Decoding Tech's Next Big Thing: VR vs. AI
In the fast-paced world of technology, sentiments can change with astonishing speed. Investors and entrepreneurs aggressively invest or build while simultaneously evaluating whether the “new, new thing” will be the promised next platform.
It’s becoming an increasingly consensus view that AI IS that “new, new thing.” It is comparable in importance to the birth of the Internet and mobile, but it’s also different. AI is different in that it is a sustaining technology, mostly benefiting incumbents at the top and bottom of the technology value chain (applications and infrastructure).
It’s not to say there won’t be any startups that benefit from AI, as there will be new software toolkits for developers, data, and product teams. However, as a sustaining technology, we believe it is designed to funnel value to semiconductor companies, hyper-scale cloud vendors, foundational models, and incumbent dominant software applications.
VR / AR used to be the “new, new thing,” but the excitement has waned due to some practical current limitations of the technology. Facebook has invested $20bn+ into the Oculus platform. Magic Leap infamously burned through $3.5bn in capital and has little to show for it. Now Apple has entered into the fray with Vision Pro, but the production forecast of ~450k units available in the first year is a big disappointment.
That historical setup doesn’t make VR sound very exciting, but it’s worth paying attention to as an investor or entrepreneur because of one sharp contrast to AI. VR represents a disruptive technology, a new place or platform where people spend time, which levels the playing field for startups or even puts them at an advantage relative to incumbents.
“Where people spend time” is an important distinction. With AI, the technology will be largely embedded into existing applications. With VR, existing and new applications will need to be reconfigured to be natively optimized for the new platform.
For better or for worse, successful platforms are the launching pad for wonderful and addictive applications. Some applications can even gain so much traction that they become a platform in themselves, creating an interesting tension with the original platform.
Therein lies the opportunity. To build the most consequential company in AI, you’ll have to enter an arena of semiconductor design and hyper-scale infrastructure, where incumbents have been building advantages for decades. To build the most consequential company in VR, you just need to build the most popular application, but everyone will be starting at zero.
But will the Vision Pro be worth building on?
How to Evaluate Vision Pro? Lessons from iPhone & iPad
One way to look at new devices is to categorize them as either “Time Theft” devices, stealing time from other legacy devices or platforms, or as “Time Expansion” devices, consuming users’ time spent on other activities. The iPhone is the best and most recent example of a Time Expansion device. The next time you’re walking around outside, start noticing how many people are preternaturally staring at their phones while navigating the sidewalk.
For us as investors, thinking about technologies along this spectrum yields valuable insight. We want to gravitate towards devices at the Time Expansion end of the spectrum because they become “must haves” as opposed to Time Theft devices which tend to be seen by users as “nice-to-haves.” Time Theft devices lack sufficient distinct, compelling use cases, whereas Time Expansion devices can actually capture enough user time to effectively generate substantial equity value at the app ecosystem layer.
Part of our investment philosophy at Flex involves seeking categories with lower competitive intensity. Looking at devices using the same lens, this is another reason we prefer the Time Expansion end of the spectrum. It’s easier to win user engagement when your device is commandeering unconquered time vs. stealing from other devices. In this light, it’s therefore no surprise that the platforms that create an abundance of unicorn applications are those that are able to command net new screen time.
The Time Expansion vs. Time Theft framework is a spectrum, and platforms or devices can move along the spectrum over time. As we described above, the iPhone initially launched as a Time Expansion device, evidenced by the fact that killer applications for the iPhone, aside from messaging, all leveraged the camera or GPS. As technology improved throughout the years, the iPhone eventually consolidated a number of legacy devices into one: camera, portable media players, handheld gaming devices, and even to some extent, PCs and TVs. Its role as a Time Expansion device is still indisputable, but its dominance as a Time Theft device has given it additional leverage as an ecosystem.
The iPad entered the market at the opposite side of the spectrum (Time Theft) and has stayed there, never achieving a transition to Time Expansion. However, even at the Time Theft end of the spectrum, the iPad doesn’t steal enough usage time from other devices and generally still necessitates the continued ownership of other incumbent devices (e.g., few of us have given up a laptop in favor of an iPad while many of us have given up a camera in favor of an iPhone). Data from the Pew Research Center reveals that only 7% of adults claim a tablet as their sole device.
When we compare the iPhone and iPad sales, we can see how this dynamic played out. Despite the iPad's robust initial sales, in the longer term, it has only become a business roughly 30% the size of the iPhone.
Source: Statista, BusinessOfApps
Of course, if the Vision Pro approaches the iPad and Apple Watch’s units sold, Apple and Wall Street would consider it to be a screaming success. Yet, as Time Theft devices, the Apple Watch and iPad haven’t created as much equity value for their app ecosystem, lacking the iPhone hits like Instagram, Snapchat, TikTok, Uber, Doordash, etc.
