The First Smart Glasses Company to Publish Audited Financials Ships 130,000 Units a Year. Meta Ships 7 Million.
XREAL's Hong Kong IPO prospectus is the first window into audited unit economics for a pure-play smart glasses maker. An original per-unit decomposition reveals gross profit and R&D cost are identical at $195 per pair, the company burns $19.7 million in cash per year, and its market share is falling even as the market grows 56% annually.
Five hundred and fifty-four dollars. That is how much XREAL earns, on average, for every pair of AR glasses it sells, a number derived from an IPO prospectus filed with the Hong Kong Stock Exchange in early April 2026 that, for the first time in the history of this product category, subjected a pure-play smart glasses company's financials to independent audit. XREAL shipped approximately 130,000 units in 2025, generating $72 million in revenue with a 35.2 percent gross margin that improved meaningfully from 18.8 percent two years earlier. Those are respectable margins for a hardware startup. They are also completely irrelevant when you divide the R&D budget by the same 130,000 units and discover a number that should keep XREAL's investors awake at night: $195 per pair, exactly equal to the gross profit per unit of $195.
The net contribution of each sold pair toward anything other than the R&D required to build the next pair is zero, and XREAL is running on a treadmill that is accelerating beneath its feet.
The Math Nobody Has Run
No reporting on XREAL's IPO filing has performed the per-unit economics decomposition that reveals the company's structural predicament, and the prospectus gives us every input we need to run it in thirty seconds.
| Metric | Value | Source |
|---|---|---|
| Revenue (2025) | $72M (RMB ~518M) | HKEX prospectus |
| Units shipped (2025) | ~130,000 | HKEX prospectus |
| Average selling price | $554 | $72M ÷ 130K |
| Gross margin | 35.2% | HKEX prospectus |
| Gross profit per unit | $195 | $554 × 0.352 |
| R&D spending (2025) | $25.4M (RMB 183M) | HKEX prospectus |
| R&D cost per unit | $195 | $25.4M ÷ 130K |
| Net hardware contribution | $0 | $195 - $195 |
This is not a rounding artifact. R&D spending of RMB 183 million against gross profit of approximately RMB 182 million (518M revenue times 35.2 percent) yields a gap small enough to disappear into currency conversion noise. Every dollar of gross profit funds the engineering team. Nothing remains for sales, marketing, general administration, or the daily cost of keeping 130,000 customers' firmware updated and support tickets answered.
R&D is also declining, which compounds the concern: XREAL spent less on research and development in 2025 than in 2024, which was less than 2023. A hardware company in a nascent category cutting R&D while shipping flat unit volumes is not optimizing. It is conserving cash to survive long enough to reach the IPO window.
The Cash Clock
XREAL ended 2025 with RMB 63.6 million in cash, roughly $8.8 million. One year earlier, the figure was RMB 205 million ($28.5 million). The difference, $19.7 million in net cash consumed over twelve months, tells a blunt story: at the current burn rate, XREAL runs out of money by the middle of 2026 without fresh capital. The IPO is not a celebration of product-market fit but a financing event, filed because private capital stopped arriving with sufficient enthusiasm.
XREAL's Series D round valued the company at approximately $833 million, or 11.6 times revenue. For context, EssilorLuxottica trades at roughly 2 times revenue, and it is the most profitable eyewear company on Earth, with seven billion dollars in annual smart glasses sales growing at a rate that its CEO called "exponential." XREAL's valuation multiple implies explosive unit growth that three consecutive years of flat 130,000-unit shipments directly contradict.
54-to-1, and Widening
While XREAL shipped 130,000 units for the third consecutive year, Meta and EssilorLuxottica sold 7 million smart glasses in 2025 alone, according to EssilorLuxottica's Q4 2025 earnings call. That is a 54-to-1 unit gap, and it is structural rather than cyclical. Meta's smart glasses start at $299 with no display and work as sunglasses. XREAL's AR glasses start above $400 and require a connected phone or computing puck. Different products serving different use cases, but competing for the same analyst reports, the same investor attention, and increasingly the same customer who wants "smart glasses" without knowing or caring about the distinction between camera-only frames and waveguide-display headsets.
Meta plans to expand production capacity to 10 million units per year by end of 2026, with Bloomberg reporting discussions to scale to 20 to 30 million units by 2027. If those projections hold, the unit gap moves from 54-to-1 to somewhere between 154-to-1 and 231-to-1 within eighteen months, assuming XREAL's shipments remain flat.
