Apple’s quiet AI bet gains traction

Apple is turning restraint into a strategic advantage as rivals pour record sums into artificial intelligence infrastructure, betting that a narrower, device-centred approach can protect margins while keeping investors focused on earnings rather than data-centre bills.

The iPhone maker’s latest quarterly performance has strengthened that argument. Apple reported March-quarter revenue of $111.2 billion, up 17 per cent from a year earlier, with diluted earnings per share rising 22 per cent to $2.01. Net income reached $29.6 billion, while operating cash flow exceeded $28 billion. Services revenue hit $31 billion, a record for the segment, and iPhone sales rose to nearly $57 billion, underlining the continued power of Apple’s installed base.

The figures landed at a moment when Wall Street is testing the limits of patience with the wider AI boom. Microsoft, Alphabet, Amazon and Meta are committing hundreds of billions of dollars to servers, chips, data centres and power capacity as they compete to supply cloud computing and build larger models. Meta has lifted its 2026 capital expenditure outlook to as much as $145 billion. Alphabet has raised its own spending forecast to between $180 billion and $190 billion. Amazon has pointed to a capital spending programme centred on AI and cloud capacity that could reach $200 billion this year.

Apple has not joined that race on the same terms. Its strategy relies on a combination of on-device processing, selective cloud inference, custom silicon and tight integration across iPhone, iPad and Mac. The company’s Private Cloud Compute architecture is designed to handle more complex AI requests on Apple silicon in data centres while limiting data retention and privileged access. That allows Apple to claim a privacy advantage while avoiding the full financial exposure faced by hyperscale cloud operators.

This does not mean Apple is underinvesting. Research and development expenses rose to $11.4 billion in the March quarter from $8.6 billion a year earlier, taking the six-month total to $22.3 billion. Annual R&D spending stood at $34.6 billion in fiscal 2025. The difference is that Apple’s spending is directed less towards selling raw AI compute to third parties and more towards making its devices and services more valuable to existing users.

That distinction matters for investors. Cloud groups must build capacity before demand is fully visible, creating pressure to prove that AI workloads can deliver returns fast enough to justify depreciation, energy costs and debt financing. Apple’s model is more incremental. It can fold AI features into devices already generating high-margin hardware and services revenue, while using software improvements to defend upgrade cycles and reduce churn.

The risks remain significant. Apple has faced criticism for moving slowly in generative AI compared with Microsoft-backed OpenAI, Google’s Gemini ecosystem and Meta’s open-model strategy. Its promised evolution of Siri has been watched closely because voice assistance is one of the clearest consumer use cases for AI. Any delay or underwhelming release could reinforce the perception that Apple is strong at packaging technology but weaker at frontier AI development.

Competitive pressure is also intensifying from device makers that are promoting AI phones and personal computers as a new upgrade category. Samsung, Google and several chipmakers are pushing features such as real-time translation, image generation, contextual search and agent-style assistants. Apple must show that its privacy-led approach can match the usefulness of more cloud-heavy systems without frustrating users who expect faster and more flexible AI tools.

Even so, Apple’s financial position gives it room to move. Its services business produced $109.2 billion in revenue in fiscal 2025, with gross margin above 75 per cent. The company’s active device base has reached new highs across major product categories and regions. Its board has authorised a further $100 billion in share repurchases, signalling confidence that cash generation remains strong despite rising technology costs.