Good morning,
I didn’t intend for this to be 2022’s metaverse week, but there are a couple of very interesting stories about Meta that have increased my skepticism about their current direction.
On to the update:
Meta’s Chip Decision
From The Information:
In late 2021, a team of Meta Platforms employees building a key chip for the second version of Meta’s Ray-Ban smart glasses was notified that the company had decided to go with an alternative chip from Qualcomm, according to two people familiar with the matter. The custom chip would power a variety of functions, such as taking high-quality photos and videos, on the glasses. But Meta’s augmented reality chief, Alex Himel, decided using it could delay the glasses’ launch, expected next year.
The decision was a setback for Meta’s in-house silicon unit known as Facebook Agile Silicon Team. Code-named Brasilia, the chip was part of a broader effort by Meta to control key technologies and reduce its reliance on off-the-shelf silicon providers like Qualcomm, which supplies chips for Meta’s Quest VR headsets, Portal smart video devices and the first version of the Ray-Ban glasses. By building its own power-efficient chips to manage its burgeoning stable of augmented and virtual reality devices, Meta would have far greater control over the features, size and battery life of its products and be better positioned to compete with rivals like Apple.
This was a very interesting story to me in multiple dimensions: there are layers within this story itself, and also interesting ways in which this story fits with other news about Meta’s ongoing metaverse bet.
Start with the decision itself: shipping hardware is hard, and it is completely understandable why Meta would decide that it couldn’t risk holding up the next version of its smart glasses for this new chip. The timing is right as well: Meta would need to have functional prototypes in hand now and be finishing final tapeout if they expected to ship products with these chips next year. That does not appear to be the case.
The more interesting question, though, is why Meta is making chips in the first place. It certainly doesn’t make economic sense at present: designing chips is a massive capital expense, and I doubt Meta is selling a sufficient volume of headsets — much less smart glasses — to justify the cost. Apple, to raise the obvious comparison point, was already selling 40 million iPhones a year when they launched their first chip.
Apple, of course, doesn’t just make their own chips; they also make the operating system, and the fact they make both means that the chips and the operating system can be designed to augment each other. Meta may have had a similar goal in mind, but recall that the company abandoned its homegrown operating system (XROS) last year in favor of its existing Android-based operating system (VROS). I thought that was the right move for Meta, and I concluded in a Daily Update about the decision:
Finally, consider the competition: I just noted above that Apple, thanks to this decision, is going to have a performance advantage. The reality, though, is that Apple was almost certain to have a performance advantage regardless. Trying to out-integrate Apple with a brand new operating system and maybe its own chips seems like a Herculean task; what makes far more sense is to go in the opposite direction and lean into modularity, building an ecosystem.
This doesn’t, by the way, say anything one way or the other about Meta’s commitment to an open ecosystem and interoperability on top of VROS; the power of the operating system is that, by sitting precisely at the point of leverage, it can facilitate completely independent ecosystems both below it, at the hardware level, and above it, at the application level. It is, though, consistent with that commitment, and the smartest way to take on Apple.
I’m going to get more into the second part of that excerpt below; what seems clear to me with regards to the first part is that abandoning Meta’s own operating system should have gone hand-in-hand with abandoning their custom silicon efforts, for many of the same reasons.
First, just as Android is already a fully-featured operating system, Qualcomm’s XR2 offering is already a fully-featured VR-focused chip. Meta would need to have a very clear idea of what they could do differently that would deliver a large enough delta in terms of performance to justify the investment and complexity, but again, by abandoning their own custom OS they reduced the potential delta.
Second, just as Meta’s Android-based VROS is being shipped and built-on today — along with Android broadly — Qualcomm’s offering is being used by other VR headset manufacturers, Snap, etc. Sure, this means that Meta’s offering isn’t differentiated, but Meta isn’t shipping at volume either: using the same chip as everyone else means pooling chip development costs in one place, reducing costs for everyone while also increasing R&D efficiency.
Third, just as Android already works across many devices, meaning that VROS gets extensibility for free, Qualcomm’s chip is, as noted, already shipping across multiple devices from different OEMs. To the extent that Meta wants to run on devices other than Oculus headsets, having a common hardware platform is a good thing.
More on this in a moment, but first another relevant story.
Horizon Worlds’ Platform Fees
From The Verge:
Meta is testing new features to let creators make money within Horizon Worlds, the company’s social metaverse platform for Quest VR headsets that is soon coming to mobile phones and possibly game consoles. Most notably, a “handful” of Horizon creators will be able to sell virtual items and effects in the worlds they create for others to explore. The idea is that creators can sell everything from access to a VIP section of their world to virtual items like jewelry or a special basketball, according to Meaghan Fitzgerald, the product marketing director for Horizon. Participants in the US will also be able to earn money from a $10 million creator fund Meta recently set up to reward creators with the most engaging worlds…
Meta will be taking a cut of what creators sell, though exactly what that take can be is a bit complex. For Horizon purchases, Meta is taking a 25 percent cut of the percentage that’s left after a platform fee. For platforms with a 30 percent fee, like Meta’s own Quest Store for VR titles, the creator will be left with a little over half of the sale price (the math there being that Meta is taking 25 percent of 70 percent).
Casey Newton explained how this works on Platformer:
Horizon Worlds is now available on the company’s Quest virtual reality headsets. Eventually, the company wants to bring it to mobile phones, game consoles, and other devices. Those devices are not made by Meta, though, and so creators selling goods on Horizon will be taxed twice: first by Horizon, then on the platform Horizon is operating on. And that’s before the creator pays regular taxes to the government.
