Revisiting XREAL Xu Chix: Making Glasses a Marathon, Luck and Leveling Up

TL;DR · AI Summary
XREAL launches new brand xbx to respond to market competition pressure, showcasing its long-term strategy and technological innovation.
Key Takeaways
- XREAL introduces new brand xbx priced at 1699 yuan, emphasizing affordability an
- XREAL's A face is high-end innovation, while the B face is youthfulness and affo
- XREAL actively responds to market competition, believing that lively competition
Outline
Jump quickly between sections.
XREAL launches new brand xbx to respond to market competition pressure.
XREAL continues to invest in high-end innovation, launching Project Aura.
XREAL introduces new brand xbx priced at 1699 yuan, emphasizing affordability and youthfulness.
XREAL actively responds to market competition, believing that lively competition benefits industry development.
XREAL has high capital utilization efficiency with low losses.
XREAL invests in autonomous chip research to improve core computational autonomy.
Mindmap
See how the topics connect at a glance.
查看大纲文本(无障碍 / 无 JS 友好)
- XREAL 发布新品牌 xbx
- A 面:高端创新
- Project Aura
- B 面:年轻化和平价化
- 定价 1699 元
- 市场竞争与应对策略
- 积极应对竞争
- 认为热闹的竞争有助于行业发展
- 资本利用效率
- 高资本利用效率
- 技术策略
- 投资自主芯片研发
Highlights
Key sentences worth saving and sharing.
XREAL introduces new brand xbx priced at 1699 yuan, emphasizing affordability and youthfulness.
XREAL's A face is high-end innovation, while the B face is youthfulness and affordability, reflecting its long-term strategy.
XREAL actively responds to market competition, believing that lively competition benefits industry development.
XREAL 将今年的第一场发布会留给了一个之前未曾听说的新品牌:xbx。其内部全称为 x, by XREAL。
考虑到 1699 的定价,xbx 的首款产品 a01 的性价比相当不错:50° 视场角,搭载 tandem OLED 显示模组的亮度高达 1600 尼特,相当于约 4 米大小的 147 英寸屏幕,支持 HDR10 和在最高 120Hz 帧率下的空间防抖。
然而,参数远不如外观和佩戴舒适度重要。a01 重量仅 62g,拥有半透明的未来感机身,并且提供多种可更换的个性化镜框。CEO 徐驰表示:“颜值即正义,尽情享受吧。”
这是成立十年来一直在拼命向上攀升的 XREAL,首次向下尝试。

在过去几年中,徐驰和他的 XREAL 团队从未走过的是一条平坦的道路。尽管中国的消费电子供应链资源丰富,使得整合能力成为成功硬件创业者的必备素质,但 XREAL 却不专注于纯粹的供应链整合。
相反,XREAL 一直在向上传递价值,打造最昂贵、最具挑战性和“极客”性质的产品。为此,XREAL 不惜投入极高比例的自研,甚至因过去两年国际半导体市场的波动而牺牲了一部分利润。
这就是为什么去年 XREAL 能与硅谷巨头 Google 合作推出 Project Aura,一台令行业刮目相看并打破了消费者对现有“智能眼镜”刻板印象的原型机(今年将正式面市)。徐驰毫不掩饰地将 Project Aura 称为他所在行业的“天花板”。
能做出这样产品,绝非靠供应链整合。为什么只有 XREAL 能做到,为什么谷歌和其他公司如 LG、ROG 也选择了 XREAL?
徐驰说,答案在于 XREAL 的 A 面:内敛、冷峻、长期主义、押注绝对的技术创新。

