Produced by | Tiger Sniff Technology Group
Author | Maruto Mountain
Edit | Wang Yipeng
Header | Visual China
Xiaozha still can't forget his metaverse.
On April 23, Beijing time, Meta announced an important decision on its official website: open the operating system of the headset Quest, Horizon OS, to third-party hardware manufacturers.
On the same day, Zuckerberg updated a video on his personal Instagram account, comparing the ecological differences between Apple's Vision Pro and Quest, which roughly means that Apple's VisionOS maintains Apple's consistent closed strategy, while Horizon OS will become the Android system in the XR field, attracting more third-party manufacturers to participate through an open ecosystem.
It's hard to imagine how much Zuckerberg went through before making this decision, because Meta has been pretty much single-handedly propping up the XR industry for more than a decade.
Since its acquisition of Oculus in 2013, Meta has invested nearly $50 billion in Reality Lab, only to receive lengthy media accusations from shareholders and now offer free licenses to other hardware vendors.
From Zuckerberg's point of view, this strategy may be the best solution at the moment: after the release of Apple's Vision Pro, Meta's advantages in hardware and interaction were almost overtaken by the latter overnight. Now that Google is gearing up to create the "XR Alliance", Meta can only rely on the content ecology accumulated over the years, and if it wants to maintain this leading edge, it is bound to attract more third-party hardware platforms and content creators.
It's just that Meta's dilemma in the XR industry today doesn't seem to be something that an open source system can solve.
There are wolves in front and tigers in the back
"I was wrong and I'm willing to take responsibility for it".
In November last year, after Meta released its 2023 Q3 financial report, Zuckerberg issued a rare "guilty edict" within the company and started a round of layoffs of 10,000 people.
And the mistake he admits is the unrestrained investment in the XR business unit in the past decade, as well as the unreasonable tilt in the allocation of resources within the company, which led to Meta "getting up early in the morning and catching up late" under the AI tuyere last year.
An easily overlooked fact is that Meta was much ahead of star companies such as Open AI and Stability AI in the layout of large models. As early as August 2022, Meta released Blenderbot 3, a large language model similar to ChatGPT, and a month later released Make-A-Video, which may be the earliest Wensheng video model, but the two barely made a splash.
All of this didn't improve until the release of Llama 2, and Meta's open-source strategy allowed Llama to evolve rapidly in a short period of time, and also made the industry re-examine the technical prowess of Meta, a long-established AI company.
In the early morning of April 19, Meta released the new Llama 3 series model, and just four days later, this model has spawned as many as 1,013 Llama variants in Hugging Face's open-source library.
It stands to reason that it is somewhat inappropriate to get such an enthusiastic response in the field of AI, and then mention the ethereal "metaverse", not to mention that just 5 months ago, Zuckerberg vowed to investors to rein in their investment in the XR business.
But an extremely real problem is that if Meta does not build an open source ecosystem for the XR industry now, then the company's real money investment in the past may really be at risk of "going down the drain".
The conundrum ahead is Apple's Vision Pro, which was unveiled at WWDC last year.
Image source: Visual China
In fact, Apple's Vision Pro and Meta's headset Quest, whether in terms of pricing or application scenarios, have no direct competition.
But the biggest blow to Meta is that Apple has deprived Meta of the right to define the XR industry with just one product. At the time of the release of the Vision Pro, Apple introduced the concept of "spatial computing", which subverted the industry's perception of VR/MR (mixed reality) products in the past, and at the same time, it also directly pulled the entertainment attributes of XR products to the dimension of productivity tools.
In addition, Vision Pro also shows the story of how a hardware manufacturer reduces the dimensionality of terminal products to attack Internet manufacturers. For example, in terms of function implementation, Apple uses the combination of M2 chip and real-time sensor processing chip R1 to reduce the delay between the display and the sensor to 12 milliseconds, which greatly improves the realism of the virtual picture.
For Meta, even if it invests another $50 billion, it is impossible to develop M-series chips on its own like Apple.
And most importantly, Apple's unparalleled appeal in the field of consumer electronics. In June last year, a week before the release of the Vision Pro, Meta suddenly officially announced a new generation of flagship product Quest 3, and its intention to "cut off" Apple could not be more obvious, but the discussion in the industry is still focused on the Vision Pro, which has not yet been unveiled.
This is clearly a red flag for Meta. While the Vision Pro's high price tag makes it doomed to become a mass consumer product like the Quest headset, how will Meta respond if Apple launches an affordable headset in the future?
If the threat from Apple makes Meta nervous, then Google's intention to "pick the fruit" definitely challenges Meta's bottom line.
At the beginning of March, foreign media The Information broke the news: Google had a conversation with Meta's senior management, the former hoped that Meta's headset would be replaced with Google's upcoming Android XR system customized for XR, Google tried to cover the Android system in the XR field in this way, Meta naturally refused.
