Xreal Says It Finally Mastered Smart Glasses. The IPO Filing Says Otherwise.

Xreal says the smart glasses industry has finally turned a corner. The numbers suggest otherwise.

Chi Xu, the founder and CEO of Xreal, spent Google I/O this week telling anyone who would listen that the company and the entire sector had reached an inflection point. "Everybody's losing money," he acknowledged to TechCrunch. "That's because it's very hard, what we're doing." But, he argued, next year is when things actually break even.

The pitch is compelling. Xreal is Google's chosen partner for Android XR, the search giant's spatial computing platform. Its Project Aura glasses offer a 70-degree field of view, run full Android apps, and integrate Gemini AI. The company has filed for a Hong Kong IPO. PitchBook data shows it raised $100 million at a $833 million post-money valuation in its January 2026 Series D. It has four consecutive years of market share leadership in AR glasses. By any measure of ambition, Xreal has arrived.

The financials tell a different story.

According to Xreal's IPO prospectus, the company generated 516 million yuan in revenue during 2025, roughly $72 million. For context: that is the annual revenue of a mid-sized chain restaurant. The company's gross margin has improved meaningfully, climbing from 18.8% in 2023 to 35.2% in 2025. But adjusted net losses were still 250 million yuan last year. Xreal is burning cash, cutting marketing spend, and shrinking toward an operating breakeven that keeps sliding right.

The unit volumes are more revealing. Xreal sold 111,000 units of its midrange One series glasses in 2025, its strongest product line. That sounds like progress until you note that its entry-level Air series collapsed from 104,000 units in 2024 to just 17,000 in 2025, an 84% drop in the mass-market tier. Its flagship Light-Ultra-Aura line, the precursor to Project Aura, moved fewer than 5,000 units for the entire year.

Against this, consider Meta. Ray-Ban Meta glasses have no display, no AR, and essentially function as a camera and audio accessory with an AI layer. They sold 7 million units in 2025. Seven million. Meta and EssilorLuxottica more than tripled their smart glasses sales year-over-year. The division still operates at a loss, reportedly billions deep, but the unit volumes are real.

Here is the inconvenient truth buried in Xreal's narrative: the smart glasses market that actually exists is Meta's market. Screenless AI audio glasses are the category that has reached an inflection point. Display-wearing AR glasses, Xreal's domain, remain a developer and early-adopter niche with unit economics that don't yet support meaningful scale.

Xreal's answer is Project Aura. Launching in 2026, it pairs Xreal's X1S spatial computing chip with Google's Android XR platform and a tethered pocket-sized compute puck. The 70-degree field of view is genuinely wide for optical see-through displays. The Gemini integration is real. The developer catalyst program is live.

None of that changes the core problem. Aura will launch as a developer kit with limited consumer availability. People cannot walk into a store and buy it today. The form factor requires a wire running from glasses to a computer in your pocket. The cheapest current Xreal product with a display costs around $400; Aura will almost certainly cost more. And Meta has a 7-million-unit head start, a retail distribution network through Ray-Ban and Oakley stores, and a per-unit cost structure built on EssilorLuxottica's manufacturing scale that Xreal cannot match at 133,000 annual units.

Xu is right that the industry has reached a turning point. He is wrong about which direction it points.

The real story in Xreal's IPO filing is not that the company has mastered a notoriously difficult industry. It is that Xreal is attempting to position itself as the Android XR hardware layer, the equivalent of a Samsung or LG for spatial computing, before the market for spatial computing hardware has materialized at consumer scale. The bet is that as Android XR matures, device manufacturers will need a reference design and a proven partner. Xreal is trying to be that partner, using Google as a validating customer and the IPO as a way to fund the runway until the category matures.

It is a reasonable bet. It is also a bet that requires the AR glasses market to actually exist, rather than remain perpetually one year away from breaking through.

The question for anyone evaluating Xreal's IPO is not whether the company has figured out smart glasses. It has not. The question is whether Google's spatial computing platform will generate enough demand for Android XR devices that Xreal's position as a preferred hardware partner becomes valuable before the company's cash runs out. The prospectus shows declining losses but also declining marketing spend and a cash position that dropped roughly 70% over the course of 2025. The break-even target of 2027 is a year further out than the 2026 target the company was reportedly circulating a year ago.

Xreal has done something genuinely difficult: stayed alive in an industry that has killed better-funded competitors, built a real partnership with Google, and improved its margins substantially. The hardware works. The technology is not vapor. But "we've finally mastered this industry" requires a market that buys what you're selling. Right now, that market is Meta's to lose.