Although Apple’s 2025 software updates appeared to shift the marketing spotlight toward the “Liquid Glass” visual interface rather than loudly championing generative AI—prompting criticism that the company was lagging in the AI race—reporting from The Information suggests that this seemingly “conservative” approach may in fact be the key to a comeback in 2026.
In recent years, Meta and Google have poured hundreds of billions of dollars into massive data centers and aggressive AI chip acquisitions to train large language models. By comparison, Apple’s moves have been notably restrained.
The report argues that this “low-spend” strategy has placed Apple in a uniquely advantageous position: cash is king.
Apple is estimated to still be sitting on roughly $130 billion in cash reserves. This war chest affords the company extraordinary flexibility heading into 2026. Unlike competitors that have already burned vast sums on hardware infrastructure without yet seeing meaningful revenue returns, Apple can swiftly bolster its capabilities through targeted acquisitions of promising AI startups or strategic investments in specific technologies. Beyond frugality, Apple’s perspective on AI models also diverges from the prevailing Silicon Valley consensus.
According to the report, some Apple executives believe that large language models will ultimately become commoditized—much like electricity or the internet—where differentiation steadily erodes. From this viewpoint, pouring enormous resources into developing proprietary LLMs is difficult to justify commercially.
As a result, Apple is reportedly planning a “leverage rather than reinvent” strategy for its AI features slated for 2026, opting to integrate Google’s Gemini models directly into its services. This allows Apple to focus its resources on refining the user experience instead of duplicating efforts already being undertaken elsewhere.
That said, Apple is not abandoning in-house development altogether. Despite a period of internal turnover, the company continues to maintain a team working on proprietary models as a contingency. Organizationally, following the retirement of former AI chief John Giannandrea, Apple’s AI efforts are now led by Mike Rockwell.
Rockwell previously spearheaded the development of Vision Pro, a detail that has fueled speculation about Apple’s future AI direction. Observers suggest the company may emphasize deep integration with spatial computing and hardware devices, rather than positioning AI primarily as a standalone chatbot.
The report also underscores that the iPhone remains Apple’s strongest card. No matter how powerful cloud-based AI models become, they ultimately require a delivery mechanism—and the iPhone stands unrivaled as the world’s most effective platform for AI applications, a hardware moat that Google and Microsoft struggle to match. In this light, Apple’s apparent “slowness” may be better understood as a calculated steadiness.
Today’s AI market undeniably carries bubble-like characteristics, with the costs of training large language models often far outpacing the profits they generate. By sidestepping this capital-intensive arms race and choosing to enter later—once the technology matures and costs decline, through integration with models like Gemini—Apple is adhering closely to its long-held philosophy: not to be first, but to be the most usable.
If large language models do indeed become a ubiquitous commodity by 2026, Apple’s $130 billion in cash and its billion-strong iPhone user base will give it unparalleled power to define what AI looks like at the application layer. For Apple, AI is simply a means to make Siri smarter and photos more beautiful—not the sum total of the company’s identity. That, perhaps, is the healthiest model of all.