While, with hindsight, the iPhone is the poster child for platform success, it’s worth noting that it wasn’t an immediate slam dunk. It took over five years for the iPhone to truly establish itself, and undoubtedly carrier subsidies played an instrumental role in its early distribution, accounting for an estimated 80-90% of sales in the initial years post-launch.
Source: VentureBeat, AndroidAuthority
Apple sold 1.2 million units of the inaugural iPhone 2G that was unveiled in 2007. Only a year later, with the introduction of carrier subsidies and the iPhone 3G, the iPhone’s growth trajectory skyrocketed. These subsidized devices were initially priced at $199, which equates to about $280 today when adjusted for inflation. In stark contrast, Apple's Vision Pro carries a price tag twelve times that at its debut, with no obvious subsidies in sight.
Will Vision Pro be worth building on?
With this Time Theft/Time Expansion framework in mind and lessons from how that has played out for the iPhone vs. the iPad, we pose two overarching questions to help us consider whether Vision Pro is the type of platform that will foster a rich ecosystem of wildly successful applications:
- How successful will Vision Pro be?
- What killer feature of the Vision Pro will be the answer to “Why now?”
How successful will Vision Pro be?
Much of this hinges on use cases and how people adopt the device in different settings. Will people be using it mostly in work settings, in which case it could steal or even replace laptops and desktops? Is it a home entertainment device, stealing time from or even replacing TVs and iPads? Or will people be walking around in the real world with the devices on, in a state of augmented reality, thus replacing the phone and expanding screen time into the remaining unconquered time?
Apple's long-term vision with Vision Pro might be all of the above. While it might merely be stealing time from other devices at inception, Vision Pro is potentially designed to eventually replace them all - and more.
While it’s fun to imagine what could transpire that far out into the future, our contention is that Vision Pro will start its life on the Time Theft end of the spectrum and as a result is unlikely to be an immediate hit as a consumption device. This is not exactly encouraging for its initial launch prospects. Adding to the challenge is its hefty price tag, which more closely aligns with the launch of the original Mac in 1984. At an inflation-adjusted price of $7,000, the first Mac could only boast one million units sold in its first three years due to a high price point and a mass market that wasn’t yet ready to adopt a new and disruptive device. Perhaps more significantly, the initial lack of software applications played an even bigger role in holding back mass adoption.
When trying to forecast early adoption, it’s also worth noting that supply chain limitations, tied to Sony's current OLED production capacity, will initially restrict Vision Pro sales to an estimated 450,000 units per year.
All that being said, what makes Vision Pro such an interesting opportunity is that as the technology improves, we could see it successfully moving over to the Time Expansion end of the spectrum and joining the iPhone as a true ecosystem capable of supporting its own application layer.
Therefore, despite our view that Vision Pro is not going to be a home run out of the gate given its price tag and initial low availability, it is certainly a device worth paying attention to. A successful Vision Pro that makes the transition into its own platform and app ecosystem could end up creating more net new unicorns than AI.
We need to stay cognizant of not only the endgame (Vision Pro potentially becoming a dominant platform) but also the timing of where we are on the adoption curve so that we are ready when opportunities arise.
What killer feature of the Vision Pro will be the answer to “Why now?”
One major opportunity is linked to the device's outward-facing cameras. It's concerning that Apple seems to have indicated it might limit developers' access, citing privacy concerns, but it also wouldn’t be the first time Apple adopted this strategy: while Siri arrived with iOS 5 in 2011, third-party devs didn’t get access until iOS 10 in 2016.
Overlooking Apple’s potential restrictions, we can’t help but look at the creator economy as the natural beneficiary of this feature. Creators have grown tremendously over the past few years, with research from Adobe estimating that roughly one in every four people define themselves as a content creator, and ~86 million creators in the US alone. While not all of them are influencers, and roughly six out of ten creators still work full-time, Goldman Sachs estimates that the creator economy could approach half a trillion dollars in value by 2027.
VR might represent the forthcoming evolutionary leap in enabling a wide range of creative talent to deliver more immersive experiences to their audience. That said, the present-day production of spatial content carries a hefty price tag. Well-produced VR content can cost up to $10,000 per minute to create, often requiring a wide array of professional equipment and post-production work.
If Vision Pro’s outward-facing cameras can be unlocked for content creation, its $3,499 asking price would suddenly appear trivial. If successful, the device could herald a new era for content creators supported by a large app ecosystem for production and distribution, paving the way for significant opportunities in equity value generation.
In this way, Vision Pro’s larger opportunity may not be as a consumption device but rather a creation device to unlock millions of “Being John Malkovich” experiences.
Apple will surely sell out of the supply-constrained headsets, which is a stepping stone to driving supply chain maturity and economies of scale. While Vision Pro goes through its growing pains, we remain cautiously optimistic for a cheaper and slicker form factor to drive more consumer adoption. In the meantime, if we make a bet, it is likely to be around the content creation possibilities that we think could drive real value in an ecosystem.