As various manufacturers dive into the AI glasses race， vivo has decided to forge a different path.
On the afternoon of August 21st， vivo officially unveiled its first MR (Mixed Reality) headset， the “vivo Vision Exploration Edition，” a device directly targeting Apple’s Vision Pro.
The day before the launch event， Huxiu was invited to an experience session for the vivo Vision. In the span of half an hour， this product left a particularly profound impression on the author.
To state the conclusion upfront: If the core reason for Vision Pro’s lukewarm sales today is its “poor wearing experience，” then vivo’s product significantly outperforms it in a 360-degree comparison.
However， the problem unfortunately does not lie here.
What truly hampered Vision Pro was its “extreme lack of content，” an issue that Apple is powerless to resolve， and vivo is even less capable of addressing.
Of course， vivo doesn’t intend to stubbornly push against this wall. According to Hu Baishan， Executive Vice President of vivo， “MR headset devices will initially satisfy the demand for immersion， and in the long term， they will become the ‘eyes’ of home robots.”
It’s worth noting that in March of this year， vivo established a robotics lab and announced its focus on the consumer market. From this perspective， the vivo Vision launched today appears more like a product paving the way for its future robotics business segment.
However， is “MR + Home Robot” truly a compelling combination?
Not a “Apple Alternative”
Prior to the official launch of the vivo Vision， many in the industry anticipated it to be an “Apple alternative.”
Indeed， in terms of product form factor and functional implementation， the two share many similarities. However， the author believes the vivo Vision is a product that is “very Apple-like” but not a direct alternative， as it has genuinely addressed some of the pain points found in the Vision Pro.
The most apparent improvement is its wearing experience.
Firstly， compared to the 600-gram weight of the Apple Vision Pro (excluding the battery)， the vivo Vision weighs a mere 398 grams， making it one of the lightest in its class currently.
It’s important to understand that MR products， compared to smartphones， contain a significantly greater number of components.
For example， the vivo Vision is equipped with a total of 11 cameras. The seven front-facing lenses are responsible for capturing images and spatial depth， two lenses on the underside are for gesture recognition， and two internal lenses track eye movement. Therefore， reducing the weight of MR products is a highly challenging endeavor.
Furthermore， the weight distribution design of the Apple Vision Pro can be considered problematic. The heavy body and forward-leaning center of gravity result in noticeable discomfort， even after just 10 minutes of wear.
Speaking of which， it’s worth critiquing Apple’s design philosophy: It seems they wanted to avoid the bulky， helmet-like form of Meta Quest without making significant adjustments to the center of gravity， resorting to a single strap to support the 600-gram weight， largely disregarding ergonomic principles.
In this regard， the vivo Vision is demonstrably superior. The reduced body thickness shifts the center of gravity more towards the rear， and the split dual-strap design alleviates pressure on the face. During the 30-minute experience， aside from a slight feeling of warmth， the author did not experience any discomfort.
In terms of interaction， the operational logic of the vivo Vision is largely consistent with the Apple Vision Pro， utilizing a combination of “eye tracking + gesture recognition.” If you’ve used the Vision Pro， you should be able to transition to the vivo Vision seamlessly.
A minor difference lies in gesture operation: Apple’s Vision Pro relies more on “finger tapping，” while vivo Vision primarily uses “pinch” gestures with the thumb and index finger for clicking. Each approach has its advantages and disadvantages depending on the usage scenario and function.
The “passthrough” experience is also worth highlighting.
This is what distinguishes MR headsets from VR/AR glasses. Essentially， passthrough involves the headset’s external cameras capturing real-time footage of the surrounding environment， which is then processed by software and projected onto the internal display. This allows users， while wearing the headset and obstructing their direct vision， to still “see the real world” and overlay virtual content onto it.
The primary challenge here is latency. Apple developed a proprietary “R1 chip” to process real-time data from sensors such as cameras， microphones， and LiDAR scanners， compressing the latency of this data transmission link to under 12 milliseconds.
While vivo， at this stage， does not possess the same level of research and development capabilities to create a similar co-processor chip， the actual experience did not show any issues of “virtual-spatial anchor misalignment” caused by excessive latency.
It is noteworthy that the vivo Vision Exploration Edition， unveiled this time， is not yet officially on sale. However， digital enthusiasts can experience it firsthand at stores in over 10 cities nationwide.
MR + Robotics: A Viable Path?
As mentioned earlier， the most significant drawback of current MR headset devices on the market is the scarcity of ecosystem content. So， how does the vivo Vision fare in this regard?
Let’s first discuss the types of applications. In the tested prototype， the vivo Vision’s applications are primarily a combination of “watching movies + viewing sports + gaming + spatial album，” essentially mirroring Apple’s Vision Pro. Functions like “efficient office work” are largely gimmicks at this stage and lack practical significance.
However， vivo Vision does introduce some innovation in certain content types. For instance， in the “Sports Live” application， when watching esports competitions， the upper half of the screen displays the match footage， while the lower half generates a real-time， functional virtual map.
Considering this is a product leaning towards a demo， it’s likely that vivo will subsequently enhance its application ecosystem.
However， it must be reiterated that the issue of content deficiency for MR headset devices is not something a single company can resolve. For application developers， setting aside the issue of profitability， even basic technical considerations like “light and shadow consistency” and “handling of virtual-real occlusion relationships” can deter many.
vivo’s proposed solution is to gradually integrate MR headsets with robots in the future.
This approach could effectively address the content shortcomings. After all， if it’s positioned as a “robot peripheral product，” there would be no need for an extensive application ecosystem; vivo could entirely manage the software side independently.
However， this might not be a perfect solution.
Firstly， it’s important to note that the combination of MR headsets and robots is not an idea originated by vivo.
Last year， after NVIDIA launched its imitation learning-based Blueprint， many embodied intelligence companies began using the Vision Pro to train their humanoid robots.
The author also had the opportunity to try this approach. However， during the experience， a notable issue was the lack of “force feedback，” which prevented the controller from achieving good perceptual understanding. While adding a vibrating controller could resolve this， it would conflict with gesture operations.
Furthermore， with current technological means， controlling robots using MR headsets still cannot achieve high-precision control. It necessitates the deployment of local edge computing gateways to reduce transmission jitter， which seems to contradict the lightweight concept that MR headsets aim for.
Nevertheless， it is worth mentioning that from a technological maturity standpoint， both MR headsets and embodied intelligence， if compared to the developmental stages of smartphones， are perhaps akin to the “Windows Mobile” era. Considering the current excitement in both fields， there may truly be unforeseen sparks of innovation that emerge from their convergence.
