The Pico virtual reality project under ByteDance is currently undergoing the largest adjustment in its history.
On November 6, Henry Zhou, founder of Pico, took merely ten minutes to announce the fate of the company: the Pico OS team is to be integrated into ByteDance’s product development and engineering architecture platform, while the entity’s marketing, gaming, and video departments will undergo significant layoffs.
According to information gathered from various sources by 36Kr, Pico had around 1,600 employees. About 400 employees are being shifted from its OS department, while nearly 400 of the 1,000 employees from the marketing, gaming, and video sections are facing layoffs. Only around 200 employees remain in the hardware department. Overall, the proportion of layoffs and transfers stands at about half.
Prior to this development, employees had sensed an impending change, prompting some to organize small farewell gatherings toward the end of October. Pico was poignantly described as “a Titanic slowly sinking” during these gatherings, reflecting the somber mood among its workforce.
When ByteDance acquired Pico, there were high expectations that its success in the VR hardware business would pave the way for a groundbreaking content ecosystem, akin to finding the “next iPhone 4” from the ground up. While Meta’s VR business set the American standard, ByteDance’s acquisition of Pico was initially considered a beacon of hope for the Chinese VR industry.
However, the trajectory has taken an unexpected turn. According to a senior executive interviewed by 36Kr, ByteDance invested nearly RMB 9 billion (USD 1.27 billion) to acquire Pico when its valuation was around RMB 3 billion (USD 423.98 million). At its peak, ByteDance injected an additional RMB 10 billion (USD 1.41 billion) into Pico in 2022. Conservative estimates now value ByteDance’s total investment in Pico at just over RMB 20 billion (USD 2.82 billion), making it the most significant investment in the domestic VR industry.
As the saying goes, all good things come to an end. Persistent rumors within Pico speculated that “ByteDance’s patience with Pico will only last three years,” and now, this speculation has turned into reality. From the substantial investment to the current retreat, ByteDance’s enthusiasm for VR seems to have waned, lasting two years and two months.
Signs of ByteDance’s changing stance were evident towards the end of 2022. Pico proposed a RMB 40 billion (USD 5.65 billion) investment plan over the next three years, a proposal that was significantly scaled down by ByteDance’s top management. Although a Pico spokesperson refuted claims of the reduction, citing false information with no disclosure of specifics, the pressure from higher-ups had a cascading effect.
Yang Zhenyuan, a senior leader of Pico’s business and a vice president at ByteDance, had previously regarded Pico on par with the Volcano Engine in his objective and key results (OKRs). However, this year, Pico was notably absent from his OKRs, signaling a shift in priorities.
This change in top management attitude directly impacted product development, causing substantial delays in the launch of Pico 5 and 6. Internal sources revealed that Pico 5, a major VR product, was initially scheduled for release in the second quarter of this year but faced a complete overhaul after the launch of Meta Quest 3.
Chen Ting, a Pico employee, attributed the delays not only to Meta monopolizing Qualcomm Snapdragon chips for half a year but also to certain aspects of Pico 5’s hardware development falling short of ByteDance’s top management expectations. Consequently, the release of Pico 5 might be postponed until 2026, as confirmed by a Pico spokesperson.
The challenges extended to the preparation for Pico 6, with core components being rejected during the application process. ByteDance went to the extent of compensating millions of dollars to suppliers who had already begun manufacturing, opting to deviate from the original plan.
These setbacks were not isolated incidents, as revealed by another Pico employee who disclosed severe budget cuts across various projects within the company. Financial allocations, which were in the hundreds of millions of RMB last year, have dwindled to just a few million this year.
For a trailblazer that once pushed the industry’s boundaries, Pico’s withdrawal invokes more regret than satisfaction among other players in the VR industry. As Li Jie, vice president of VR startup Xiaopai, aptly puts it, “It’s not easy being a leading figure in the industry… If you don’t agree, try starting your own company before critiquing others.”
Fall from favor highlights mismatch of Meta’s strategy in China’s VR industry
As the “pioneer” of the VR industry, Pico’s first step was to heavily invest in gaming resources.
The VR industry faced a daunting challenge: a dearth of content leading to stagnation in VR shipments. The vicious cycle ensued—limited VR content resulted in developers struggling to monetize the platform, creating a smaller content pool and discouraging user purchases.