那么,XREAL 的 B 面又是什么?
在徐驰带领公司不断优化头戴式显示技术性能和实现极致轻量化的同时,他的背后却危机四伏:
在通过 Apple Vision Pro 试水后也犯错,苹果随后立项了更多轻量级 AI/AR 眼镜产品,预计将在 26H2-27 全年逐步问世;小米、阿里千问、Rokid、VITURE 等纷纷杀入市场。
其中一些产品用 AR 显示眼镜围攻 XREAL 占据已久的光明顶,更多的则是用 AI 眼镜(屏显/无屏)来抢占行业领头者尚未明确布局的新空间——无论何种产品定义,价格都被压得越来越低。
对此,徐驰并不紧张。经过深入交谈后,爱范儿感到 XREAL 的 A 面朝前太久,以至于同行们似乎误以为它没有或不屑于展现另一面。
“怎么说呢,A 面没立住的时候,就没有 B 面。我们现在来了,虽迟但到。”
XREAL 主品牌的势能已积累到位,徐驰和他的产品团队终于有余力去做另一种风格的产品——一个更年轻、外放、价格也更亲民的牌子 xbx。
这就是 XREAL 的 B 面,与那个永远创新不止的 A 面相映成趣。
他说,自创业以来,见过 VR/AR 的泡沫破灭,然后元宇宙到来,也破灭了。一路走来,这次创业越来越像一场万米长跑——这也是他一开始就坚信的赛制。“抢跑一点也不重要,跑对方向才重要。”
因此,徐驰并不担心这些新来的竞争者。问他怕不怕大厂和其他创业公司一拥而上,他回答:“我们最怕的,是这个行业只有自己。如果没有别人,可能说明我们走错了方向。所以热闹一点挺好的。”
2016 年,徐驰离开混合现实先驱企业 Magic Leap,回国创业,制作一副当时无人看好的眼镜。
十年过去了,他庆幸 XREAL 能活到现在,运气占了相当大的因素。
“感谢这十年赛道的起伏,让我有机会不断打怪升级……等到真的要跟大厂掰手腕的时候,不至于一上来就是总决赛。”
爱范儿等与徐驰、XREAL 产品负责人刘宗楷进行了专访,从全新的子品牌 xbx 和第一副价格定为 1699 元的 AR 眼镜产品 a01 开始聊起,一路谈到他对竞争的看法、如何比较自己与同行之间的资本效率、AI 眼镜是否会最终取代手机,以及一个第一次创业的人凭什么活到今天。

“年轻人最好的,就是不迷信传统”
爱范儿:XREAL 这些年的特质就是高端化,为什么要在这个时间点推出 xbx 这个品牌?
徐驰:我们一直说,今天的智能眼镜行业很像 05、06 年的手机行业,很碎片化,系统、应用生态、交互范式都不统一。在这个相对早期的阶段,没有哪个品牌能够覆盖所有的价位段。所以我们就想,有没有可能做两种风格不一样的产品,像 A 面与 B 面一样。
这个行业充满波折,是出了名的难做。很长一段时间里,大家都在摸索,我们自己也(一段时期内)没有一个特别清楚的定位。但是“XREAL”在我心里就是那个极致创新的品牌,更冷酷、更经典。但是,一个品牌尚未立住的时候,我没办法再去做一个更大众、更宏观的东西。
慢慢地,XREAL 成为了我想要的那种更内敛的品牌,这时候,就可以有一个更绽放的品牌跟它相互映照了。这就是我说的 A 面与 B 面。A 面没立住的时候,就没有 B 面。
这件事虽迟但到。从今往后,我们不只是一家叫 XREAL 的公司,还是一家 x by XREAL 的公司。
爱范儿:年轻人想要什么样的 AR 眼镜?难道年轻人就不想要极致的产品?
刘宗楷:对年轻人来说,个性与自我表达是每个人心里的渴望。市面上不管是 AR 眼镜、AI 眼镜,还是 XR 头盔,很多人下意识觉得这东西就该不好看、不轻便。但我们偏要反着来,为什么不先做出一副好看、够轻、年轻人愿意戴在头上的眼镜?一副愿意戴出门的眼镜,是所有事情的第一步。
徐驰:年轻人最好的,就是不迷信那些传统的大道理。颜值就是正义,好看就好,好用就好。我们希望用 a01 这副眼镜让大家明白,一千多块的价格也可以做到两千多块的体验。我们会把它长期做下去。
爱范儿:必须戴到外面,才能影响更多潜在受众。
徐驰:没错就是这样。我们希望这个产品可以在地铁上,在咖啡馆里,在飞机上,在各个地方,更多的年轻人把它戴到外面,而且是不尴尬的。所以我们做了极致的轻量化和个性化设计。

