Meta's AI Device Dreams:  Challenging Apple and Google

Clare Pleydell-Bouverie from Liontrust Asset Management thinks Meta could challenge Apple and Google to become the "AI king on our devices," citing the company's access to 3.48 billion users across its family of apps. It's a compelling theory—until you examine the fundamental structural barriers that make this more wishful thinking than strategic reality.

Meta's ambition to dominate device AI isn't just facing an uphill battle—it's climbing Everest in flip-flops. The company's recent struggles with AI development, combined with inherent platform disadvantages and architectural challenges, reveal why this fight was over before it began.

The Operating System Reality Check

The most glaring obstacle in Meta's path is one they can't buy, build, or acquire their way around: they don't own an operating system. As Ian Fogg, tech industry analyst at CCS Insight, puts it bluntly: "The challenge in the consumer space that Meta has compared to Apple or Google is the idea of personal intelligence. This means that you need to understand the individual's data information and tailor the AI to that. If you're on a device as part of an operating system, which is what Gemini is or whatever Apple does, you have a clear advantage there."

This isn't a minor technical hurdle—it's an existential moat. Apple and Google control the foundational layer where AI integration happens. They determine what data AI systems can access, how they interact with users, and which features get prioritized. Ben Barringer, global technology analyst at Quilter Cheviot, crystallizes the problem: "The problem Meta has is it does not own a platform like Microsoft or Apple does – it is a collection of apps that are hosted on others' operating systems. As such, it is far easier for Google, Apple and Microsoft to push its AI tools to users in a way that Meta simply can't."

Meta's apps may reach billions, but they're ultimately tenants in someone else's building. Apple and Google are the landlords who control the infrastructure, utilities, and house rules.

The Talent War Meta is Losing

Meta's recruitment struggles tell a more damaging story than their optimistic investor pitches. Despite offering packages worth hundreds of millions—including a $200 million deal to poach Apple's Ruoming Pang—"some potential candidates have been hesitant to join Meta Platforms' efforts because of the challenges that its AI efforts have faced this year, as well as a series of restructures that have left prospects uncertain about who is in charge of what."

When your compensation offers are making headlines and you're still struggling to attract top talent, that's not a negotiating position—it's desperation. Meta's struggles reached a head in April, "when critics accused the company of gaming a leaderboard to make a recently released AI model look better than it was. They also delayed the unveiling of a new, flagship AI model, raising questions about the company's ability to continue advancing quickly in an industrywide AI arms race."

The tech industry is small enough that reputation matters, and Meta's reputation in AI development is currently toxic among the very people they need most.

The Technical Reality Behind the Marketing

Meta's AI performance in practice reveals the gap between ambition and execution. Users report significant reliability issues: "Meta AI struggles with key drawbacks, including limited technical reasoning and inconsistent memory" and "struggles with multi-part queries within a single conversation." For trip planning scenarios, "It will likely forget your budget or destination preferences at some point mid-conversation, forcing constant backtracking."

These aren't minor bugs—they're fundamental limitations that make Meta AI unsuitable for the complex, contextual interactions that define next-generation device AI. When "users have reported that when it comes to image generation, Meta AI struggles to accurately incorporate different art styles, often producing results that feel generic or off-target," you're not seeing a product ready to challenge industry leaders.

The technical gap becomes more obvious when examining Meta's approach. "Meta AI's integration with Facebook, Instagram, and WhatsApp is a huge advantage for some, but it can feel limiting if you're not deeply invested in Meta's platforms... Many of its features are designed to steer you back towards Meta's ecosystem." This circular, platform-first approach fundamentally misunderstands what device AI needs to be: universal, contextual, and seamlessly integrated across all user activities, not just social media engagement.

The Distribution Mirage

Pleydell-Bouverie's argument about Meta's "extraordinary distribution advantage" because they're "in the pocket of 3.48 billion users" sounds impressive until you realize that those users are accessing Meta through Apple and Google's platforms. Meta's distribution is entirely dependent on the goodwill of their competitors.

