Hype Takes Over: Smart-Glasses Industry Faces a Wave of Exaggerated Claims

At the end of October, the founder of a Shenzhen-based AI smart-glasses manufacturer—originally an OEM—expansively listed a string of “good news”: a multimillion-yuan funding round reportedly about to close, contract manufacturing orders worth hundreds of millions this year, order volumes set to double next year, and mass production claimed to be “no longer a problem.”

Sitting across from him was a client urging delivery. That customer had ordered over a hundred smart-glasses units in June but had repeatedly been notified of shipment delays, prompting some overseas clients to cancel orders outright. Faced with these bold claims, the customer simply replied, “Is anyone else here?”—instantly puncturing the spell and redirecting the conversation to other matters.

Such grand declarations are hardly isolated. The founder of one “Top Four AR” company openly told media and investors multiple times: “The entire smart-glasses industry will ship around 500,000 units this year—half of them ours and Xiaomi’s.” The number alone is staggering—enough to electrify any investor audience.

By late 2024, after Ray-Ban Meta found product-market fit and surpassed one million units in sales, China’s “Hundred-Glasses Battle” ignited overnight.

Startups—Rokid, Thunderbird, XReal, VITURE, Even Realities, INMO—each sounded their charge in niche verticals. Meanwhile, Huawei, Honor, vivo, Meizu, and Transsion encircled the field, with major internet platforms including Alibaba, ByteDance, and Tencent hoping to seize a fresh entry point.

Expectation sprinted ahead of reality. No one wanted to miss the chance to back the next breakout device—or even a future computing terminal. And so the game table filled with players claiming massive order books. Investors, however, are not blind. They know someone eventually pays the bill for inflated performance—just so long as it isn’t them this round.

I. How the Numbers Get Inflated

A defining trait of this “Hundred-Glasses Battle” is the chasm between what companies aspire to do—and loudly claim to do—versus what they can actually deliver.

First, a clarification: AR glasses focus on visual augmentation—overlaying digital information onto the real world. AI glasses rely on AI + voice interaction, often without displays. Smart-glasses development is drifting toward integrated AR+AI.

In mid-October, the founder of the aforementioned “Top Four AR” firm planned to fly to Beijing to pitch financing and partnerships to a major telecom operator using its latest AI-smart-glasses prototype. But the device could not be stabilized in time—forcing him to repeatedly reschedule his flight.

That same model had just debuted as the star product at a launch event, complete with polished slides and demo videos showcasing immersive use scenarios. Yet one employee admitted that, by early November, the company struggled to find a single fully working unit suitable for key influencer reviews.

Earlier in the year, the company publicly claimed Q1 orders exceeding 100,000 units. When the new product launched in mid-October, it announced over 20,000 preorders in the first three days and projected 300,000 annual sales by 2026.

But employees across multiple departments say that, as of November 5, total offline + online shipments this year are under 20,000 units.

A major upstream supplier—its exclusive provider of core display components—confirmed that total annual orders from the company are under 10,000 units. Because most AR-optical chips are supplied by only a few vendors, these suppliers can reasonably infer end-market performance by tracking component purchases.

They shared real order numbers: recently, Rokid’s monthly demand for optical chips exceeded expectations, requesting 50,000 units per month, though actual deliveries max out at ~20,000 due to limited capacity. Even Realities placed annual orders of ~65,000 units.

Upstream suppliers possess the most accurate market view because they risk real capital when purchasing raw materials. Their numbers tell the truth—unlike the zero-cost marketing stories floating downstream.

Another factory, specializing in white-label AI glasses, claimed ODM order sales near 500 million yuan in nine months. Based on component pricing, suppliers say this figure exceeds the combined order volume of all their major customers—and is unlikely to be real.

According to a hardware-focused private-equity manager, smart-glasses companies frequently inflate sales figures. Some classify projected “framework contract” demand—non-binding, non-billable—directly as completed sales. Preorders are counted even if orders aren’t fulfilled or are later returned.

Returns are especially high in this category. Top manufacturers face return rates of 40–50%. Some offer zero-deposit preorders, further muddying numbers. In extreme cases, one major brand allegedly transferred stock between warehouses and counted that as fulfillment.

One investor recalled backing an optical-waveguide startup that forecast 100 million yuan in annual revenue but delivered just over 20 million. Year after year, targets were recycled—and never met—until the company collapsed.

Every consumer-electronics boom begins with frothy narratives. As one investor put it, the question isn’t whether there’s foam—it’s whether the foam is healthy. A little makes the beer enjoyable; too much ruins the drink.

