Good morning,

This interview was conducted yesterday morning as Meta was broadcasting its pre-recorded Connect keynote. One of the biggest announcements in that keynote was that Meta was partnering with Microsoft (and Accenture) to bring the metaverse to the enterprise.

To that end, I had the chance to speak with Microsoft CEO Satya Nadella and Meta CEO Mark Zuckerberg about their partnership; this is the first part of the interview. Nadella had to leave after twenty minutes, and the rest of the interview was with Zuckerberg about not just the Microsoft partnership but the company’s metaverse efforts generally, plus some additional questions about AI and competing with TikTok.

For subscribers: this is the weekly Stratechery interview, just on a Wednesday instead of the usual Thursday. There will not be an additional post tomorrow.

To listen to this interview as a podcast, sign-up for a Stratechery account to add Stratechery to your podcast player.

On to the interview:

An Interview With Meta CEO Mark Zuckerberg and Microsoft CEO Satya Nadella About Partnering in the Metaverse

This interview is lightly edited for clarity; in addition, the interview started with only Nadella; Zuckerberg joined about two minutes in.

Meta + Microsoft

Satya Nadella, it’s good to see you. I do have to say one of my favorite things about the shift to the Microsoft 365 strategy and Teams in the cloud is I can now talk to a Microsoft executive and not feel guilty that I’m sitting here with an Apple laptop because it all aligns with your strategy. It’s a nice shift from a decade ago.

Satya Nadella: Yeah, you’re using Teams and I’m thrilled about that.

Excellent. I’m going to start with you, Mark will be joining us in a little bit, but I feel very nervous about this announcement of your and Meta’s partnership because it seems to involve a very high risk of confirmation bias on my part, given that my concluding thought about the Metaverse when it was a big topic last year is that it seems that enterprise is the obvious place for it to start. That is why I tipped Microsoft as being the favorite in the space.

Still, it didn’t occur to me then that, well, if you had just worked together with Meta, you would get all the pieces together. To that end, I’m curious, how has your view of this evolved over time? Was this partnership something that you had your eye on for a while or was it just a happy coincidence that you came together? Walk me through how your thoughts have shifted, particularly given the context of having HoloLens? How does this partnership make sense for you.

SN: Obviously a lot of what we are doing in Mixed Reality has been informed with what we have done in HoloLens and what we are seeing in terms there, especially since we focused it very quickly after its initial launch on the enterprise and the business use cases and we’ve learned a lot. But the way I come at it, Ben, is that I like to separate out, “What is the system, what are the apps”? Of course, we want to bring the two things together where we can create magic, but at the same time, I also want our application experiences in particular to be available on all platforms, that’s very central to how our strategy is.

For example, when I think about the Metaverse, the first thing I think about is it’s not going to be born in isolation from everything else that’s in our lives, which is you’re going to have a Mac or a Windows PC, you’re going to have an iOS or an Android phone, and maybe you’ll have a headset. So if that is your life, how do we bring, especially Microsoft 365, all of the relationships that are set up, the work artifacts I’ve set up all to life in that ecosystem of devices? That’s at least how I come to it and that’s where when Mark started talking to us about his next generation stuff around Quest was pretty exciting so it made a lot of sense for us to bring — whether it’s Teams with its immersive meetings experience to Quest or whether it’s even Windows 365 streaming, and then, of course, all our management and security and even Xbox, that’s what is the motivation behind it.

Zuckerberg joins the interview.

How about you, Mark? How did this evolve to where this partnership made sense? It’s interesting because while Meta has had their workplaces, the reason I’ve been optimistic about enterprise in this space it has just made sense from a go-to-market perspective. Specifically, it’s hard to grok the value of this VR stuff until you try it, and a good way to try it is your employer gives you a headset and thus you have no choice. But then if you like it you might want to use it at home, which mirrors the PC adoption cycle. Is that something that’s been in your head all along? Where you wondering “What’s the best place to go-to-market and get people to try this out?”.

Mark Zuckerberg: I think work software is pretty different from social software, and then there’s a range within work. Enterprise is almost its whole own beast, and Microsoft I think is really the leader and the best at this, and has been for quite some time. When I was thinking about how this platform is going to develop overall, taking a step back, I think that there are going to be two major price points, if you will, for devices and bundles of technology. I think there’s going to be a consumer device like Quest 2 and Quest 3, the next generation that we’re working on, we’re not releasing it right now.

Oh, breaking news!

MZ: Well, I don’t think that that’s a surprise to anyone, it’s not this year, but there will be a Quest 3 and that’s in the price range of $300, $400, or $500, that zone. Then you have this Pro work line where I think high-end professionals and knowledge workers are willing to pay $1,500 for a laptop, $2,000 for a laptop for your workstation, and that basically gives us the ability to build in a lot of technology there, and also improve that technology and build out the development ecosystem before we can get that down to the price point that’ll fit with consumers.

So there are these two logical zones to build out the next platform and consumer is obviously where we focus the most, but there are 200 million new PCs that are bought every year primarily for doing work. That’s a huge part of the market and what we realized early on was that we had a place in helping to build out that platform. The technology is eventually going to be universal across all different parts of VR and eventually AR, but we’re not the leading enterprise company, Microsoft is. I think this is actually quite a natural partnership for the companies where a lot of companies I think want a device like this but loaded up with all of the Microsoft stuff that they know how to use and trust and built that relationship with Microsoft over, in many cases, decades.

Satya, does my theory on VR potentially following the PC adoption curve of the ’80s resonate with you or am I over-fitting?