Meta was the only company that adequately bridged this gap, notably with the Quest 2. It broke through by aggressively acquiring VR gaming content and employing pricing strategies that made content affordable. The results were significant, propelling Meta to push VR shipments into the tens of millions—an unprecedented achievement in the decade-long history of the VR industry.
Naturally, Meta became the benchmark for China’s content giants, with Pico closely following suit and making substantial investments. From its acquisition by ByteDance in September 2021 until June 2022, Pico mirrored Meta’s approach by prioritizing gaming content. To address the content gap, Pico consistently introduced imported games and developed in-house titles.
The pace of game additions on Pico was notable, starting with approximately 200 games post-acquisition—half of Meta Quest’s offerings. However, Pico swiftly doubled its game library, introducing 2–4 new games weekly, aligning with Meta’s catalog.
Huang Wei, an employee in Pico’s gaming department, said that “Pico has imported almost 60–70% of the games on Meta,” emphasizing the competitive nature by incorporating this comparison into its objective and key results (OKRs).
Games were categorized into four levels: S, A, B, and C, with Pico willing to pay substantial entry fees for S-level games. According to Huang, beyond introducing new games, Pico invested generously in gaming operations and organized detailed activities to collect feedback and enhance user engagement.
Pico’s investment in new game titles amounted to hundreds of millions of RMB, paralleled by ByteDance’s strategy to imitate Meta through extensive marketing and aggressive price subsidies for hardware devices.
Pricing-wise, Pico mirrored Meta’s Quest 2, launching Pico Neo 3 and Pico 4 at a loss in the RMB 2,000 (USD 282) price segment, initiating a price war in a market where VR devices typically sold for RMB 3,000–4,000 (USD 420–560).
In marketing, ByteDance leveraged its entire ecosystem, saturating Douyin, the Chinese equivalent of short video platform TikTok, with Pico advertisements and influencer endorsements. Ambassadors Leah Dou, a singer-songwriter, and Sun Yingsha, the world’s top table tennis player, adorned high-traffic areas in Beijing and Shanghai.
According to an individual from the mobile industry, the marketing expenses for Pico 4 rivaled those of mobile manufacturers that ship millions of devices, signifying ByteDance’s commitment to Pico’s success.
Excitement buzzed within Pico in 2022, but ByteDance’s top management closely monitored the situation. Sales and weekly user activities were scrutinized frequently, with high expectations set for Pico 4—aiming for 2 million units sold annually and a 50% weekly user engagement rate.
However, Pico soon realized that blindly emulating Meta’s approach wasn’t effective in China. Chinese gamers, accustomed to free mobile gaming options with in-app purchases, posed a challenge to the paid gaming model adopted by Meta. Users on Pico preferred playing pirated games for free, spending only an average of RMB 40 (USD 5.6) per game, a third of Meta’s average.
Attempts to transpose the mobile gaming model to VR, exemplified by iQiyi Smart’s failed content aggregation approach, highlighted the challenges. Pico faced difficulties everywhere in adopting this path, with the VR gaming development ecosystem primarily in Europe and the US, limiting the negotiation power of Chinese companies.
Pico’s failure to grasp the intricacies of the bottom-level content model rendered its high marketing and discounts ineffective. The marketing conversion rate stood at a mere 2%, with Pico 4’s marketing cost per unit reaching RMB 1,300 (USD 183), despite being sold for RMB 2,500 (USD 352).
Despite Chinese VR companies looking up to Meta as a role model, there is an inherent ambivalence, with Meta being simultaneously viewed as a nemesis. Within Pico, employees express their ambition with phrases like “battling Meta and trampling on Apple” and “taking two years to fight against Zuckerberg,” underscoring their fervor and determination.
Ultimately, Pico’s decision to withdraw serves as a poignant reminder that merely replicating the Meta model carries limitations and does not guarantee success. The complexities of the VR landscape, coupled with the unique dynamics of the Chinese market, demand a more nuanced and strategic approach beyond emulation.
A collision of VR strengths and applications
ByteDance is known for its pattern of swiftly entering and exiting new ventures. How the Chinese tech giant has dealt with Pico, its virtual reality business, is testimony of this dynamic.
According to several employees from Pico’s gaming department, in July and August 2022, when realization dawned that gaming wasn’t a viable path, ByteDance imposed strict control on the budget of the department. This even extended to denying approval for server costs related to introducing overseas games and the budget earmarked for game operations.
