their AI risks, leaving them even more vulnerable to any shift in sentiment toward tech and AI.
We’ve known for some time that the AI ecosystem is heavily interconnected through complex vendor financing, where tech giants and chip makers use debt backstops, loans, and compute guarantees to support weaker data-centre operators and startups. The whole thing is reminiscent of the telecom bubble, which burst in 2001. Back then, round-tripped capital created the illusion of organic demand but exposed participants to systemic risk.
If demand for AI compute falters today, or subprime borrowers default, suppliers will simultaneously face collapsing chip sales, unpaid debt, and depreciated second-hand hardware. The interconnected bets will lead to severe losses for investors.
And so, now, the broader technology market has entered a period of scrutiny. Investors who previously cheered on aggressive infrastructure buildouts are now questioning when – and if – the return on invested capital will arrive.
The background
Hyperscalers have accelerated their capital spending to historic levels. Combined capital expenditure guidance across the largest infrastructure providers – including Amazon, Alphabet, Microsoft, Meta, and Oracle – has surged past US$700 billion for the year.
Amazon raised its annual cash capex projections to approximately US$220 billion, driven heavily by data centre infrastructure and rising memory component costs.
Alphabet increased its capital allocation to between US$195 billion and US$205 billion, even as its trailing free cash flow dipped into negative territory.
Oracle also announced aggressive capital targets of up to US$95 billion gross for the upcoming fiscal year.
This spending spree has begun to weigh on financial metrics. Aggregate debt issuance across the sector has risen sharply to finance the gap between operating cash inflows and massive data centre commitments.
While cloud revenues have expanded, external estimates from analysts suggest that incremental direct revenues tied to new AI services remain significantly lower than the total annualised cost required to build and maintain the hardware stack.
As a result, we’ve seen stock valuations pull back.
Enter Apple
Despite the turmoil, investors have routinely treated Apple as a safe harbour. While competitors poured hundreds of billions into data centres, Apple put its massive cash balance to work through share buybacks and dividends, staying capital-efficient while avoiding direct infrastructure risk. The result: a share price up 29 per cent year to date.
But overnight, following its September quarter earnings report, Apple’s status as a defensive tech investment received a reality check.
While top-line figures looked solid – revenue grew 12 per cent year over year (YoY), and gross profit margin reached 47.9 per cent, both slightly beating consensus Wall Street expectations – the company projected that its high-margin Services division will continue to slow, and gross profit margins will face ongoing pressure.
And then there are rising input expenses, as escalating memory component costs are beginning to strain margins while Apple’s legacy stockpiles of lower-cost inventory dwindle.
Apple shares fell as much as 4 per cent in extended trading market.
The news places significant focus on Apple’s new Chief Executive Officer, John Ternus, who inherits a business valued at a ‘premium’ price-to-earnings (P/E) multiple of 36 times next year’s projected earnings.
That’s a steep multiple for earnings projected to grow by only 1 per cent.
It also worth remembering global court rulings and regulatory mandates have increasingly capped App Store distribution fees, while OpenAI – historically a major generator of App Store fees via the ChatGPT app – has started redirecting prospective subscribers to complete transactions on its direct site, bypassing Apple’s standard commission cuts.
Pricing power fading
Apple’s traditional defence against rising supply costs has been its pricing power. During a global cost-of-living crisis, however, that strategy is meeting resistance. While the US$1,999 foldable iPhone Duo has seen high initial consumer demand, Apple’s overall AI offering is facing criticism.
The “updated” Siri AI chatbot continues to be limited to basic device controls, while competing autonomous assistants such as Instinct and Muse, can carry out complex multi-step tasks across application ecosystems.
At a minimum, the lack of a compelling new artificial intelligence capability will see consumers extend their hardware upgrade cycle.
With its core growth driver – Services – decelerating, memory costs increasing, and its primary AI software features lagging behind market leaders, Apple’s slowing momentum is worthy of note, and it’s the icing on the cake for a tech market that’s recalibrating its expectations around the economics of AI.