The market share math is equally unforgiving, and it gets worse with every quarterly forecast revision. IDC projects 22.67 million smart glasses shipments globally in 2026, a 56.3 percent year-over-year increase driven primarily by Meta and a growing roster of Chinese competitors. XREAL's 130,000 units against that total gives it 0.6 percent of the market. Even in the narrower AR-glasses-with-display subsegment, where TrendForce forecasts 950,000 units in 2026, XREAL's share drops to about 14 percent, assuming it holds volume while Rokid and others surge.
Rokid's Counter-Datapoint
Rokid, a Chinese competitor making display-equipped AR glasses, shipped approximately 300,000 units in 2025, triple its initial forecast of 100,000. Rokid CEO Miao Zhu's viral demo at a January 2025 conference drove 40,000 first-week sales, with 30 to 40 percent of orders facing delivery delays into 2026. The company raised its price by RMB 800 (approximately $111) mid-campaign and still could not keep up with demand.
Two details from Rokid's 2025 data matter more than the unit count. Only 16 percent of buyers self-identified as working in technology, down from 50 percent the prior year. Mainstream consumers are buying display AR glasses. And Rokid outsold XREAL by more than two-to-one despite XREAL's three-year head start and superior brand recognition in Western markets. XREAL generates 70 percent of revenue overseas. Rokid is eating the domestic Chinese market that XREAL vacated.
The Samsung Problem in Reverse
XREAL's strongest strategic asset is its position as the "lead hardware partner" for Google's Android XR platform, which is scheduled to launch on glasses hardware in late 2026 with XREAL's Project Aura, a device with a reported 70-degree field of view. This partnership is the bear-case killer, the reason XREAL's Series D investors paid 11.6 times revenue for a company with flat unit growth and declining R&D.
The Samsung analogy is instructive and also cautionary. Samsung became the dominant Android phone maker by being first to build on Google's platform with industrial manufacturing capacity that dwarfed every other early Android OEM. Samsung shipped 300 million phones per year by 2013. XREAL ships 130,000 glasses per year. Samsung had tens of billions in cash reserves to fund the scaling sprint. XREAL has $8.8 million. Samsung's Galaxy S launched into a market with no dominant smartphone platform competitor, because iOS was exclusive to Apple. XREAL's Android XR glasses will launch into a market where Meta has already shipped 9 million cumulative smart glasses and is ramping toward 20 to 30 million per year with manufacturing by EssilorLuxottica, the world's largest eyewear producer, in Italian factories that avoid the 27 to 65 percent tariffs hitting Chinese-made optics.
Being the first hardware partner for a winning platform mattered when you could scale production to match demand. XREAL's IPO prospectus contains no evidence that the company can scale production to match demand, because demand has not materialized in the first place. Three years of 130,000 units is not a supply constraint. It is a demand reality.
What the Numbers Actually Say About Independent Smart Glasses Companies
XREAL's prospectus is the first audited test of a thesis that much of the venture capital community has accepted on faith: that independent smart glasses startups can build viable businesses in a market dominated by the largest technology and eyewear companies on Earth. The data does not support the thesis.
| Metric | XREAL (2025, audited) | Meta/EssilorLuxottica (2025, estimated) |
|---|---|---|
| Units shipped | 130,000 | 7,000,000 |
| Market share (all smart glasses) | ~2.6% | ~56.8% |
| Revenue | $72M | Not disclosed separately |
| Gross margin | 35.2% | Not disclosed separately |
| Cash on hand | $8.8M | $43.9B (Meta, Q4 2025) |
| Manufacturing partner | Chinese ODM | EssilorLuxottica (Italian) |
| Tariff exposure | 27-65% on US imports | Minimal (Italian manufacturing) |
| Production capacity target | Not disclosed | 10-30M units/year |
The 35.2 percent gross margin is genuinely impressive for a hardware startup at this scale and deserves acknowledgment. Two years ago that number was 18.8 percent. XREAL's engineering team has driven significant manufacturing cost improvements even without volume leverage. If you could inject 10 million units of demand into XREAL's supply chain tomorrow, the margins would likely expand further and the per-unit R&D amortization would drop to $2.54 instead of $195. The problem is that "inject 10 million units of demand" is not a strategy. It is a wish.