SN: I think the way even Mark described it, one of the things during the pandemic we learned a lot is what I’ll say the first line worker scenarios which is, if I look at even what happened with HoloLens, in the enterprise, a lot of the use cases around remote training or remote field service, all of those things became pretty mainstream use cases. Similarly, right now the use cases around knowledge work I think can absolutely happen in the VR headset. In fact, when Mark and I first talked, he talked about, “Hey, I want my Outlook to be there in Quest so that I can do both my heads up email work and be in meetings,” and do all the other things that he wants to do. So I do fundamentally think that thinking of this as another form factor of personal computing where knowledge work happens and with mixed reality, even frontline work happens, the pass through piece is also pretty exciting. I think that some of these things will start taking hold and some of the apps we built for HoloLens we want to bring to all platforms like Quest.

Is this your Virtual Reality strategy going forward? I know you have HoloLens, and you’re talking about different partnerships, but there’s no one out there that’s investing to the level that Meta is. You have built the HoloLens, but I don’t think you’ve invested $10 billion in the last year on it. Is it fair to say that this is the center of your approach and, yes you’ll support other stuff, but you’re going to start here going forward?

SN: We always look to who is investing and most importantly who has the highest volume install base of any form factor, and we want to make sure our software is really first-class in it. So that’s why the way Mark described it is, he wants us to be a great ISV [Independent Software Vendor] on his platform and we want to be a great ISV on his platform.

I think I’ve always felt that unless and until you are doing leading-edge work, Ben, and I think you’ve heard me say this, I grew up in a Microsoft where Office was born on the Mac before Windows even was there, so therefore I’m grounded in that lesson. It doesn’t mean we won’t do other things, just like how Mark and the team will also have other enterprise applications and business applications on their platform. To me this is early — I think Quest represents a big new innovation cycle and we want to be there learning, innovating, and making sure our software lights up in great ways.

I’ve never heard you Satya talk so much about being an ISV and “We’re just a mere app maker”. I think the first time I met you, you would describe Teams as the operating system in the cloud.

Mark, when you think about this partnership and a big part of this I think it’s fair to say is building a platform that Meta owns and controls, do you see any tension between these two company visions, or is this a natural thing where you’re more on the device layer, you’re focused on consumer, and you’re fine with Microsoft orthogonally sitting on top of that with their enterprise platform?

MZ: Overall, I think that this is a very natural partnership where I think our strategic interests are pretty aligned, and a lot of the things that we care the most about like the consumer experiences and the sense of expression and the ability to maybe make the version of yourself that you want to express the most, and then be able to hang out with your friends, those are the things that we care about the most and then we just really want to make sure that we have the best work tools on there. So we’ll build our own stuff just because we want to make sure that it exists and we want to push the platform too, but I think work is really Microsoft’s bread and butter, and making sure that hooks up to Azure and that all the enterprises can have this whole 365 suite of tools that Microsoft provides, which I really think to understand the partnership you have to wrap your head around what Microsoft is trying to bring all of that, that full experience for enterprises.

I have a question for Satya on that so you don’t have to list it out.

MZ: I think the fact that it’s not just Teams, but it’s Microsoft 365 and the ability to stream a Windows PC from the cloud and all the Intune work and Azure Active Directory, it’s just the full set of things that Microsoft brings.

But taking a step back, I think in addition to the strategic alignment, I also think that there’s a very deep philosophical alignment on the direction that we want the next generation of computing to go. My brief take on the history here is that in every major computing platform there’s been an open ecosystem that’s focused on partnership and a more closed and integrated ecosystem. So with PC, Windows was the leading open ecosystem and, of course, Mac was the leading closed integrated one. On phones, Android is the leading open one and iOS is the leading closed one.

I think one of the things that’s interesting in the history is that I don’t think it’s predetermined whether the open ecosystem or the closed one ends up being the primary ecosystem. I actually think it quite well could be the case that a partnership-driven and open ecosystem ends up being the primary one that the majority of people and the majority of the value is created on for the next computing platform. So I think our goal in our alignment here is not just about specific use cases, but I think we share a vision of wanting to not only help build the open Metaverse and the open version of the next set of platforms, but also make sure that the open ecosystem wins in this next round of computing.

SN: One thing, Ben, I would add. First of all, I like the way Mark frames the open and integrated, and both have a place. I also feel, at least the lesson I’ve learned, sometimes the hard way perhaps, is I think we overstate all the zero-sum battles that are in our industry. Everybody talks about everything just being one of everything, the reality is there are three or four of everything, take the database market. Sometimes I think companies and leaders in particular miss out big opportunities where you go in and assume everything is zero-sum.

I want to make sure, especially when something new and big is being born, that you are really innovating — the way I think about Quest is, for a Microsoft customer, it’ll feel like a Microsoft device. It’ll be managed by Intune, you’ll log in using AAD, you’ll have Teams, you’ll have Microsoft 365, you can even stream Windows 365. To me, that’s a beautiful way. And, of course, it’ll have all the other stuff that Mark and team and other ISVs build too, but that’s a good world for that platform.

That’s why I was setting you up from the beginning with the Mac reference, because it does seem to align with that vision.

Satya, I know you have to go soon, I have a couple things I just wanted to get your point of view on. You’ve mentioned Teams in VR, Microsoft 365, the Intune Device Management, Azure Active Directory, Windows Streaming, I know those are packaged together, and also the gaming aspect — that was the order they were presented in. Is that also the order that appeals to you strategically? Teams leading the way, but then getting the enterprise bits? Then of course you are making heavy investments in gaming. Is that a ranking from your perspective?