Simultaneously, Pico quickly pivoted, redirecting its focus to video content and livestreaming—domains where ByteDance has excelled. Spearheading this strategic shift was Ren Lifeng, president of Xigua Video and a prominent figure at TikTok, who is deemed a “hero” and a “firefighter” in ByteDance’s hierarchy. A mid-level employee told 36Kr that Ren once even received a 100-month year-end bonus.
The immersive experience of wearing a VR headset, likened to “sitting in a living room,” led Pico to explore video consumption as a potential application scenario. However, this shift toward content did not resonate with the majority of Pico’s experienced VR industry employees, who questioned the viability of VR videos based on their experience.
An illustrative precedent is Meta’s Oculus Go, a VR video device, which failed to gain traction in the market, selling only 250,000 units in a year, well below the anticipated one million units. Despite being a leader in the industry, Meta abandoned the venture.
“Ren Lifeng comes from the video sector. Him coming to Pico is like ‘carrying a hammer, seeing everything as a nail.’ Short videos and livestreaming exploded on phones, so ByteDance also wanted to gamble on whether a miracle in VR was possible,” said Huang Wei, an employee in Pico’s gaming department.
Pico’s video team expanded rapidly, adopting an ambitious strategy that involved significant spending to build a VR video ecosystem. This included a hefty RMB 1 billion (USD 140.5 million) investment in acquiring broadcasting rights for the Qatar World Cup to be broadcast on Pico. Concerts featuring heavyweight stars like Zheng Jun and Wang Feng were also paid for, along with collaborations with popular titles like “The Three-Body Problem” to plan various 2D and 3D programs.
However, reality unfolded as VR hardware struggled to provide an exceptional video experience.
“Why would I use VR to watch what I can watch on a projector, TV, tablet, or at a movie theater? The increment of experience that VR brings to video is immersion, but it comes at the cost of strapping a heavy half-kilogram lump on your head,” said a Pico employee. His Pico 4 has been gathering dust for over six months.
VR hardware faces constraints in optics, display, and battery capabilities, limiting mobility and portability. The nature of short video consumption, fragmented and scattered throughout the day, poses a natural contradiction to VR’s immersive experience.
For long videos, Huang told 36Kr that the production cost of VR videos is exorbitant, involving issues like network transmission, 5G penetration, and more. These challenges are beyond Pico’s sole capacity to manage.
Currently, the content that aligns with VR is primarily adult films that require an immersive experience. While this industry is mature and profitable in other countries, it operates in a gray area in China. Pico gradually shifted its content toward obscure “borderline” live shows, departing from its initial focus on mainstream video content. An employee from Pico’s hardware department shared a surprising experience while using the Pico 4 to watch the Qatar World Cup. To his amazement, the Pico VR’s livestreaming experience underwent a notable transformation after 2 a.m., featuring a shift toward content of “girls gyrating.”
Despite a long period of validation, livestreaming did show results for Pico in terms of usage time and revenue, constituting 30% of Pico’s usage time, with rewards almost matching gaming income. However, these results failed to translate into the sales that ByteDance’s high-level management prioritized.
An examination of backend user data revealed that Pico’s actual sales in 2022 were less than 600,000 units, with the highly anticipated Pico 4 falling short of 300,000 units. Daily active users were less than 100,000, significantly below ByteDance’s expectations for Pico. A spokesperson from Pico rejected the insights, labeling them as inaccurate. However, no specific data was provided to substantiate the refutation.
Li Jie, vice president of VR startup Xiaopai, offered a pragmatic perspective. “VR and AR are still in their early ‘brick phone’ stage compared to smartphones, following the objective laws of development… don’t fantasize about miraculous overnight success,” Li said.
A VR company divided into factions
The VR industry, in its perpetual pursuit of innovation, grapples with uncertainties marked by constant exploration and iteration. Amid strategic shifts, inevitable alterations in personnel and team configurations ensue.
The term “factional integration” has become a common refrain among Pico employees who engaged with 36Kr. Multiple internal sources hint at the existence of two factions within Pico: the “Pico faction,” comprising original employees, and the “Byte faction,” consisting of those transferred from ByteDance to Pico.
