Apple is confronting growing pressure to redefine its artificial intelligence (AI) strategy after a turbulent first year for its Apple Intelligence initiative. While competitors like OpenAI, Meta, and Google continue to charge forward with hardware and software integrations powered by advanced generative AI, Apple has struggled to deliver compelling features that justify its massive installed user base and proprietary hardware ecosystem.
When CEO Tim Cook takes the stage at Apple’s Worldwide Developers Conference (WWDC) in Cupertino, all eyes will be on how the company plans to course-correct its AI ambitions. Apple’s challenges have shifted from mere execution problems to existential questions about the future relevance of its hardware.
A Rocky Debut for Apple Intelligence
Apple Intelligence introduced in 2023 as the company’s response to the rise of generative AI, aimed to embed AI-powered functionality directly into iOS. Analysts saw promise in Apple’s 1 billion+ iPhone user base, coupled with custom silicon like the M1 and A-series chips. But the rollout has disappointed.
Initial features included revamped Siri animations, a text rewrite tool, and the ability to generate slideshow videos using photos. However, these underwhelming capabilities failed to meet user expectations. A more advanced feature—AI-powered summarization of long message threads—was turned off after media outlets discovered it altered headlines in misleading ways.
The most damaging setback came in March when Apple delayed its highly touted “More personal Siri” update. Promised to integrate deeply with iPhone apps—retrieving email details or making reservations—it was featured prominently in iPhone 16 advertising. The delay led to Apple removing its promotional content and triggered a class-action lawsuit by consumers, who alleged false advertising.
Investor Concerns and Market Pressure
Investors, who were initially optimistic about Apple’s AI direction, are now showing signs of impatience. JPMorgan Chase analyst Samik Chatterjee described expectations for WWDC as “lackluster,” noting that Apple has yet to deliver on its promises from last year.
Compounding the issue is intensified competition in the AI hardware market. OpenAI acquired startup io for $6.4 billion and recruited former Apple design chief Jony Ive to develop next-gen AI devices. Meta’s Ray-Ban Meta Glasses have sold over 2 million units, integrating the Llama language model for real-time interaction. Google, meanwhile, made its Gemini AI the default Android assistant, demonstrating video summarization and multimodal capabilities.
Google also inked a $150 million deal with Warby Parker to create its own AI glasses. These moves starkly contrast Apple’s limited public-facing AI advancements and underline its trailing position.
The Threat to the iPhone Itself
Apple’s dominance has always been closely tied to the iPhone. However, according to Apple Services Chief Eddy Cue, that dominance could be upended. In testimony during the government’s antitrust case against Google, Cue warned: “You may not need an iPhone 10 years from now.”
Cue’s point underscores fears that AI-powered devices could replace smartphones altogether. OpenAI and Ive’s collaboration, focused on building an AI-native device, signals a potential paradigm shift—one that could erode Apple’s core business if it doesn’t respond aggressively.
Behind the Curve in Frontier AI
While Apple trains its models for Apple Intelligence, it has steered clear of developing “frontier models”—those robust systems that lead the pack in capability and require billions in GPU-powered infrastructure. Instead, Apple leases computing power from third-party cloud providers, such as Google Cloud.
In fiscal 2024, Apple spent $9.5 billion on capital expenditures, which accounted for approximately 2.4% of its revenue. In contrast, Meta, Amazon, Alphabet, and Microsoft are collectively spending over $300 billion this year, up from $230 billion in 2023. Amazon alone plans to invest $100 billion, while Microsoft plans to invest $80 billion. The disparity in investment signals a different level of commitment to AI.
Is Acquisition the Only Way Out?
Apple has a long history of leveraging strategic acquisitions to enhance its competitive strengths. It bought PA Semi in 2008 to kickstart its chip division. Before launching the Vision Pro headset, it acquired over 10 startups focused on AR and VR. Siri was itself a startup acquired for $200 million in 2010.
Analysts, such as Gene Munster of Deepwater Asset Management, suggest that Apple may need a bold acquisition to catch up in AI. “They probably need to acquire Anthropic,” he said. With a $61.5 billion valuation as of March 2025, Anthropic would be Apple’s largest acquisition by far.
Another potential target is Perplexity AI, which recently partnered with Motorola and is reportedly in talks with Samsung. Apple has held discussions with Perplexity, as revealed during Cue’s testimony in court.
Still, even with $133 billion in cash, acquiring a company like OpenAI—valued at $300 billion—is likely out of reach, especially given regulatory scrutiny and competitive entanglements. Ive’s partnership with OpenAI further complicates any strategic collaboration.
Plug-and-Play Partnerships?
Apple may also pursue a modular approach. As Craig Federighi hinted last year, Apple could allow third-party models, such as Google Gemini, to plug into the Apple Intelligence ecosystem. This approach resembles how Apple integrated Google Search and could buy it time to develop its frontier models.
Documents from the Google antitrust case reveal that Apple’s senior executives, including Cue and Adrian Perica, were directly involved in negotiations regarding the integration of Gemini. Such flexibility could help Apple leapfrog interim development gaps.
Apple’s Hardware Advantage
Despite falling behind in AI software, Apple retains unique strengths. Its silicon, especially the M-series chips, is optimized for AI. David Stout, co-founder of WebAI, says Apple’s unified memory and efficient GPU setup makes it ideal for running large models.
“We picked Apple Silicon because we think it’s the best hardware for AI,” said Stout. His team ran Meta’s Llama models on Mac Studio clusters and found that Apple chips could deliver 100 million tokens per dollar—far more efficient than Nvidia’s H100, which outputs just 12 million.
Apple also introduced “App Intents” to enable seamless Siri integration into third-party apps. But while the system powers widgets and lock-screen tools, the full implementation for Siri remains unreleased—another sign of delay.
Consumer Loyalty Buys Apple Time—For Now
Even as AI threatens to upend the smartphone-centric computing model, Apple enjoys extreme brand loyalty. Gene Munster believes Apple has time but not an infinite runway.
“This is still something that has an existential risk to all these companies, including Apple, but I don’t think we’re at some break point in the next year around it,” he said.
Still, with every passing quarter, Apple loses more ground to rivals that are setting the pace of AI development and deployment. If Apple doesn’t act decisively—whether through partnerships, acquisitions, or product reinvention—it risks becoming a spectator in the next tech revolution.
About Author