SN: Yeah, for sure. For me, I think the number one experience we want to get right is Teams, because we want to make sure people can have great meetings on Quest. Teams, as you know, is a scaffolding where it’s just not about meetings, it’s about collaboration, it’s about even business process applications that terminate inside of Teams. There’s a lot of stuff that happens inside of Teams — it’s like a browser, essentially — and so we want to get that experience right.

We also want to look at the avatar system that Mark and team are building. How do we really take advantage of what’s the native avatar system? It’s like the native UX on these platforms will be like the native avatar system so you want to be able to, as an application developer, really take advantage of that. Then all the applications, one of the things I’m also keen to see is what’s the use of 2D canvases inside a 3D world look like? I think it’s going to be important to get that right, get the resolution stuff right, so that I can do even heads down work. Lots of stuff to look forward to learn and really refine.

Just to go back to something Mark said a moment ago, if you want access to the enterprise, you need to partner with Microsoft and in this case you have this major investment in a consumer product whose first major partnership is with Microsoft. Does this feel like a validation of having doubled down on being an enterprise company? I know you never actually said that per se, but looking at it from the outside, it’s like, “Look, this is what we are, this is our bread and butter.” Now, because of that focus, this is almost like your way back into the consumer market in this new platform, because you’re perceived as a good partner, you’re not necessarily a competitor for what Meta’s trying to do. Thus it makes sense to partner and then suddenly you have these Microsoft products on these devices and Meta doing a whole bunch of work for you in many respects.

SN: First of all, I think Mark said it well, I think we’re very, very strategically aligned. We clearly are a strong commercial and enterprise company and we want to bring all of that value to Quest, but we also want to bring our gaming assets. We’re definitely all in on gaming and we today have Minecraft on Quest and we have Flight Sim on Quest and we’ll continue to push on that, so we have picked our places where we are a consumer company and a commercial company. Our commercial strength is much bigger, but at the same time having a strong consumer business that probably is something that I want to make sure, especially in gaming, we’re doing some of our best work as these platforms emerge.

MZ: Getting the whole Game Pass catalog available with an Xbox controller and just a big screen in Quest is I think that’s going to be really awesome when it’s available.

SN: It’s actually pretty cool. One of the first demos was with Xbox controller and on a big screen. So in some sense, people are looking at, “What’s the device that I’ll really play a lot of cloud gaming on?” And I’m looking forward to seeing how Quest really works out for that.

Well, I will save the “Common enemies make good friends” question for Mark since I know you have to go. Thank you, Satya, for your time, I really appreciate it.

Nadella leaves the interview.

Meta’s Partnerships

Well, I guess that leads into the question. I mean, as you think about partnering with Microsoft, how much of it is the motivation of avoiding the current system where basically Apple and iOS is this dominant player? They’ve obviously had huge impacts on your business, do you think this partnership happens or happens as early without that?

MZ: I do because this isn’t the only partnership that we’re doing. If you think about all the different use cases that we’re trying to unlock, for work I think that this one makes particular sense. But if you think about some of the other big partnerships that we have, I don’t think that they’re primarily competitive driven, I think that they’re opportunity driven.

Qualcomm is a big one, we’re basically designing and working with them to help shape the direction of the custom silicon that’s going to exist for VR and eventually hopefully Augmented Reality as well. I don’t think that’s primarily driven by a competitive thing, I think that that’s just it’s a big market and we have some silicon design efforts internally and we’ll keep pushing on those. But at the end of the day, Qualcomm is really the powerhouse there, and we figured we’d get more leverage and create more value by working with them, and I’m really happy and grateful for the partnership with Cristiano [Amon] and the team there.

If you look at what we’re doing with smart glasses, EssilorLuxotica is I think clearly the best glasses design and optical manufacturer now with Essilor and distributor of glasses around the world, so they’re a great partner with that and I don’t think that that one’s primarily competitively driven. So I think that there are plenty of other examples here that basically show that what we’re really trying to do is drive innovation, push these markets forward and open up new use cases.

If you want to create a general computing device, you kind of go use case by use case. So with VR, we started with gaming and then the next set of use cases that we worked on were primarily social apps and that’s still kind of in-flight and underway with our efforts with Horizon and all that. But already today the top I think two or three apps in the Quest store are social metaverse apps where people primarily are just hanging out, maybe playing casual games, but they’re social experiences. That I think is sort of a validation of the use cases expanding and of the original strategy that I kind of expected that this was going to be primarily a social platform.

Some of the fitness stuff I think is pretty interesting too, but I think it’s pretty clear that work is really just the next big step function and usage. This is just a huge opportunity and in some ways I think Quest Pro is really the V1 of this effort and it’s an effort that we’re committed to and we’re going to keep pushing on.

I said this to Satya before, you arrived a couple minutes late so you might have missed it, but I’m really biased here because I found the enterprise portions of the keynote far more compelling than the more consumer focused stuff. But again, it’s kind of like I had this preexisting thesis that the enterprise makes sense as the first market for VR, so you get to be the judge. Is this confirmation bias talking? Or do you feel a similar use case is just really tangible for enterprise, particularly from the meeting perspective, and there’s still a little bit of figuring stuff out of what’s the killer app for a consumer?

MZ: I think that the keynote basically breaks down into an update on things that we’ve already launched, new things, and then future things. So there’s no doubt that the work part of the keynote was the new use case that we were trying to talk about. I think that there is an excitement there in the sense that it is the new use case. It is kind of the new, kind of open terrain, it’s kind of the new area that we’re trying to open up and we haven’t talked about a lot of this stuff in depth before.