The Strongest Case for XREAL
Android XR could be the injection event. If Google's platform generates the developer ecosystem and consumer demand that makes AR glasses a mainstream category, and if XREAL's hardware is good enough and available early enough to capture even a modest share of that wave, the unit economics transform overnight. At 1 million units per year with the same 35.2 percent margin, R&D cost per unit drops to $25 and the company generates roughly $170 million in gross profit against $25 million in R&D. That is a viable business, possibly a very good one.
XREAL's margin trajectory also suggests manufacturing competence that would compound with volume. Moving from 18.8 percent to 35.2 percent gross margin in two years while holding unit prices roughly constant means the team knows how to reduce component and assembly costs. If Android XR demand arrives, XREAL is not starting from a position of manufacturing incompetence.
The question is timing. XREAL's $8.8 million in cash buys roughly five months of operations at current burn rate. The IPO needs to close and fund the company before the cash runs out, and then Android XR needs to generate demand before the IPO capital runs out. Two sequential bets, each dependent on external forces XREAL does not control, with a margin of error measured in months rather than years.
What This Analysis Does Not Know
XREAL's exact unit breakdown by model is not available from the prospectus reporting. The 130,000 figure is an aggregate across product lines, and display AR glasses command higher ASPs than simpler models, so the $554 blended average may mask meaningful variation. Meta does not separately disclose smart glasses revenue or margins. The 7 million unit figure comes from EssilorLuxottica's earnings call and could include a small number of non-Meta smart glasses models, though Meta dominates EssilorLuxottica's smart glasses output. Revenue splits between Meta and EssilorLuxottica on each pair sold are not public, making a direct per-unit profit comparison between the two companies impossible.
More fundamentally, XREAL makes AR glasses with optical displays, while Meta's high-volume products are camera-and-speaker glasses without displays. These are genuinely different products in the same market report. The 54-to-1 gap is real, but the comparison is somewhat apples-to-oranges, a distinction that becomes less meaningful as Meta adds displays to its product line with the Ray-Ban Display shipping in 2026 and as XREAL potentially moves into simpler form factors.
Finally, the cash burn estimate relies on year-end balance differences. Actual quarterly burn rates may vary significantly if the company front-loaded or deferred spending during the IPO preparation period.
The Bottom Line
XREAL's IPO prospectus is the most important document in the smart glasses industry right now, not because XREAL is the most important company in the space but because it is the only company that has been forced to show its books. What those books reveal is a company with genuinely strong manufacturing margins, a product that 130,000 people per year want to buy at $554, and a financial position so precarious that the IPO itself is the product.
If you are an investor evaluating smart glasses startups, XREAL's filing is your first empirical benchmark for what independent smart glasses economics actually look like at low volume: 35 percent margins consumed entirely by the R&D required to stay competitive, zero free cash flow, and a market share trajectory that declines even as the market grows. If you are a developer choosing a platform, the unit base matters: 130,000 installed devices is a rounding error against Meta's 9 million cumulative, and developer tools follow installed base, not aspiration. If you are a consumer deciding between AR display glasses, watch for the Android XR launch in late 2026. That is the event that will determine whether XREAL becomes the Samsung of smart glasses or the Palm of smart glasses, and the difference between those outcomes is measured in months of cash runway and millions of units of consumer demand that do not yet exist.
Sources
- ChinaBizInsider / EqualOcean (April 2026). XREAL HK IPO prospectus analysis: $72M revenue, ~130K units shipped, 35.2% gross margin, RMB 63.6M cash, $833M Series D valuation, CICC + Citigroup sponsors. EqualOcean
- UploadVR (February 2026). EssilorLuxottica Q4 2025 earnings: 7 million Meta smart glasses sold in 2025, production capacity expansion plans. UploadVR
- IDC (2026). Global AI/smart glasses shipment forecast: 22.67M units in 2026, 56.3% YoY growth. IDC
- TrendForce (January 2026). Global AR glasses forecast: 950,000 units in 2026, 53% YoY growth. Meta Ray-Ban Display orders revised up twice. TrendForce
- Bloomberg (February 2026). Meta and EssilorLuxottica discussions to scale smart glasses production to 20-30 million units. Bloomberg
- ChinaBizInsider (2026). Rokid 2025 shipment data: 300,000 units, 16% tech buyers, $111 mid-campaign price increase. ChinaBizInsider