Rokid's 800% Sales Surge Reveals Two Competing Smart Glasses Playbooks
Rokid reported 800% year-on-year sales growth at its June 26 Open Day, with 50-60% of users wearing the glasses daily — a retention signal that suggests smart glasses are moving beyond novelty. The company unveiled YodaOS, an Agent-first operating system that abandons the app paradigm, while CEO Misa Zhu compared the current market moment to the pre-iPhone BlackBerry era. The deeper story is a fundamental split between Western fashion-first and Chinese AI-first product philosophies that will shape how the entire category develops.
## Why This Matters for PMs The question this forces you to confront is which paradigm you're designing for — and they're not compatible. If you're building an experience for Meta Ray-Ban's ecosystem, you're optimizing for ambient, low-friction AI that doesn't demand attention. If you're building for the Chinese AI-first model Rokid is running, you're designing for a user who chose the glasses specifically because they want an AI-native computing experience. Same form factor, completely different intent signals, completely different UX bar. The concrete thing you should do right now: audit your product roadmap for any 'wearable AI' surface area and get explicit about which paradigm assumption is baked in. A lot of teams are accidentally designing for both and will end up serving neither. On timeline — Rokid's 2 million unit threshold for a viable developer ecosystem is your trigger event. We don't know exactly where they are today, but if that number gets announced in Q3 or Q4 2026, the platform calculus changes quickly. That's when third-party development becomes investable, not just experimental. If you're waiting to see if this category is real before committing resources, the window for early positioning is shorter than it looks.
Rokid hit 800% year-on-year sales growth with products shipping to 166 countries as of June 2026.
Forget the hardware specs. The most interesting thing Rokid CEO Misa Zhu said at the company's Open Day on June 26 wasn't about sales numbers — it was the BlackBerry comparison. He's saying we're pre-iPhone right now.
The rocket has launched, but the killer app hasn't shown up yet. That tension is exactly what makes this moment worth paying attention to.
The Number That Actually Matters
Yes, 800% year-on-year sales growth is a headline. But the stat you should be writing down is this: 50-60% of activated Rokid users wear the glasses throughout the day. That's a retention signal, not a sales signal.
Consumer hardware is a graveyard of products people bought once and never touched again. Daily wear is the benchmark that separates a gadget from a platform, and Rokid is clearing it.
The company now ships to 166 countries. That's a distribution footprint that forces you to take them seriously as a global player, not just a Chinese market story.
China vs. the West: Two Totally Different Bets
Here's the angle that didn't make the press release: China and the West are building completely different products that happen to look identical from the outside.
Meta Ray-Ban's playbook is fashion-first. You buy the glasses because they look good, and the AI assistant is a bonus feature you might use occasionally. It's a Trojan horse strategy — get the hardware on faces, build habits slowly.
Rokid and the Chinese market are running the opposite logic. The glasses are an AI computing device that happens to sit on your face. Users are buying for the AI capability, full stop.
Form factor is a delivery mechanism, not the value proposition.
Neither approach is wrong. But they're going to produce very different ecosystems, very different developer incentives, and very different monetization models. If you're building in this space, which paradigm you're designing for matters enormously.
YodaOS and the End of the App Paradigm
Rokid's biggest technical announcement was YodaOS — what the company is calling the world's first AI operating system built specifically for smart glasses. The architecture is worth understanding because it's a direct challenge to how we've built software for the last 15 years.
YodaOS ditches the app-based model entirely. Instead of launching apps, the OS uses an Agent-first architecture where AI dynamically generates interfaces based on what you're trying to do. JD.com, Amap, Alipay, and WeChat have all opened their underlying APIs for integration, which means the system can pull from real commerce, navigation, payments, and messaging infrastructure — not sandboxed demo environments.
Rokid users are already consuming over 1 billion tokens daily. The company is exploring token-based billing as a monetization model, which is a fascinating structural bet — essentially treating AI inference as the utility layer, the way mobile carriers once sold data.
The honest caveat: Misa himself acknowledged the developer ecosystem is still nascent. His threshold for a sustainable third-party ecosystem is 2 million units of installed base. He didn't say where Rokid is today relative to that number, which is the question every analyst in the room should have been asking.
What the 'BlackBerry Era' Frame Actually Tells You
When a CEO voluntarily compares their own product moment to the pre-iPhone BlackBerry era, that's either unusual humility or a very calculated expectation-setting move. Probably both.
What Misa is really saying: the infrastructure is real, the retention is real, but the product that makes this category obvious to everyone hasn't been built yet. The killer application is missing. That's not a reason to dismiss the space — it's actually the most important signal for where the opportunity sits.
The companies that win in smart glasses won't necessarily be the hardware makers. They'll be whoever figures out the use case that makes taking the glasses off feel like a downgrade. That application doesn't exist yet.
Which means the window is open.
Watch for Rokid's installed base disclosure in Q3 2026 — if they announce crossing 2 million units, the developer ecosystem story accelerates fast and the platform competition with Meta gets real.
Frequently Asked Questions
YodaOS uses an Agent-first architecture, meaning there are no discrete apps to launch. Instead, the AI reads user intent and dynamically generates the interface and pulls the relevant data — from partners like Alipay or WeChat — in real time. Think less 'open an app' and more 'the system figures out what you need and surfaces it.' It's a fundamental departure from how we've built software since the smartphone era.
It's a real structural difference with downstream consequences. Meta Ray-Ban leads in Western markets by selling fashion eyewear with AI as a secondary feature — the purchase decision is aesthetic. In China, Rokid users are buying explicitly for AI capability, which means higher engagement expectations and a different tolerance for friction. These two user bases will generate very different developer incentives and monetization models over time.
One billion tokens daily across the user base is a meaningful signal that people are actively querying the AI, not just wearing the hardware passively. It also explains why Rokid is exploring token-based billing — if inference is the core value exchange, pricing around it makes structural sense. The risk is that heavy token consumption is expensive to subsidize at scale, so the billing model needs to catch up to usage before margins get ugly.