I do think that the growth that we’re seeing in gaming, the development of the social metaverse platforms, even a lot of the fitness stuff is pretty compelling too, but those are all things that we’d sort of introduced in the past, so we were just updating on. Then I think some of the future stuff is just wild. The neural interface, EMG wristband, showing an actual demo of that working I think is crazy, and I think that that’s going to be pretty amazing and a necessary component for where we eventually are going to go with AR.

You kind of have to imagine that with Augmented Reality, one of the big questions in a general computing platform is what’s the input? With VR, you have the controllers and you have hands, but you’re probably not going to be walking down a sidewalk with augmented reality glasses with your hands up. Then as good as a voice assistant is, there are times where that doesn’t make sense if you want to be more discrete or private or you just don’t want to annoy the people around you. So I do think stuff like the neural interface thing — I think that that’s some of the wildest and most exciting stuff that we’re working on. But I think for someone like you who’s focused more on what products are actually shipping, obviously that’s still a few years off.

Metaverse Dimensions

One of the things that I always thought about the difference between AR and VR is VR is this — it’s kind of in the name — it’s this sort of immersive experience, like you’re going to another place. I always put it in the same sort of line as movies or video games or this idea where when you’re doing it, your attention is all there. Then AR is sort of in the line of the phone, where it’s something that’s with you and accompanying you. That’s one of the reasons why I thought VR in work was interesting is because you go to work, that’s a physical thing that we do. In this case, you can go to work by putting a headset on.

The second dimension is the single-player experience versus multiplayer experience, where if you’re playing a game, that tends to be a single-player experience. If you want to get to the social aspects, that’s more of a multiplayer experience. If you want to walk down the street and check your AR glasses with your neural link or whatever, single-player experience.

I’m curious how you’re thinking about these dimensions, do you have a particular focus right now that’s the key to getting this off the ground? I could imagine it being the dedicated multiplayer experience, just because of technical capabilities and you want to build a social experience, but I’m curious how you think about those different dimensions and what’s possible now, what’s possible in the future and where you want to get to.

MZ: I think general computing platforms need to excel at both, but clearly a lot of the time, you’re just doing stuff by yourself. So the case of whether it’s you put on your Quest Pro or eventually your AR glasses and you just kind of snap your fingers and you have three huge monitors up in front of you no matter where you are, or it doesn’t have to actually be your desk setup, it could be Starbucks or you’re sitting on a bus somewhere, I think that that’s pretty compelling just to have your kind of perfect workstation, which by the way will probably be a better workstation than most people can get physically anyway, but have it not only there, but have it anywhere you go is going to be pretty compelling.

But if it were just single person productivity, that probably wouldn’t be a big enough reason for Meta to be in this space. The kind of philosophical reason why we care about this is about human connection and the fact that delivering this sense of presence with another person is the ultimate expression of the type of services that we’ve been building for eighteen years now. That’s the part that’s exciting to me. I think that there’s a consumer version of that and there’s a work version of that. Meetings also are human connection, but I think both of those are going to end up being important, and for what it’s worth I think both of them are going to be important in both virtual and augmented reality.

There’s nothing that I think would stop us in five years from now from having this podcast and basically just having augmented reality glasses and it’s like if you’re still in Taiwan, basically your hologram is here at Meta headquarters, and it just feels like we’re there. And I can kind of have a 3D model of the device and I can hand it to you — so it’s not just a video call, it’s like we can actually interact with things together, that’ll be compelling. Now, you can start to build those experiences with Mixed Reality with Quest Pro, which I think is exciting.

Some of the things like what Satya was talking about, I think that there will be real interesting use cases, like training or let’s say you kind of have someone and they’re repairing an expensive jet engine or different industrial applications, having the ability to have someone back at headquarters see what they’re seeing and be able to draw or annotate on the world and give them advice on what they’re doing or help train them to do that, that’s actually a very compelling Mixed Reality use case that I think — we didn’t touch on the partnership with Accenture, but we’re working with them and with Microsoft to basically build a lot of the last mile software to actually build out different applications, whether it’s training or specific applications like that.

But the other part of building Mixed Reality with Quest Pro is that developers will start to be able to build the types of Augmented Reality experiences that will exist, both consumer and business that I think will eventually be critical for full glasses when we just get there.

I think that your bit about multiplayer experiences and that it is in Meta’s DNA of connecting people, it’s very compelling. But I guess the question is why Meta? Because when you’re online, the network effect is very real, in part because anyone can use Facebook. It’s free-to-use, it’s one of the beauties of the ad-supported model. In this case, having this sort of experience is gated by acquiring a headset. Now, yes, you had discussions in your keynote about having a better experience on your computer, on your mobile device, but do you think that network effect or that moat that you already have extends into the Metaverse? Or do you have to build it again from scratch and get people on board one by one?

MZ: I think it’s some combination, but I think our business model, like you said, the ad-supported model was hand-in-hand for social networking — if you care about serving as many people as possible, having your product be as affordable as possible is a really important point. So I think that that basic point extends to this generation of computing and hardware as well. Now, we’re not going to be able to make the hardware devices free —

I will say $1,500 was less than I expected.