爱范儿:轻量化肯定有取舍。一个产品想做更高的分辨率、帧率、视场角,模组就会变大;模组大了,重量就会大、配重也会失衡。
刘宗楷:做轻的同时还要保证体验,真的非常难。镜片和外壳的厚度能不能再降一点,但强度还能保住。每一个器件既要轻还要保住性能,我们抠了很多细节。
这条路没有尽头,就是一个个夜晚,一次次较劲和争吵。当然,我们觉得还可以做得更好。
爱范儿:AR 眼镜能做到的 FoV(视场角)物理极限是多少?以及不考虑极限,只说在不同场景下人类佩戴的人体工学舒适度,最优解是多少?
徐驰:我给你个最直接的答案,最好的视场角应该是在 85° 左右,但这是在不计成本堆料、不考虑重量的前提下。
在 Project Aura 上,我们做到了 70°,在这个产品形态下已经是非常不错了,但是仍然有差距。什么时候我们能做到 85°,并且仍然是轻量化的,那么我们会觉得至少在显示端做到极致了。
刘宗楷:根据场景来看,比如你戴上 VR 头显去火星,画面主体是一艘宇宙飞船,背景则是宇宙星空——你需要同时看到主体和背景才能获得最大的沉浸感。但是对于 AR 眼镜,最好的背景其实是真实世界。如果是打游戏或者看球赛可能就不需要很大的角度;但如果是看电影,或者附着在真实环境里的 AR 显示,那么宽视角的沉浸感就更重要。所以最终还是取决于内容是否沉浸。
至于人眼的注意力聚焦视角,从眼科学上来说的确有极限,一般就是水平方向的 50° 左右,垂直的 30-40° 这个区域内。
视场角并不是唯一的关键因素,还有电致变色、性能续航等等。在我们定义不同产品的时候,会有无数个取舍的拨杆,往不同的方向去拨。

