Apple Was Supposed to Lose the AI Race. So Why Is It Worth $5 Trillion?
Studio Orris · · 3 min read
By every measure Silicon Valley uses to keep score in the AI era, Apple should be losing. It has spent a fraction of what its rivals have, its most important AI product arrived years late and partly on someone else's technology, and it paid to settle claims over how that product was advertised. Yet in October 2026 Apple closed above a $5 trillion market value for the first time. The contradiction is not a quirk of the share market. It is a lesson about what customers actually buy, and a warning for any brand tempted to lead with "AI" because everyone else is.
The case in numbers
- ~US$650–700b : reported combined 2026 AI and infrastructure spending of Amazon, Alphabet, Meta and Microsoft
- ~US$14b : Apple's planned 2026 capital expenditure, roughly fifty times less
- June 2024 → September 2026 : from the promise of a smarter Siri to its release, as an opt-in beta with a waitlist and daily limits
- ~US$1b a year : what Apple reportedly pays Google for Gemini-based models (reported, not confirmed by Apple)
- US$250m : settlement over claims about how Siri's AI features were advertised; Apple denied wrongdoing
- US$5.01t : Apple's market value at the close on 14 October 2026
Sources are listed at the end of this article. Spending figures are company guidance and analyst estimates reported in early 2026.
### The arms race Apple refused to run
Over the past two years, artificial intelligence has turned into a capital contest. The largest technology companies have committed hundreds of billions of dollars to data centres, chips and model training, on the logic that whoever owns the most compute will own the future. Reported 2026 guidance puts Amazon near $200 billion, Alphabet between $175 and $185 billion, and Meta between $115 and $135 billion, with Microsoft estimated at well over $100 billion.
Apple has not played that game. Its planned capital expenditure for 2026 is around $14 billion, up only modestly from $12.7 billion the year before. Early in 2026 it was the only major technology company whose capital spending had actually declined year on year. Rather than build frontier models at any cost, Apple signed a multi-year agreement with Google in January 2026 under which Gemini models help underpin the next generation of Apple's own foundation models.
By the standards of the AI boom, that looks like abdication. Analysts and commentators spent much of 2025 asking whether Apple had missed the most important platform shift since the smartphone.
### Late, rented and litigated
The product story did little to quiet those questions. At WWDC in June 2024 Apple presented a more personal, more capable Siri. That Siri did not arrive for more than two years. When Siri AI finally shipped with iOS 27 on 14 September 2026, it came as an opt-in beta, in English first, behind a waitlist, with daily usage limits on cloud-powered features, and unavailable at launch in the European Union and China.
Along the way Apple agreed to a $250 million settlement over claims that it had advertised AI features before they existed, while denying wrongdoing. Its most important AI upgrade relies, at least in part, on a competitor's technology. On paper this is a cautionary tale: a company that overpromised, under-invested and then outsourced.
By Silicon Valley logic, Apple's AI strategy looks like a failure. The market does not appear to agree.
### The twist: customers do not buy "AI"
The more revealing evidence sits outside the technology industry, in how customers respond to the word itself. A Washington State University study of more than 1,000 US adults, published in the Journal of Hospitality Marketing & Management , found that describing a product as using artificial intelligence lowered emotional trust and, through it, purchase intention. The effect was strongest for high-risk purchases, including expensive electronics.
Brands that lead loudly with AI have learned this the hard way. When Duolingo announced in April 2025 that it would become "AI-first", the reaction was swift: media-monitoring data cited by Marketing-Interactive found around 41 per cent of global online conversation about the announcement was negative, with words like "delete" and "quitting" rising sharply. The company spent weeks clarifying that it was not replacing its people.
Apple's marketing has taken almost the opposite approach. It rarely sells "AI" as an idea. It sells privacy, design and the experience of a product that works, and it describes Apple Intelligence as running on-device or on its own Private Cloud Compute. Whatever one thinks of the delays, that positioning speaks directly to the anxiety the research identifies.
Customers do not want artificial intelligence. They want the outcome, delivered by a brand they already trust.
### What the experts say
Opinion on Apple remains genuinely split, and that disagreement is useful.
Gene Munster of Deepwater Asset Management argued in 2025 that Apple had roughly two years of runway before AI shortcomings would cost it customers, pointing to the depth of its ecosystem: more than half of Apple customers own more than one Apple device. After WWDC 2026 he remained cautious, noting that Apple had previously overpromised with demos and still needed to deliver what it showed.
Others see the restraint as strategy. Max Weinbach of Creative Strategies argued that if Apple's new features prove good and come free with its devices, standalone AI companies should be worried about their consumer ambitions. Ben Bajarin, also of Creative Strategies, described Siri as potentially becoming the control layer for consumer AI. Wedbush's Dan Ives framed WWDC 2026 as Apple finally delivering on its two-year-old promise.
Both camps can be right. Apple's position depends less on winning the model race than on owning the relationship with the customer at the moment AI becomes useful to them.
### The hidden cost of being "AI-first"
There is a broader business lesson here, and it applies far beyond trillion-dollar technology companies.
First, spending is not strategy. A capital-intensive arms race rewards those who can sustain it, and depreciation on rapidly ageing chips is unforgiving. Apple's restraint means it carries far less of that risk if AI models become interchangeable commodities, which many analysts now expect.
Second, labels carry risk. "AI-powered" has become the default descriptor in product launches and pitch decks, but the research suggests that for considered, high-value purchases the label can reduce trust rather than build it. For brands selling homes, financial products, professional services or premium goods, that should give pause.
Third, delay is survivable; broken trust is not. Apple's two-year Siri delay was embarrassing and expensive, but it did not appear to break customer loyalty, because the brand's underlying promise of privacy, quality and an integrated experience remained intact. Duolingo's problem was not that it used AI; it was that its message implied the people customers valued were being replaced.
The question is not whether to use AI. It is whether AI is the reason anyone should choose you.
### What this means for Australian businesses
Most Australian businesses will never fund a data centre. Their AI decisions are smaller but no less consequential: whether to automate customer service, rebuild marketing around AI-generated content, or rename a product line to signal innovation. Apple's case suggests a simple discipline.
- Lead with the outcome, not the technology. Describe what is faster, safer or better for the customer, and let the AI stay in the background.
- Protect the trust signals you already have. Privacy, human expertise and accountability are assets. Messaging that undermines them for the sake of appearing modern is a poor trade.
- Avoid overpromising. Apple's settlement shows the commercial cost of advertising capability before it exists, even for the world's most valuable brand.
- Measure value, not adoption. As our companion analysis on Australia's AI ambition argued, the meaningful question is whether AI investment produces measurable commercial return, not how visibly it is used.
### The Studio Orris perspective
At Studio Orris we see the same tension in brand work every week: the pressure to look current pulls against the slower work of being trusted. Apple's AI story is not proof that restraint always wins, and its delays carried real costs. But it is a powerful reminder that brands are valued on the relationship they hold with their customers, not on the technology they announce. Before you put "AI" in your headline, ask whether it strengthens the reason people already choose you, or replaces it.
So here is the question we would put to any leadership team: do you lead with AI, or with the reason people trust you?
For related reading, see our analysis: Australia's AI Ambition Is Growing. But Where Is the Return?
### Sources and research
Research note: the ~US$1 billion annual Gemini figure is reported by media and has not been confirmed by Apple. Capital expenditure figures are guidance and analyst estimates. Expert views are paraphrased from published commentary.
© 2026 Studio Orris. Original editorial analysis. This publication provides general commentary and does not constitute financial, investment or legal advice.