MZ: I think the business model will be disruptive, in that it’s typically people build hardware and they try to make a profit off of it, where if you’re Apple, you build hardware and you charge as much as you can for it. I do think that having someone come into the space and basically say, “We’re going to build the best hardware in the space and we’re going to basically sell it at a break-even point and in some cases, maybe even slightly at a loss in order to basically help grow the ecosystem with the business model of basically having the revenue come through software and services”, that business strategy I think is aligned with the mission of basically connecting people and having people there because if you want to build a social experience, you have to have the people there.

Quest Pro and Expression

Now that Satya has gone, I can ask you. Do you have any sort of nervousness about Microsoft having their own — there is a certain network effect of being on, whether it be Teams, whatever it might be, that if you can go in and have this full Microsoft experience, that it becomes diluted somewhat and, “Oh, just go get an HTC headset and you get the same sort of experience”? How do you think about managing that? Or do you feel pretty confident that you’re going to be so strong in the consumer spaces, that’s just going to be really easy to flip back and forth and it won’t be an issue?

MZ: I think the pros of the partnership way outweigh the risks. Obviously nothing is risk-free, but at the end of the day, we also have to do our job and deliver world-class services and hardware. If we don’t do that, then obviously we will lose. I do think though that there is this alignment that we talked about before between the things that we primarily care about, which are the aspects of the platform around expression.

And I will say I was extremely impressed by the facial tracking. It was much better than I expected.

MZ: I appreciate that — yeah, I’m proud of that.

It goes back to the example, it’s impossible to appreciate until you actually try it, which is one of the challenges here.

MZ: Yeah, I also think it’s interesting because if you look at what even others in the space are doing, there’s a cost to including those sensors, it adds some depth to the device, it adds cost, so the device is now a little more expensive.

Plus battery life.

MZ: Yeah — battery life, CPU, and GPU. It’s basically processing on the device is limited and now you’re processing basically video streams from multiple more cameras.

It is actually very interesting, because it ties into the work thing, because my one disappointment is I want higher resolution. It’s higher resolution than the Quest 2 was, but it’s still not as high as I want, especially if I want to have my whole computer screens up, and jump in and out of meetings. It’s interesting because that feels like one area where you and Microsoft may not be perfectly aligned. If they were going to build a headset, it’s almost a more single-player experience and it’s going to be all productivity.

MZ: I think basically everyone else in the space would focus more on the the single-player experience. Our bet in this is a deep bet that the connection aspect matters more and this has been part of the experience of running the company all along, is that even just growing up we’re told, “Do your homework, then go play with your friends.” I just think at some level that’s wrong. The connection between people is the point, not the thing that you do after everything else.

Yeah, and that resonates with me. To me, the killer app for VR has been this meeting. I was talking about this on a podcast, the presence thing is real, it’s really meaningful, but it’s so hard to convince people and explain people without trying it.

MZ: We’ll just get there over time. And to your point about the network effect in community, I think part of why making the stuff available on computers and phones and why plugging into Workrooms or Teams or Zoom, which we also announced in the keynote that we’ll support, that’ll work across computers and VR. But I think part of the bet is that more people will be in the ecosystem, and people have the best experience in VR, and AR as well eventually and eventually people will migrate over. But it’s never going to be fully that, these platforms don’t tend to replace the other one completely. Just because we got phones, we still have computers, and our watches haven’t replaced our phones. I do think we’ll be living in a world with a bunch of these things for a while.

What do you see as the biggest risk factors for this vision? There’s almost like a hierarchy. Maybe level zero is the most fundamental question of all, which is will this even be a meaningful product category? Obviously, as you know, there’s lots of skeptics about that. The second is timing. Is this stuff years or decades away? You said in five years we’ll have a hologram sitting across from you. That seems pretty aggressive, but hey, that’d be incredible. The third is the question as to whether Meta is the company to bring this vision to market. Which of these concern you the most as far as is this going to actually work out?

MZ: I think that they’re all risks. I try to think about the things that we can actually control. Is this going to be a big thing? I’m highly confident it will be a big thing. Timing I think is harder to predict but the thing that I feel pretty confident about is that if you look at the other big tech companies, they typically have decades of building out their own operating systems and this kind of computing platform infrastructure, they just have a lot of other technology to bring to bear. Which I think means that if we develop this at the same time as an Apple or a Google or an Amazon, then there are a lot of advantages that they might have. So on timing what that suggests to me is we need to be on the early edge of this, not the late edge or showing up at the same time if we want to help push this forward and really help to shape what the standards are.

For example, if we want the default to be that there are facial expression sensors because we care about having that part of the experience, I don’t think it’s enough for us to show up after people have already set the standard for what this is. I think basically being in there now and helping to shape this is an important part of that, that’s kind of how I think about this.

My impression, and again I was very impressed by the Quest Pro, it feels like the hardware is a fair bit ahead of the software. Do you feel that way as well? Which is maybe counterintuitive to what people would expect.

MZ: I think in some ways you could say that and I think part of the reason is that Quest 2 was just a bigger success than we expected. Basically we built a bunch of versions of VR headsets tethered to a PC and then Quest 1 was really the first one that got to the form factor that we knew we needed to get to, which is what wasn’t as small as we needed to get to, but it got to the de facto —

Standalone status. Yeah.

MZ: And then we come up with Quest 2, and it sells seven times more than Quest 1 or something like that, it was this really big step function. I think part of what happened was the software ecosystem started getting developed around Quest 1, so that by the time that Quest 2 came and it had a better processor, better graphics, somewhat better form factor, the ecosystem was ready for it to take off. On the back of the success of Quest 2, we’ve doubled down in a bunch of other investments, including around things like Horizon and the avatar platform. Because at some level, it’s not worth building a whole kind of social platform on a device that there aren’t a lot of them out in the world. But with Quest 2 it’s like, “Oh, this is happening. Now is the time to go build this.” But of course, those are investments that take multiple years.