Amazingly, this is true even on platforms that Meta owns. If you’re a creator selling $10 worth of virtual goods through the Quest, Meta’s Oculus store will keep $3, and Horizon will keep 25 percent of the remaining purchase — in this case, another $1.75. That leaves the creator with $5.25.
Newton is more bothered by the fees than I am. I think that both Worlds and Quest have a long ways to go before they stop being console-like and start being computer-like; were the iPhone a niche product instead of, you know, the single most important platform in the world, I wouldn’t be bothered by its current console model. Moreover, the the dual-tax model makes more sense when you consider that Meta plans to have a Horizon Worlds apps on iOS and Android; in that case Apple gets 30%, and Meta 25% on top (a la Roblox, for example).
At the same time, I definitely get Newton’s confusion about Meta collecting the fees twice. Sure, Meta is treating Quest like iOS, and Horizon Worlds is a platform that sits on top of that, but again, these products are from the same company! Apple isn’t forcing Meta to collect 30% on its own hardware — if anything, Meta, after all of its harrumphing, is validating Apple’s approach by imitating it, even as it simultaneously insists it is offering a cross-device platform. That right there gets to the root of the confusion: Meta can’t decide on a strategy.
Meta’s Missing Metaverse Strategy
I have been skeptical of Meta’s VR strategy from the day the company bought Oculus; I wrote in 2018’s The Problem with Facebook and Virtual Reality:
If Facebook wanted a presence in virtual reality the best possible route was the same it took in mobile: to be an app-exposed service, available on all devices, funded by advertising. I have long found it distressing that Zuckerberg, not just in 2014, but even today, judging by his comments in keynotes and on earnings calls, seems unable or unwilling to accept this fundamental truth about Facebook’s place in tech’s value chain.
I’ve softened a fair bit on a lot of my rhetoric in that Article, and certainly Apple has shown just how fragile it is to be “just an app.” At the same time, the fundamental tension between being a horizontal and vertical company can’t be wished away. This is one of the oldest themes on Stratechery; I wrote in 2013 in the context of Android:
Look again at the Mobile Hierarchy of Needs:
Apple invests in software, apps, and services to the extent necessary to preserve the profit they gain from hardware. To serve another platform would be actively detrimental to their bottom line. Google, on the other hand, spreads their services to as many places as possible — every platform they serve increases their addressable market.
The point of that Article was to note that Google had become too focused on Android, and to the extent that that focus hurt Google’s iOS business it was bad for the company — the proverbial tail wagging the dog. This is always the danger of services companies building hardware, and was the root of my criticism of Meta’s Oculus acquisition. To further explain:
- A vertical VR strategy would mean building your own differentiated hardware, further differentiating it with your own software and services, selling it at a profit, and charging for access to the ecosystem. The Apple model, in other words.
- A horizontal VR strategy would mean building an experience that ran across devices from all kinds of different manufacturers and even different operating systems; you could charge for access to the ecosystem, or run an advertising business. The current Facebook model, in other words.
Building hardware certainly pushes a company towards a vertical strategy, which as I noted in the excerpt above, didn’t make sense for Meta; this was my worry with the acquisition. At the same time, Google did eventually find the right balance when it came to hardware, just as Microsoft has with Surface (and, more broadly, Windows). And, by extension, I can understand why Meta feels it needs to build hardware: the company doesn’t want Apple to lead the way, but then who else is going to make the investments necessary for the ecosystem to reach critical mass? To that end I can understand a strategy where Meta uses Oculus to push the ecosystem forward, while explicitly avoiding a vertical business model.
This, though, is why the above story is disconcerting: a model where Oculus takes a fee and Horizon Worlds takes a fee is Meta trying to have its cake and eat it too — it wants its hardware to make money, and its service to be everywhere. It evinces a company that doesn’t want to make choices — the essence of strategy — and at this point, it’s starting to be a pattern. Go back to the operating system decision: this wasn’t a six month bake-off before committing resources; the now-abandoned XROS was in development for four years and had over 300 employees working on it. It’s the same thing with the chip question: Meta isn’t choosing one direction, it’s trying to do both at the same time.
This, I would note, is one of the biggest reasons why established companies with unlimited funds struggle with new initiatives: a lack of resources forces choices, and choices are strategy; unlimited funds, on the other hand, increase the temptation to eternally prioritize optionality without ever actually making hard decisions.
For what it’s worth, I still believe that Meta’s ultimate goal should be to be a horizontal service that works across different devices from different manufacturers. Finally deciding on VROS is a step in that direction, and Meta should commit to the Qualcomm ecosystem — yes, Qualcomm has struggled for a while, but Meta’s ability to do chip design better, particularly if its not building the lowest layers of its OS, is limited (and this is a decision that can be revisited in the future); better to pull on the same oars as everyone else outside of Apple. This also means that platform fees should come from the horizontal service, not the underlying hardware.
I would go further than that: Meta’s goal should be for VROS to be the default operating system for all VR headsets, not just the one it makes. More than that, the hardware the Meta is developing should be made available to other manufacturers (as long as they use VROS, of course). Not only does this approach fit with Meta’s fundamental nature as a service, it is also the best strategy to compete with Apple: you’re not going to do vertical integration better than the iPhone maker, but that means there is an opportunity to build out a modular ecosystem while holding onto the piece that ties it all together. It will be a hard thing to accomplish, to be sure, but until Meta fully commits to a strategy they are making success that much more difficult to achieve.
This Update will be available as a podcast later today. To receive it in your podcast player, visit Stratechery.
The Stratechery Update is intended for a single recipient, but occasional forwarding is totally fine! If you would like to order multiple subscriptions for your team with a group discount (minimum 5), please contact me directly.
Thanks for being a subscriber, and have a great day!