Apple's new chief executive is losing patience with the company's cautious approach to artificial intelligence. CEO John Ternus has been pressuring senior vice president of software engineering Craig Federighi to accelerate Apple's AI roadmap and move quickly beyond incremental improvements to Siri, according to Bloomberg's Mark Gurman, writing in his Power On newsletter on Sunday.
The report describes a closed-door dynamic in which Ternus has goaded Federighi to make changes, frustrated by what he views as hesitance to embrace AI with the same enthusiasm as rivals. For a company that has spent two years defending its deliberate AI strategy, the internal pressure campaign is the strongest sign yet that Apple's leadership sees the cautious approach as a liability rather than a virtue. For context on how the rest of the industry is moving, check our latest AI developments.
What Gurman Reported
According to the Power On newsletter, Ternus has pushed Federighi to "accelerate the company's road map and move quickly beyond merely making necessary improvements to Siri." The phrasing matters: it suggests Apple's internal debate is no longer about whether to invest in AI, but whether the company can afford another cycle of polish-first, ship-later releases that improve existing features without producing anything categorically new.
Gurman cited sources familiar with the matter, and characterized the CEO as frustrated by Federighi's caution. Federighi, who runs the software organization that owns Siri and Apple's on-device AI features, has a well-documented reputation inside the company as an AI skeptic — Gurman notes he has voiced concerns about the technology's privacy, security, societal and employment implications.
That skepticism once aligned neatly with Apple's brand. The report adds that "the world has changed around Federighi — and around Apple itself," arguing that Apple no longer has the luxury of dismissing the AI efforts of Meta, Google, OpenAI and Anthropic.
Why This Fight Defines Ternus's Tenure
Ternus's rise to CEO was itself a signal about Apple's direction. The former hardware chief took over from Tim Cook, who moved to an executive chairman role, with a mandate to ship products rather than manage transitions. His public comments so far have been consistent: in a recent interview, Ternus argued Apple won't ship "AI for AI's sake," insisting the company would use AI aggressively where it makes products better but refuses to add features simply to match competitors' marketing.
The Gurman report suggests that philosophy is now colliding with a new reality. Agentic AI — systems that complete tasks rather than answer questions — has become the industry's central product narrative, and Apple's rivals are shipping it at a pace that makes Cupertino's incremental Siri updates look increasingly out of step. OpenAI, Google, Meta and Anthropic are all releasing agents that browse, code, book, and buy on the user's behalf.
Apple, by contrast, has spent the Siri era rebuilding foundations: migrating the assistant to hybrid on-device and cloud architecture, licensing external models, and promising a more personal context-aware experience that has repeatedly slipped. Each delay has narrowed whatever technical lead Apple's privacy-first design once implied.
Federighi's Position Is Not Simply Obstruction
It would be a mistake to read the report as a simple tale of a skeptic blocking progress. Federighi's concerns — privacy, security, societal impact, employment — are the same issues that have generated regulatory scrutiny, lawsuits and public backlash across the AI industry this year. Apple has faced its own reckoning with AI-assisted features that misfire, and regulators on multiple continents have opened investigations into how AI systems handle user data and minors.
An engineer's caution in that environment is a defensible product philosophy. The problem Gurman describes is one of pace and positioning: Apple's brand of careful AI only wins if the company eventually ships experiences that are clearly better, and each year of incremental Siri releases makes that promise harder to credit.
There is also an organizational reading. Federighi's software organization controls the surfaces where AI features land, and any acceleration plan will run through his group. Ternus's pressure campaign is, in effect, a public-in-internal signal about priorities and staffing — the kind of message CEOs send when they want reorganizations to be understood as inevitable rather than optional.
What Acceleration Would Actually Look Like
For Apple, a faster AI roadmap likely means several concrete shifts. First, deeper reliance on external frontier models for cloud features while Apple's own models handle on-device tasks — a hybrid approach the company has already begun. Second, shipping agentic features in narrower domains where failures are contained: scheduling, photo curation, on-device search and writing tools, rather than open-ended assistants. Third, loosening internal review cycles that treat model behavior bugs as ship-blocking defects, which historically has been the biggest structural brake on AI feature velocity at Apple.
The competitive math is unforgiving. Apple's most valuable assets — the installed base, the silicon, the distribution — remain unmatched. But AI leadership is accruing to companies that ship agents people use daily, and the iPhone's app-and-Assistant model was designed for a search-and-tap internet that AI is actively replacing.
Neither Ternus nor Federighi has publicly commented on the report, and Apple declined to discuss internal deliberations. But the direction of travel is clear from the outside: a new CEO elected to move fast is telling the company's most cautious senior executive that the era of moving carefully is ending.
Stay Ahead of AI
Track every AI strategy shift — from Apple to OpenAI — with independent, multilingual coverage of the AI industry.
Read more AI news →