Is that part of the whole thing about which operating system are you going to use, which chip are you going to use, was that decision made for you because of the Quest 2 success?

MZ: I don’t know if I’d say that was made for us, but I do think with Quest 2, it was basically time to stop just experimenting and lock on what we thought the long-term architecture was going to be. Which is why before that, there were a lot of different things going on, and we weren’t kind of sure what shape would be the one that got to product-market fit. But with Quest 2, I think it was pretty clear what the basic shape of this was going to be for at least the next five years, what things we needed to do. I think that was a forcing function.

The Meta Brand

Well, the thing that I like about this keynote and your real focus on the partnership and we’re going to be open, is there does seem to be a much greater sense of alignment overall. We’re not going to build our own chip, we’re going to partner with Qualcomm. We’re not going to necessarily build our own operations from scratch, we’re going to leverage Android. All those come with certain sacrifices, particularly in terms of performance and how cutting-edge you can be if you’re not having that deep integration. But it’s like, if we’re going to go ecosystem, we’re going to go all out ecosystem. This Microsoft partnership feels like almost a manifestation of this vision and locks you into it. When did you come to that realization? Like, it’s not going to be fully integrated, we’re not going to do everything? Was that directly linked to, “Okay, we will partner with Microsoft”? Or were those just discrete events that I’m drawing connections that don’t exist?

MZ: A lot of this is just getting the other companies excited about it. Before Quest 2, I think the conversations were very different. A company like Microsoft wouldn’t have wanted to put this many resources on this, because there just weren’t that many units. But I think that now they can see the trajectory of it and they’re like, “Okay, so this isn’t huge yet. There’s millions of units, it’s not billions of units, but it’s kind of going in a direction where this is going to be a major computing platform”. I think similar with Qualcomm, similar with other folks. The terms on which we could work with them, if at all, were not sufficient before we demonstrated the success with Quest 2, which meant that we had to go build a lot of these things ourselves.

What we’ve basically come to — take Qualcomm, for example, this is a partnership that I think is completely unique for Qualcomm in terms of the depth of collaboration that we’re doing together to make sure that we’re designing really great chips. I don’t know if they would’ve been open to that before we demonstrated the success of Quest 2. I do think to some degree we always knew that we wanted the ecosystem to be open, but we also need to do that in a way that gets to the quality level that we need to get to. I just think it was, in some ways, the success of Quest 2 that allowed us to make the partnerships on terms that we’re confident will actually deliver the products that we need to over time, if that makes sense.

I think one of the arguments that can be made as to why Apple has not just gotten away with sort of knee-capping the ad market but been praised for it is it’s a good brand. They have decades of goodwill. They’ve really pushed a narrative about privacy that while I could argue is self-serving, it is also appealing, particularly given the frame they put around it with their ads and things on those lines. The flip side of that argument is that Meta is weaker from a brand perspective. Do you worry that will impact your potential in this space? Or is it just like, “If we just make a great product, it’s going to be fine and will actually refashion the way people think about us”?

MZ: I think it’s somewhere in between the two. I do think that having a track record of decades of building hardware and platforms is helpful. You just have a relationship with that company doing that thing.

That’s a point too, about having operating system expertise. When mobile phones came out, iOS scrolled well, Microsoft phones scrolled well, and Android was janky. It’s like, well, which two have been doing window managers for the last 30 years? You kind of know how it works.

MZ: Yeah, I think that there’s the technical version of that, which I’ve talked about before and then there’s the kind of customer relationship version and brand, which you’re talking about now, which I think is equally if not more important. But I think that those two go together. To me, what this all suggests, I think it is natural that until a person has had one or two of our devices and realize that they’re high quality devices, that they’re going to be reliable, and do everything that the person wants. I just think that’s something that we have to build up over time, and there’s no substitute for both the time but also doing high quality work for building that up. I think the question is, how do you counteract that?

To me, the tradeoff that I think we need to make is be willing to be a little bit early on this, which of course means that the program will cost a little bit more. Because obviously the timing, if you’re trying to get it exactly right to when the market is ready, if we’re spending billions of dollars a year on this, we’re going to be spending billions of dollars more than another company in order to basically be early. But I think we need to do that in order to build up the brand and credibility around these efforts.

That’s kind of my worry though, right? I drew the analogy earlier this year talking about General Magic — bunch of former Apple people, they’re going to make this personal communicator device that ends up looking a lot like an iPhone. But they were in the 90’s, they were a decade too early, and it didn’t work out. How is your confidence about timing right now? Do you have more clarity even than a year ago, the last time we talked, about when this stuff is actually going to be meaningful and useful?

MZ: Again, I think that the Quest 2 success was a good signal that this is, at least for consumers, it is in the ballpark of gaming consoles in terms of the popularity of it. It’s not in the ballpark of mobile phones yet or computers, but at least in the ballpark of gaming consoles, that’s a meaningful first step to break into. I think a lot of these things take a couple of generations. So Quest Pro is the V1 of the work line, I wouldn’t be surprised if Quest Pro 2, when we come out with that —

(laughing)Wow. More breaking news!

MZ: I mean, this isn’t breaking news. Obviously we’re working on the next generations of all these things. I wouldn’t be surprised if that sold significantly more, sort of the relationship that we saw between Quest 2 and Quest 1. But yeah, I think that there’s a difference between being three to five years ahead and being ten to fifteen years ahead.