Wooing Meta didn't work, and Google turned its attention to other hardware manufacturers, touting the Android XR system to manufacturers including Samsung, intending to establish an "XR alliance" that excluded Meta.
Google's move made Meta completely break its defenses. Andrew Bosworth, the company's CTO, publicly accused him on social media, "Google hasn't focused on XR for so many years, and now they're starting to sell Android XR to their partners and imply that we're the ones who destroy the ecosystem, but in fact they are." ”
Under the attack of wolves in front and tigers in the back, Meta's XR business, which it has painstakingly operated in the past, can be described as embattled.
开源Horizon OS,能搅起多大风浪？
Since the release of Llama 2, Meta's demonstrated "general infrastructure capabilities" have quickly put Meta in the top tier of this AI arms race, and the company's strategy of using generative AI technology to improve return on ad spend has also regained the favor of advertisers.
In terms of performance, in Q4 2023, Meta achieved revenue of US$40.111 billion, a year-on-year increase of 25%, and a net profit of US$14.017 billion, a year-on-year increase of 201%. Since the beginning of the year, Meta's market capitalization has increased by 40%.
In the context of such a quiet and good advertising business, wouldn't it be good to directly cut off the XR business that has been losing money all year round?
The answer is no.
In addition to the marginal "metaverse" strategy, the biggest significance of the Quest series headset is that Meta has a hardware entrance for the first time as an Internet company.
In the era of mobile Internet, Apple only needs to modify privacy protection settings at its fingertips, which can make Meta lose tens of billions of dollars a year, which is the advantage of mastering the hardware entrance. Although the XR industry is currently completely incomparable to smartphones in terms of the number of users, if the XR industry can usher in a round of technological revolution in the future, its potential market is still considerable.
That's why Zuckerberg is holding on to the XR business.
In the face of today's threats, Zuckerberg's choice of open-source Horizon OS is indeed a smart strategy.
First of all, Meta has a rich technology accumulation in the XR industry, which can provide great help to third-party hardware manufacturers. For example, in the video on the 23rd, Zuckerberg mentioned that the new Horizon OS will provide four interaction methods: hand, eye, face, and body tracking.
Image source: Youtube
It should be noted that the current interaction mode of the Quest series products is still based on the traditional handle interaction mode, but Zuckerberg's statement undoubtedly gives a dose of peace of mind to potential third-party hardware platforms: you can do whatever you want in terms of interaction, and Meta can provide support at the system level.
Relative to Google's Android XR system, if Horzion OS can solve the underlying technical problems, then there is no doubt that other hardware platforms will choose to "vote with their feet".
Secondly, in the face of Apple's huge advantage in hardware, Meta is basically impossible to overtake, not to mention that investors will most likely not allow Reality Lab's R&D budget to increase again. Then simply give full play to the existing advantages - a relatively rich content ecology.
According to statistics from Wellsenn XR, as of 2023, there are 559 content apps on the Meta Quest content platform, and 5,247 content apps on the Quest-based third-party platform, SideQuest. For comparison, Pico has the second largest content app in the global market with 516 apps.
On Horizon OS, Meta will allow Steam and Xbox cloud games to run on its operating system, which will further enrich Meta's content ecosystem.
It is worth mentioning that in the past, Meta mostly followed the strategy of console game manufacturers in XR content, that is, relying on "buying, buying, buying" to obtain exclusive supply advantages, but after the release of the new Horizon OS, Meta's open-source ecological strategy is bound to reduce this part of the expenditure.
It is foreseeable that after Meta releases the open source Horizon OS, although it will lose some users who were able to be deeply bound in the past, in the long run, this strategy is likely to be able to stifle the offensive of Apple and Google.
Of course, the release of the open source Horizon OS doesn't mean that everything is going well, and there is no escaping the fact that the market demand in the XR industry is not growing steadily. According to IDC statistics, global VR/AR headset shipments will decline by 23.5% year-on-year in 2023.
Although the release of star products such as Vision Pro has brought the XR industry a high degree of attention, it still faces big problems in attracting new users or retaining them. Taking Vision Pro as an example, on April 8, Ming-Chi Kuo, a well-known analyst in the industry, lowered the annual shipment target for this product to 400,000 to 450,000 units. It is worth mentioning that this is the second downward revision of sales expectations since the release.
The reason for this is the two problems that have been repeatedly mentioned: the lack of killer applications, and the lack of a fundamental solution to the problems of field of view and motion sickness. Subject to the current development status of the industry, these problems are not something that Meta can solve in a short period of time.
In this context, if the open source Horizon OS only allows Meta to gain a firm foothold in the stock market through ecosystem-based vertical and horizontal integration, its future impact on the industry still needs to be marked with a question mark.
People who are changing and want to change the world are all in the Tiger Sniff APP