This dependency creates an inherent strategic vulnerability. Apple has already demonstrated its willingness to limit Facebook's tracking capabilities with App Tracking Transparency. Google could implement similar restrictions at any time. Meta's "distribution advantage" exists only at the pleasure of companies whose business models they're trying to disrupt.

Moreover, Apple has passed on partnerships with Meta, with "Bloomberg reported Monday, June 24 that Apple passed on the deal months ago" after brief talks. When your potential partners are actively choosing not to work with you, your distribution strategy needs serious reconsideration.

The Ecosystem Reality

Apple's approach reveals why platform control matters. "The one advantage Apple has that its tech rivals don't: the world's most monetized installed base, more than two billion users running its iOS operating software on devices." Apple doesn't just have users—they have an integrated ecosystem where AI can access calendar data, location history, communication patterns, and device usage across all touchpoints.

Meta's data, while extensive, is fragmented across social platforms and lacks the device-level integration necessary for truly personal AI. Apple is reportedly using "2025 as a 'gap year' for AI, which means all eyes will be on enhancing Siri and more services through advanced AI technology in 2026." This strategic patience, combined with ecosystem control, positions Apple to deliver AI that understands users holistically rather than just their social media behavior.

The Hardware Reality Check

Meta's attempts at hardware diversification through smart glasses and VR headsets face their own challenges. While "Meta has made a splash with its Ray-Ban Meta Glasses, selling over 2 million units since launching in 2021," this pales next to Apple's iPhone ecosystem that ships over 200 million units annually.

The smart glasses market, while promising, remains niche. New competitors like Halliday Glasses are emerging with superior display technology, and "Xiaomi's AI Glasses are feature-rich with AR display, smart payments, and deep ecosystem integration" shows that hardware innovation alone won't secure market dominance.

Meanwhile, "Apple is reportedly reorganized some of its AI teams" and working on its own smart glasses for a 2027 launch, giving them time to learn from Meta's mistakes while leveraging their superior hardware manufacturing capabilities and ecosystem integration.

The Strategic Delusion

Meta's AI strategy suffers from fundamental strategic confusion. Stratechery's Ben Thompson notes: "There is a case to be made that Meta is simply wasting money on AI: the company doesn't have a hyperscaler business, and benefits from AI all the same." Meta is spending billions to compete in a space where their core advantages don't apply and their structural disadvantages are most apparent.

The company's open-source Llama strategy, while generating positive PR, doesn't solve their device integration challenges. Meta's approach of "open-sourcing these compressed models and partnering with chip makers Qualcomm and MediaTek" attempts to bypass platform gatekeepers but ultimately puts them at the mercy of hardware partners who also work with their competitors.

Math Doesn't Lie

Meta's AI device ambitions face insurmountable structural barriers:

Control: Apple and Google own the platforms where device AI lives

Talent: Top AI researchers are actively avoiding Meta due to reputation issues

Integration: Meta's social data, while extensive, lacks the device-level context needed for personal AI

Distribution: Meta's reach depends entirely on competitors' platforms

Technical execution: Current Meta AI performance reveals significant capability gaps

The most damning evidence isn't in the analysis—it's in the actions. "Investors have grown increasingly concerned with Apple's lack of AI strategy" yet Apple stock has maintained relative stability while Meta spent billions chasing AI dominance. The market understands what Meta's boosters don't: platform control trumps user counts.

Meta may have 3.48 billion users, but they're all using Apple and Google devices, running Apple and Google operating systems, subject to Apple and Google's rules. That's not a competitive advantage—it's a structural dependency that makes Meta a service provider, not a platform controller.

The dream of becoming the "AI king on our devices" requires owning the devices, or at least the platforms that control them. Until Meta solves that fundamental problem, they're not challenging Apple and Google—they're just renting space in their kingdoms.

Ready to build AI strategies based on realistic market positioning rather than wishful thinking? Winsome Marketing's growth experts help companies identify genuine competitive advantages and avoid expensive strategic dead ends. Let's discuss how to leverage AI effectively within your actual market position—not the one you wish you had.