Right.

MZ: I think ten to fifteen years ahead is really tough. Three to five years ahead is expensive, but I think you’re also developing a lead and a real technological moat. I think that in our case, that actually plays into an advantage here. The other way in which we’re doing this is actually just by focusing on VR at all instead of just Augmented Reality. Because I know there’s this bias that a lot of people in the industry think that Augmented Reality is going to be the real prize —

It goes back to the accompanying technology versus the immersive technology.

MZ: Yeah. First of all, I think VR is probably going to be bigger than people think. I tend to agree, I think AR will end up being — it’s like you’re not going to walk down the street with a VR headset on.

Right.

MZ: Whereas eventually you will with normal-looking AR glasses. But the average American I think still spends more time between their TV and computer than they do on their phone. So if you’re talking about a sit-down experience of computing and looking at bigger screens, VR is probably the closer analogy to that, whereas AR might be closer to a phone. I don’t know, I think that there’s a reasonable shot that people end up ten years from now spending almost as much time in VR headsets as they do with Augmented Reality. I think VR is a big deal, but I also think that VR is sort of a good way to ramp up a lot of the 3D immersive development ecosystem, because a lot of the technologies are the same. The Mixed Reality stuff that you’re going to be able to develop on Quest Pro is going to look a lot like the software that you will eventually have on Augmented Reality glasses and a lot of the basic technologies around the avatar and how you show up in 3D, hand-tracking, all these different technologies are going to be I think very similar. So by focusing on VR now, even before AR is possible, and I think it’s quite likely that AR before you get to the V2/V3 that are really the mature products, that could be later in the decade. But I think VR gives a clear path for developing those technologies in a way that kind of fits the market.

Meta and AI

I have a couple of quick AI questions. First off, I have my own theories, but what’s your view on the role of AI in the Metaverse, particularly as there’s more creative capabilities? Is that going to be an important factor, or is that TBD?

MZ: Oh, yeah, AI is basically touching everything at this point. I think you’re specifically talking about content generation, and I think that will be the case too. I don’t know if you saw this demo that we did where you basically can create a Horizon world by talking to an assistant that’s like, “All right, put two palm trees here. Change the clouds. Actually, I want them to be cumulonimbus clouds. Okay, let’s start playing some music.” and it’s kind of building the whole thing. That stuff is just going to get better and better and better, so there’s that version of that. I think for having AI agents that you interact with, I think eventually you’ll end up having –

Single-player games need NPCs.

MZ: Yeah, the NPCs will get better and better —

That sounds terrifying, but yes.

MZ: No, I think that that’s a real thing.

Yep.

MZ: So I think world-generation, content-generation, customizing your space, I think part of the basic economics of the whole thing is how do you make it really easy for people to create worlds in games?

To me that’s the big thing, because content generation is such a limit on games today. That’s why you have this huge bifurcation between AAA games and indie games that are pixel art. It’s because someone has to draw all that stuff, but if an AI can draw it for you—

MZ: This is a big part of when we think about in Horizon, okay we want to make — obviously there are going to be a set of people, there are going to be studios that have hundreds of people working on them and they want to create AAA content and they’re going to sell it as their own game and it’s going to be its own app and we need to support the APIs to be able to go deep to the metal of the device to be able to really tune that.

I think about Horizon as more the web of the Metaverse as opposed to the App Store is more the kind of standalone apps, the native apps. A lot of what we’re trying to do is just make it as easy as possible to create worlds. So let’s say you create a simple battle royale world, and maybe that takes someone a few weeks right now, but with better AI tools, you can maybe do that in an afternoon and that’s really cool. Okay, and then the next question is, “All right, in order to play a multiplayer game like that, you need to get a bunch of other people around”. So if you have your friends with you, that’s the best case. Second best case is that there’s other people who you don’t know, but they’re respectful and fun and not jerks while you’re playing. And that’s —

Not the norm online.

MZ: Somewhat hard to come by online, but you can create some environments that do that. But the other thing, like you said, is a lot of people just build games with NPCs, but if you’re a person and you’re not an engineer and you’re trying to create a game or a world, you don’t have the ability today to create a world that has NPCs that can play some kind of game, whatever kind of game you want. So I think the ability to actually make it so that a normal creator who has a vision and idea for what they want to do can make something that’s pretty reasonable that includes some AI components, I think that that’s going to be really compelling.

Looking more broadly, one of my theses has been that if Meta can make it through its current challenges, particularly I think with TikTok, then it’s structural advantages in AI, thanks to both capital and data, would give it a leg up, not only in the Metaverse, but also in the creation of entertainment content, whether it be in the feeds and ad-serving as ads shift for being super deterministic to being much more probabilistic. However that thesis has been shaken a bit this summer by — I mean, no one’s doing this that stuff yet, but I think the amount that people are doing — the democratization of AI has been pretty stunning to me. Do you still see that original thesis that AI and your capabilities is going to be a long-term sustainable advantage? Or is it something that everyone’s going to have access to, not just in terms of feed management, but also all the way down to content generation or whatever it might be?

MZ: I think that there will be a mix of this stuff. I think that there’s a bunch of examples where having just a lot of infrastructure and a lot of examples to train the machine learning on is a huge advantage, but I also don’t think that if you’re not one of the big tech companies you won’t be able to do anything interesting at all. I think that people tend to go extremes on this stuff, and that often it just finds a more reasonable balance, but it’s also not necessarily bad for us if the generative tools end up being more democratized, where at some level I don’t really care that much if you create a photo that you want to post on Instagram within the Instagram tool or in some third party app. You just want the good content.