II. In the Investor Arena

After reviewing nearly every smart-glasses OEM over two years, some investors decided to bow out. One firm invested heavily in early-stage companies—deploying roughly $70 million over three months—but remains cautious about AI glasses.

In September, VITURE secured $100 million in additional Series B financing, partly from that same cautious firm—yet their team insists they will invest in only one smart-glasses company, regardless of how hot the market becomes.

One investor’s product evaluation standard is simple: Which users were invited to the launch? Could people actually try the product? Many companies only invited local officials and investors—not tech enthusiasts. Prototypes were displayed under plastic cases, with no real hands-on testing allowed.

“In this environment,” the investor said, “every company is inflating numbers. So you look for whoever tells the biggest, most convincing story. Bluffing is also a skill. As long as a company can raise the next round, investors won’t take the blame.”

Instead of preorder hype, he focuses on real, sustainable indicators:
• Actual shipment volume
• Net promoter score
• Real return rate
• User feedback on Amazon, JD, Taobao, Xiaohongshu, and Douyin

Many glamorous product promises fall apart under real user reviews.

Another VC said the priority isn’t technology—it’s valuation and exit strategy. If valuations rise too high or state-backed capital enters, exits become difficult.

III. Still “Half-Finished”

Smart-glasses are seductive because they support all-day wear, hands-free interaction, and multimodal input. Add a CMOS sensor and camera, and they could become powerful personal-computing endpoints with much higher potential than earbuds.

But adding the necessary components—chips, optics, cameras, batteries, motors—creates a brutal hardware equation: heat, weight, and battery life.

One investor noted that adding a camera requires legal inclusion of a flash unit; waveguides + optical engines alone add at least 10 grams; double-battery solutions double weight. The result: bulky glasses that rest uncomfortably on the nose and ears.

He cited RayNeo Air 4—its spec sheet impressed him until he learned it weighs 76 grams, two to three times that of ordinary glasses. Human facial tissue is extremely sensitive to weight and heat—making such designs impractical for everyday use.

Currently, most AR/AI glasses remain half-finished products. Qualcomm’s ARGen chip is effectively reworked from smartwatch silicon with significant power leakage. Users get 30% functionality for 100% heat and battery drain.

To mitigate this, manufacturers adopted dual-chip configurations, as seen in Xiaomi and Quark models. But investors call this a stopgap solution.

Ray-Ban Meta sidestepped the battery challenge by making sunglasses—users remove them indoors, naturally charging in between. Achieving full-day wear demands at least 12 hours of battery life, which no product currently delivers.

Meanwhile, no brand has produced glasses capable of recording more than five minutes of video without overheating.

CMOS space constraints result in suboptimal imaging, forcing companies to use AI upscaling to emulate 1080p quality—and performance remains inconsistent.

Two contrasting philosophies define leading players:
• Rokid: “Tech first—users can be educated.” A familiar approach, yet smart glasses remain niche, expensive toys after 13 years.
• Even Realities: “Fashion first.” They removed low-usage features like cameras and voice assistants, focusing on aesthetics, which helped attract European buyers.

IV. Avoiding the Giants

Some investors avoid backing general-purpose smart-glasses companies altogether. Their logic: Startups are simply scouting the path for smartphone giants. Once upstream tech—micro-LED, waveguide, chip architectures—matures, big phone brands can easily take over using their distribution and supply-chain dominance.

Startups would be crushed.

The real opportunity, they argue, lies in niche verticals—gaming, cinema, outdoor sports—large enough to sustain business but too small to draw full-scale competition from giants.

Others prefer to invest upstream—waveguides, micro-LEDs—playing the “picks and shovels” game with lower risk and clearer exit paths.

One investor noted that despite the hype, most optical and component suppliers have not earned meaningful profits from the smart-glasses boom. Only those with strong technical moats—camera modules, PCB, FPC, frames—see consistent revenue.

Battery suppliers confirm this reality: their main business still comes from earbuds and audio glasses, not AR/AI products.

Ultimately, many end-user features sound imaginative—AI assistants, gesture payments—but remain impractical. Few activate AI assistants more than three times a week; paying via glasses is awkward, battery-draining, and socially conspicuous.

Consumers, expecting 100-point next-gen computing terminals, end up receiving 60-point prototypes—hot, heavy, low-use, and battery-strained. This disappointment is captured by return rates as high as 40–50%.

For startups still seated at the table, raising the next round is a race against time. They must find a final investor willing to buy their vision before the bubble bursts—and before giants take the field.