“我们最怕的,是这个行业只有自己”
爱范儿:苹果也入场了,国内大厂的竞品也已经上市,价格越压越低。你怎么看?
徐驰:大家进来,我认为是好事。我们最怕的是什么?是这个行业只有我们自己——那说明这个赛道没人关注,没人看好。
我们始终认为,眼镜是最有机会替代手机的下一代计算终端。虽然已经创业十年了,我们也才刚刚开始,我们的渗透率可能还不到 1%,后面还有百倍甚至更大的成长空间,所以大家一起来把蛋糕做大是件好事。
我们这个行业是有泡沫的,但泡沫不一定是坏事,说明大家对行业的期待值很高。过去在每个阶段,都有过想挣快钱的人,发现不好赚就走了。泡沫褪去,受害的其实是消费者。而真正推动行业往前走的,是那些把“用户期待”和“产品体验”之间的差距一点点缩小的人。
打个比方,今天的 AI 眼镜就像五岁小孩,而我们定义的全天候佩戴的 AI 眼镜就像贾维斯。这中间的差距得靠底层创新去一点点推动。这些创新不会无缘无故发生,背后一定有人负重前行。
问:你们跟 Meta 的距离还有多少?
徐驰:举个不那么恰当的对比:2025 年 Meta 的 Reality Labs 业务营收是 22 亿美元,亏损接近 200 亿美元。我们今年做到了 2 亿美元营收,差不多是它的十分之一,但我们的亏损不到 2000 万美元。
十分之一的营收,千分之一的亏损,我觉得我们的资本利用效率还可以,这也是我们的优势。
爱范儿:你们有自己的全栈自研芯片、光学,但 Project Aura 的部分算力用的骁龙,两者这两者是什么关系?将来 XREAL 会否提高核心算力的自主性?
徐驰:X1S 是一颗完整的 SoC。在 Aura 上,所有对延迟和带宽敏感的计算,放在我们的 X1S 芯片上,其他的部分给骁龙。
我们的芯片就是纯端侧计算,骁龙芯片放在 puck(外挂的计算单元) 上。这两者不是处理器和协处理器的关系,而更像是“端侧”和“云”之间的关系。有些计算需要发生在离你更近的地方,更加及时。
我们一直说眼镜会取代手机。在可预见的将来,puck 会消失,直接换成你的手机就行了;更长远来看,如果眼镜真的取代手机,它需要自己能够处理所有的计算。这才是我们为什么押注自主芯片的意义所在。
前段时间美国出台了禁令,先进制程的晶圆不能直接运进中国大陆。这件事挺流氓的,我们的芯片在这个范畴内,本来要在大陆做封装,结果必须在台湾封装完才能运回大陆。当时国内一大批芯片厂商都在争抢台湾的封装资源,造成了一次性的短缺,跟今天的内存短缺很像。为此,我们的业绩也少了蛮大一块,否则去年 Q3、Q4 的增长会非常明显。
但从长远看,这反而驱动我们继续往前走。还好我们今天销量不是很大,总比卖了几百万台突然被卡脖子要好(笑)。我们希望中国有越来越多的先进制程握在自己手里,谁也卡不住。
爱范儿:Project Aura 在国内能上吗?你们会选择哪些国内模型厂商一起探索?
徐驰:因为 Android XR 和 Gemini 强绑定,而 Gemini 在国内用不了。所以很遗憾,要么你海淘吧(笑)。
我们不会放弃国内市场,如果 Android XR 能够和 Gemini 解耦,连上国内 AI,就是 Project Aura 进入国内市场的时候,但不是今天。就像当年 iPhone 也不是刚问世就进入中国。我觉得这个结果是可以接受的。It's okay.
对我们来说,阿里是我们的股东,我们也一直跟字节跳动保持交流。在模型方面,我们不会排斥任何一家。我们的终极理想,是 AI 能像搜索引擎一样换着用。未来的大模型会变成基建,谁家的 token 效果好就用谁,可以无缝切换。
“眼镜凭什么取代手机?”
爱范儿:你自己也说,AR 眼镜这个品类存在很多年了,但渗透率仍然很低。让更多人接纳它的“入门毒药”会是一个怎样的产品?
徐驰:大概率还是主流两大类:更加全天候的 AI 眼镜、带显示但不够全天候的 AR 眼镜。
这个“全天候”(always-on) 有两层意思:一是全天候佩戴,二是全天候使用。今天的问题是,AI 眼镜的主要场景不是 AI,而是听歌拍照;你打开相机拍个 30 分钟,产品就没电了。如果说眼镜是你的个人助理,但它每天只能睁眼 30 分钟……那就不是一个全天候的助理。
在将来的某个时间点,会有一款 35g 以下、全天候续航的产品,作为 AI 交互的载体。这样的产品,我认为是能做到的。如果做到了,它绝对会是人手一个的设备。
另一条路就是 AR 眼镜,追求更高清、更多内容。这个路线今天还是分体机形态,能做到 60g,但终极形态可能会是一体机。
There are two products, one like the iPhone, which everyone carries and has an annual shipment of over a billion units; another like our current devices, which could reach the size of tablets and laptops combined, with shipments of 15 to 25 million units annually, also very impressive; and traditional headsets, which might have a volume similar to desktop computers—these three will coexist for a long time.
As for the product that will completely ignite the category and push glasses onto the path of replacing phones, I believe we will see clearer answers by 2027 or 2028.

