Exclusive | Burned 2.2 billion over ten years, only 64 million left in the account: XREAL will either go public or go bankrupt

When XREAL, wearing the halo of “global AR glasses market leader” and “first in market share for four consecutive years,” filed its IPO prospectus with the Hong Kong Stock Exchange, what the capital market saw was not the glory of a tech wunderkind, but a decade-long capital disaster that burned through $2.2 billion, left only $64 million on the books, and racked up cumulative net losses of over $2 billion in three years.

Founded in 2017 and established by Xu Chi, a former NVIDIA GPU architecture engineer, this company used to be one of the brightest stars in the AR track—cumulative funding of $320 million (about $2.3 billion RMB). Its shareholder roster was star-studded: Alibaba, Kuaishou, Sequoia, Hillhouse, Shunwei, Yunfeng, NIO Capital, Luxshare Precision… almost gathering half the internet and hard-tech landscape.

With a 27% global market share, products covering 40 countries, and overseas revenue accounting for 71%, it appeared to stand atop the industry. Yet the financial figures tore open the cruelest truth: over a decade investing $2.3 billion resulted in $500 million in annual revenue, 361 employees, a net loss of $38 million per month on average, and cash flow that could last only 1.7 months…

This is not the growing pains of a tech enterprise. It is the complete collapse of a technology ideal built on capital—shattered by the iron law of commercialization. XREAL’s IPO prospectus is a “critical condition notice” exclusively for XREAL itself: either the IPO succeeds to keep the company alive, or the decade of effort turns to nothing, with no middle path.

1. $2.2 billion vanishes into thin air—where did the money really go?

Looking back at XREAL’s decade, it is a survival saga of “funding—burning money—raising again.”

From the $3 million angel round in 2017 to the $67.76 million Series D in January 2026, 12 rounds of financing came in waves, only to be consumed to exhaustion at an astonishing speed. Breaking down the $2.2 billion funding flow, every single tranche bears the label of “technical obsession” and “strategic loss of control.”

As a technology-driven company, XREAL poured nearly half of its resources into R&D—practically “built for technology, destined to die for technology.”

The question is: did the $2.2 billion really burn into defensible moats?

The prospectus shows that among the company’s 361 employees, there are 188 R&D personnel, accounting for as much as 51.6%. In 2023–2025, R&D spending was $216 million, $204 million, and $183 million respectively. Over three years, R&D totaled more than $600 million, accounting for over 40% of total revenue.

This money went into three major areas: the self-developed X-Prism optical engine, the X1 on-device coprocessor, and the NebulaOS operating system, with a cumulative 481 patent applications. It looks like it builds a technological moat, but in reality it falls into the trap of “high investment and low output.”

For example, in 2023, R&D expenses accounted for 55.3% of revenue. In other words: earning $1 requires spending $0.50 to develop R&D. And this still doesn’t include personnel, hardware costs, or selling expenses. Its gross margin from core business cannot cover fixed costs at all, and R&D investment far exceeds the capacity that revenue can support.

Actually, if it were only a matter of spending more money, it might be tolerable. The key is that even after spending all that money, the technological moat it produced is not truly unassailable: the core waveguide and Micro-OLED solutions converge with companies such as 雷鸟 and Rokid. Patent counts are high but quality is questionable. In February 2026, it even got pulled into core patent litigation with Viture, facing risks of product bans and compensation;

And to create the so-called technological moat, it chose: a self-developed route with extremely low cost performance—building an optical factory in-house and custom chips. Costs are 2–3 times those of supply-chain procurement. Moreover, the AR industry scale is still small, so it cannot spread production capacity costs at all.

XREAL used “technical ideals” to mask a harsh reality: $200 million per year in R&D investment did not bring monopolistic advantage—it only brought continuous bleeding.

2. Number one worldwide, 71% overseas share—can’t afford 361 employees?

To build the halo of “global #1,” XREAL aggressively pushed globalization, turning selling expenses into the second-largest money-splattering machine.

In 2023–2025, selling and distribution expenses were $210 million, $140 million, and $130 million respectively. Over three years, they totaled more than $480 million. The company built a sales network covering 40 countries. Overseas revenue share rose from 58% to 71%. It looks like internationalization succeeded, but in fact it is stuck in the trap of “high revenue, low profit.”