As Pico transitions its content focus from games to live video, the influence of senior executives from the “Byte faction” has surged, posing challenges for “Pico faction” employees within the evolving evaluation system. This shift has resulted in a significant exodus throughout 2022.
The disparities in development stages and business characteristics give rise to mutual misunderstandings between the factions. Pico, initially a hardware startup, favored a flat structure, leading the “Pico faction” to lean toward faster, simpler approaches. Conversely, ByteDance, a massive company, prioritized process safety over speed.
Employees within the “Pico faction” lament the complexity introduced post-acquisition, with complaints about extensive audits, document writing, inter-departmental communication, and blame games. The toll on wellbeing is exemplified by a reported case of a “Pico faction” employee requiring emergency treatment due to elevated blood pressure after prolonged hours and stress.
On the other hand, criticism from the “Byte faction” points to perceived crudeness within the “Pico faction,” emphasizing the need for precision and correctness in decision-making. “What’s the daily active users of your original Pico product? My Byte product has tens of millions of daily active users. I don’t need a quick response, but I have to ensure that every decision is correct, rethinking and demonstrating it repeatedly,” said Li Ming, an employee in the “Byte faction”.
This clash of cultures reflects the differing pace preferences between the two factions.
The ascendancy of managers from the “Byte faction”, rooted in an internet background, brings new challenges. As a hardware company, Pico oversaw a steady pace of iterations, but with new managers taking over, this momentum was eventually lost.
Chen Yong, an employee from Pico’s hardware department, said that ByteDance was almost tailing Meta in its footsteps, even timing product releases specifically before the Meta Quest Pro.
According to Chen, the loss of confidence among Pico’s upper management was potentially triggered by the rushed release of Pico 4, which has been perceived as a contributory factor of its failure.
On one hand, Pico 4 was launched shortly after a price war with Pico Neo 3, the latter, boasting better sales, had already fulfilled a certain consumption potential and piqued curiosity among buyers due to the extended replacement cycle typical in the VR market.
On the other hand, the swift release of Pico 4 resulted in a myriad of issues, including streaming problems, clarity issues, battery life concerns, and software update complications. According to a mid-level Pico employee, these challenges contributed to a high return rate of 30% from sales channels.
“Executives from content companies are well-versed in the logic of ‘launch now, iterate later,’ but this isn’t suitable for the stable and gradual pace required in the hardware industry,” said Chen, alluding to the mismatch of capabilities between the “Byte faction” and the requirements of Pico.
Despite this, Pico’s substantial investment across content, marketing, channels, and hardware underscores its significance in the Chinese VR landscape.
As major internet factories reassess priorities in a global economic downturn, Pico faces scrutiny. The dwindling patience of these major players signals a broader trend seen in the collective withdrawal of domestic content companies from the VR landscape in 2023. Tencent’x extended reality business department faced massive layoffs and eventually dissolved. iQiyi Smart changed CEOs for the first time since its establishment amidst financial issues. The XR lab of Alibaba Group’s Damo Academy and Kuaishou’s VR panoramic video business have quietly dissolved as well.
The VR industry’s upheavals in 2023 signal a necessary downturn and adjustment. The once prevailing notion of achieving miracles with sheer strength in the internet industry proves unsuitable in the VR domain, particularly in China.
Apple’s entry into the mixed reality space with the Vision Pro—a headset that allows viewing of the outside world and is no longer closed off—has also introduced a new variable. Meta apprentices, AR, and VR companies are now pivoting toward Apple’s model, acknowledging a potential shift from VR to AR in the near future.
“AR will likely surpass VR next year, this is not a suspense. It’s just a pity that the ‘defeats’ of a few big players this year has brought negative emotions to the industry. Both the industry and its ecosystem partners need to endure,” said Misa Zhu, founder of Rokid.
Huang Feng, founder of Playdream, added that while paying homage to Meta and Pico for advancing the industry, the next step will be to learn from Apple and Huawei.
According to 36Kr, Pico, despite scaling back, has retained a team exploring MR to investigate how to align with Apple’s Vision Pro.
Ultimately, Pico’s withdrawal doesn’t mean it is the end of the road. While it may mark the conclusion of the Chinese VR industry’s emulation of the Meta route, indicating a transition is underway, the actual battle has yet to truly commence.
(Note: Some employee names may have been changed to protect their identities.)
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This article was written by Qiu Xiaofen for 36Kr.