iF: Even if you achieve the ultimate lightweight design, how do you convince customers who still find it too heavy?
Xu Chi: I think today people tend to easily fall into stereotypes, such as 'I won't wear anything without 35 grams.' In today's industry, apart from subsidized products, there is no product that sells more than a million units without government subsidies. If it really reaches 35 grams, it would already be at 150 million units.
We need to take things step by step: first, get a single product to a million units, then to a million, then to ten million, and finally to one hundred million. There are many steps in between. I believe that today, with good enough experience, a pair of glasses weighing 50 grams should not hinder it from selling a hundred million units. The only thing holding back acceptance and sales is that the experience has not been polished enough.
iF: Smartphone manufacturers feel that smartphones will still be the main event in the next five to ten years. But at the same time, smartphone manufacturers are also making glasses. What do you see as the future competitive landscape?
Xu Chi: Indeed, what exists today will continue to exist for a long time. But the core is who can stand at the highest point of the value chain. Just as we once thought internet giants' super apps were amazing, but today their momentum is certainly less than that of AI companies. Smartphones are the same. As technology advances, there will always be new fields and companies that rise to higher positions in the value chain.
We believe within two years, there will be consensus: glasses are the best native terminal for AI, and they may be the closest thing to multimodal AI. This is why we went together with Google to imagine a new interactive paradigm and what the terminals under this new paradigm might look like.
This excites me because it is difficult, and if done right, it would be highly rewarding.
iF: Other forms of AI hardware, such as pins and headphones with cameras, are inferior to glasses, right?
Xu Chi: Not only do I think so, but Demis Hassabis also says that glasses are absolutely the most central device in all AI. Because only glasses can capture the key contextual information of where someone's attention lies.
When you put on a pin, it can see a bunch of people in front of you, but in the future, glasses will have eye-tracking, knowing exactly what I am looking at at any given moment, and peripheral information may not be as important. Only glasses can bring end-to-end closed-loop data chains, something other terminals lack. Of course, other forms can assist, but glasses are definitely the most critical entry point.

'Lucky to Fight and Upgrade'
iF: Entrepreneurs and entrepreneurs have different sources of fear, whether it's internal organizational efficiency lagging behind the times, competition from peers, or disruption from other industries. What is the fear that wakes you up from sleep?
Xu Chi: Running a business is just like being a person; sometimes we feel lost, and someone gives us guidance to find our benchmark. But I believe ultimately, worries come from ourselves.
I believe great businesses are driven by values. At the core, it's finding a comfortable state of organization where the entire company recognizes your set of values—whether leaving or staying, they will continue to operate under these values. As long as this is achieved, competition or other issues become secondary.
Personally, my sleep is quite good, which I think is a trait entrepreneurs should have (laughs).
If there is truly something that worries me, it's whether the values I hold dear can be fully implemented. I fear that as XREAL grows larger, its culture may dilute. I need everyone to genuinely believe in one thing: we want to be innovators and leaders. This is not easy, especially in China. In China, people generally follow hierarchical systems, 'what the boss says goes,' but I hope that everyone can form a flat and efficient mechanism both top-down and bottom-up.
iF: As you said, through several waves of bubbles rising and falling, XREAL has survived until today.
Xu Chi: In 2016, I returned to China from Magic Leap, and it has been ten years since then. Back then, it was grassroots entrepreneurship, and I simply wanted to make a pair of glasses. Being able to survive until today is really fortunate. This was my first startup, and I thank the investors along the way (as well as other competitors) for helping me understand how to run a business, an organization, and a venture.
To be honest, if this industry had developed even faster and gained momentum earlier, we might not have had the opportunity to hone ourselves and prepare for the intense competition when big players entered the market.
Every startup probably goes through this phase: you need to fight and upgrade before standing on a bigger stage. If you start with a total boss and face opponents like Alibaba and ByteDance right away, there's nothing left to fight. So, I actually thank the ups and downs of this industry's first decade, allowing me to compete with big companies one day.
The AR industry is notoriously difficult, and I've been doing it for quite some time, so I now view these things more calmly. As long as everyone is still at the table, this is a long-term issue.
I believe AR is a marathon, and getting the direction right is more important than starting early. If the industry is still in its early stages but everyone rushes towards the same place, that supposed consensus might just be a bubble. On the contrary, early non-consensus ideas that later prove correct through time. History has taught us this countless times.

By | Du Chen
Interviewed By | Du Chen