Sept 2024

Reality Labs leads Meta's fight to define the future of digital interaction. Competition from Apple and Google are on the fronts of training data, hardware, and silicon, instead of the mobile, cloud, and platform battles of the past decade.

The LLM paradigm shift has weakened the long-standing moat of incumbents like Apple in mobile UX. A tight integration between hardware and software is no longer enough, AI is now required for the best possible user experience, creating a space for Google to reclaim mobile marketshare if it can integrate it's Pixel + G-series chips with Gemini into something more useable.

If Google and Apple boast opposing strengths between mobile hardware and software, Meta introduces as it's strength a third variable: having the only AR/VR devices truly deployed at scale.

While lacking a mobile platform of it's own, the Oculus and Meta Ray-Bans cover both ends of the AR/VR spectrum, with convergence of the two into an eventual single device seemingly likely amid rumors of Nazare.

Given Meta's lead in AR/VR distribution and the success (and indeed mere existence, in the case of Apple) of open-sourced LLaMa, I think it makes sense for Meta to:

1) take the mobile market with an AI + AR/VR + call-enabled wearable

2) build the mobile platform it missed out on, positioned as a more open App Store

A unique opportunity in mobile UX

Despite Android's customizability, iOS dominates because Apple's seamless user experience is more valuable than additional power and control to the average user. But AI has introduced a new UX component with which value can be provided outside of this traditional HW-SW integration realm.

Primarily a service-based offering, AI requires Apple to partner with an external firm like OpenAI to build that which it can't. Apple's functionally-organized firm builds beautifully integrated devices at the expense of mediocre (relatively) services, which require a more product-oriented org like Google (where service quality trumps device).

This presents a rare opportunity for Google to capture mobile market share by tightly coupling their Pixel phone, Gemini, and 5th-generation G-series chip into an Apple-meets-AI user experience fueled by the AI paradigm shift.

But instead, Google announced support for Android as a whole, likely due to needing to balance relationships with their OEMs. Prioritizing the Pixel probably wouldn't sit well with Samsung, who manufactures Android phones and until recently, their G-series chips as well.

Therefore a path exists for Meta to jump forward and create a tightly-integrated, Meta-silicon-to-LLaMA wearable.

Since Meta isn't tied to any OEMs or widespread operating system like Android, it can tightly couple it's hardware into product offerings for an Apple-like experience, while also integrating the high-quality service products (LLaMA, Instagram, WhatsApp, ads) that Apple lacks.

Platform positioning as a service company

Another advantage Meta has over Apple is the ability to offer a more open platform.

Apple's revenue overwhelmingly comes from it's devices, which means it needs to incentivize the walled garden, where it can get its 30% App Store cut.

Whereas Meta already has a strong service-based revenue stream:

meaning it can trade the guaranteed App Store cut for a more open model akin to Google, where merely keeping users on Meta's platform directs them to their other service lines.

Meta has already experienced the goodwill from their releasing of LlaMA, encouraging developers to improve the platform and become familiar with the ecosystem.

All this in conjunction with the tightly-coupled hardware/software/LlaMA integration mentioned earlier provides a powerful differentiator against both Apple and Google.

Risks

While Google has announced support for Android over just their Pixel, they've also laid the groundwork for reducing their OEM relationships. Google shifted design for their G5 chip from Samsung LSI to in-house, and is using TSMC instead of Samsung Foundry for manufacturing, so the groundwork may be laid for a tighter Pixel-Google coupling in the future.

Similarly, Apple is sitting on $40 billion in cash and likely has the technical ability to train a foundational model to compete with LlaMA, if not the organizational ability.

Regulatory headwinds are a looming threat, especially in areas like the EU, and with the uncertainty around AI and alignment. Meta's done a great job so far in securing goodwill through open-source releases and Mark Zuckerberg's overall PR.

Manufacturing is a risk to consider, with the decline of Intel we have no serious contender for US-based manufacturing, with TSMC the default. But this carries geopolitical risk due to China's relationship with Taiwan, with the other option being Samsung Foundry.

Which brings us to a final possibility…

Moonshots

As a long-shot perhaps Meta will develop the ability to manufacture chips on U.S. soil.

With the decline of Intel, America has lost it's ability to manufacture competitive chips domestically. With the contentious relationship between China and Taiwan, this has become a national security issue, which actually provides an opportunity to get government backing.

As the cost will likely be hundreds of billions of dollars, and require supply chain revamps to secure EUV and raw materials, training, and construction -- Meta could work with the U.S government to subsidize the development in conjunction with Meta's technical expertise.

While a steep learning curve and long time horizon, the payoff could be a much deeper integration for Meta, as well as a possible new revenue stream and risk mitigation on the manufacturing front.