For things like ranking the feed well or recommending the best content or having the best ad system, that’s where we need to be better, but that’s where I think we do both have a track record and have the ability to invest to build technology that is probably better than our other peers in the industry, which I think will give us a sustainable advantage on those things that I think are more core to what we do.

Meta’s Moat

Yeah, that makes sense. This kind of leads to my final question, which is as you sit here today in 2022, what is Meta’s moat generally? Everyone’s always talking about network effects, but then you see this sort of being devalued with the sort of TikTok-ization of everything. Did this tie into the Instagram slow-down or roll-back a little bit? It turns out that those network effects still matter and people missed it? I’m actually curious about that decision specifically, what you’ve been learning from Reels and all those sorts of things, and also big picture how you think about Meta and if that sort of shifted at all or over the last couple years.

MZ: Okay, so there’s a lot in there.

Save the best for last!

MZ: Yeah, let’s see. So in terms of advantages, I think that this stuff typically gets back to values and what the company is focused on. So one is we are deeply focused, I think, on just the sense of connection and I think that leads us to build different products than the other big tech companies that I think tend to be focused more on your relationship with technology. I think that has always taken us in a somewhat different direction than Google or Apple or any of these other companies.

Good job excluding Microsoft.

MZ: Microsoft too! I just think that’s sort of taken us in a different direction than others. Not always good, but I think it makes it more interesting to have companies that focus on different things. The second thing is we are a deep technology company that just invests a ton of not just capital, but where the cultural center of gravity is in the company on building deep technology to be able to build new capabilities. So I think that differentiates us from a lot of the other smaller social companies that I think are frequently building a feature or something like that, but I think the ability to make very transformative changes is somewhat more limited. But then that’s sort of different from your question around Reels, because on that I would say I actually think TikTok — they’ve proven to be a very effective competitor. I think we were somewhat slow to this because it didn’t fit my pattern of a social thing, it felt more like a shorter version of YouTube to me.

Which I think is true.

MZ: Yeah, I mean obviously we have a ton of video on our service. I mean more than 50% of it, or I think it’s at least 50, maybe more, percent of the time that people spend on the Facebook app is watching video at this point. So it’s not like we’re against it, but it’s not the thing that we wake up in the morning and are like, “This is what we are uniquely here to do.” If we don’t do video, then there are other services. If we don’t focus on connecting people, then it’s not clear that there are going to be others who are doing stuff at the scale and the quality that we are, even though there are plenty of other competitors out there.

The thing that I think I sort of missed there though is there’s a different loop around how people interact with discovered content. So before, the way this worked is you had your list of friends that you followed and you got their content and feed and you commented in line and the interaction was there. Now I think that there’s still some of that, but I think it is by and large shifted to you use your feed to discover content, you find things that are interesting, you send them to your friends in messages and you interact there. So in that world, it is actually somewhat less important who produce the content that you’re finding, you just want the best content.

Now what is the best content? It’s still locally relevant and personalized to you. So you’re still going to want to know when your friends who you know have updates in their life, or depending on how much you care about the person, you’re going to want either a higher resolution or lower resolution on those updates. I want to know when my cousin has a baby or something happens in their family, and I want to know with my close friends what they did this afternoon. Those social connections are always going to matter, but by and large, at this point, I think just being able to produce a feed where over time more and more of the content is going to be recommended by AI across the whole corpus that exists. That actually creates social interactions and that’s a loop that’s primarily in messaging now, so we’ve been very focused on that.

There are a couple of big trends that are going on. One is that more and more of the time and content is shifting to video. So I’m going to share the stat around Facebook, which is, it’s not a new stat, but I’ve been talking for a while about how there’s been this shift to video and that continues and we’re now at the point that it’s half or more of the time on Facebook is that. That’s continuing, but the other trend is just that I think we’ve reached a point where the state of AI technology allows you to effectively recommend content from across the whole corpus that is of the same quality and increasingly better than some amount of the Follow Graph stuff that we would’ve shared, that was shown before. The thing for us is I call this vision internally that what we need to produce is a discovery engine, where it’s not just Reels, because video is not the only format. Even as more becomes video, it’s not like photos are going away or text is going away, or news links are going away, or groups are going away, or any of these things. They’re going to still be there. The addition of more content tends to be generally additive to the overall ecosystem, but no one is really building the AI recommendation system or the discovery engine, as we call it, across all of these different types of things and blending them together. I think that that’s a huge area where we can create a lot of value because I think at the end of the day, a lot of the competition that we’re going to have, whether it’s with TikTok or others, is —

Medium specific.

MZ: Yeah, you have a few minutes and you want to discover the best content. Sometimes I want to watch specifically videos, but a lot of the times I just want the best stuff. So I think having the ability to kind of intermix video, photos, text, all these different things, content from friends, content not from friends is going to be a really big advantage, but again, that’s more a question. It’s not a technological advantage as much as it’s an advantage from the way that we look at the world and the problem that we’re trying to solve, which is a different problem than what others are trying to solve.

I appreciate you taking the time. Congratulations on the Quest.

MZ: Happy to and thank you.

Quest Pro, not the Quest Pro 2, that’s the secret news that’s coming, it’s very impressive and look forward to seeing what comes next.

---

This Daily Update Interview is also available as a podcast. To receive it in your podcast player, visit Stratechery.

The Daily 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 supporter, and have a great day!