Taking the European and U.S. markets as an example, channel commissions and marketing expenses are 3–5 times those in China, but sales in a single market are only a few thousand units. There is simply no meaning in spending huge sums to open the U.S. market.

Moreover, despite the “global #1” brand, the main One-series products have an average price of RMB 3196, which is a niche high-end product. It relies on geeks and overseas enthusiasts, and penetration in the mass market is still below 1%.

To shape the image of global #1, XREAL heavily spent on ads across various platforms and numerous tech media. However, the results have been minimal. Its sales growth has effectively stalled. Total units sold in 2023–2025 were 137.2k, 124.9k, and 133.7k respectively.

More ironically, the overseas revenue of 71% still can’t support the cost structure of 361 employees at home.

In 2025, XREAL revenue was $516 million. On an adjusted basis, net losses were $250 million. Every dollar earned overseas was consumed almost entirely by the costs of global expansion—leaving only a story of “internationalization success” and a large loss figure on the books.

3. Diseconomies of scale: 27% market share, yet can’t escape “the more you sell, the more you lose”

XREAL’s biggest paradox is that “global #1” can’t even keep itself alive.

Based on sales revenue, XREAL holds the global #1 position with a 27% market share. But its annual unit sales are only about 130k, far from the scale threshold needed for hardware profitability.

Compared with the smartphone industry, annual sales in the tens of millions are required to break even; in the AR industry, due to high costs and a small market, at least 137.2k units per year are needed to profit. XREAL’s scale is only 1/10 of that threshold.

This “diseconomy of scale” shows up in the cost structure: in 2025, the gross margin rose to 35.2%, but fixed costs (R&D + selling + general and administrative) exceeded $310 million. Gross profit was only $181 million. Gross profit cannot cover fixed costs, so the more it sells, the more it loses.

At the same time, inventory becomes an implicit “cash-burn point”—in 2025, inventory increased sharply, causing net operating cash flow to be negative $203 million, directly dropping book cash from $260 million to $64 million.

A 27% market share is only “the #1 position in a niche market.”

For all of 2025, domestic consumer-grade AR units were only 484k. XREAL’s share is under 20%. The industry’s total volume is too small, so even a higher share cannot support a full commercialization closed loop.

4. Profitability is out of reach: only $64 million on the books, already in a death countdown

XREAL’s decade of development relies entirely on the “funding—burning money” cycle and has never had the ability to generate cash internally.

From 2017 to 2026, the company almost needs one round of financing every 1–2 years: $125 million Series C in 2021, $60 million Series C+ in 2024, RMB 200 million on May 2025, and $67.76 million Series D in January 2026. Every time the financing hits the account, it is quickly consumed by R&D, marketing, and inventory use, leaving cash on the books for a long time “below the warning line.”

More fatally, the company has never formulated a clear “bleeding-stop plan.” The prospectus does not disclose any timeline for reaching profitability, cost-control targets, or strategic contraction plans. Management remains immersed in the illusion of “technological leadership, global #1,” using investors’ capital to satisfy its own technical ideals, but never being accountable to investors.

As disclosed in the prospectus, XREAL has累计 raised about $2.3 billion, with cumulative losses of over $2 billion. Ninety percent of the financing has turned to ash, leaving only $64 million in cash to limp along.

Strictly speaking: XREAL’s life-and-death countdown has already begun.

Based on the net loss of $456 million in 2025, average monthly losses are $38 million. With $64 million, it can only support about 1.7 months of operations. And its: short-term liabilities such as accounts payable, employee compensation, and R&D prepayments exceed $300 million. The cash-to-short-debt ratio is only 0.21, far below the safety line. Once a supplier accelerates payment collection or financing is interrupted, the capital chain breaks immediately.

Even if you include the RMB 460 million Series D financing in January 2026, it would still support only about one year. The IPO is the only lifeline; if it fails, everything goes to zero.

5. Threefold death risks: if the listing fails, it’s a sheer cliff

Besides the cash-flow crisis, XREAL faces three deadly risks. Each one could become “the final straw.”

First death risk: the risk of IPO failure

The Hong Kong market is extremely cautious in valuing loss-making hardware companies. XREAL’s current valuation is $833 million (about RMB 5.7 billion). Yet annual revenue is only $500 million and cumulative losses are $2 billion, creating a severe disconnect between valuation and fundamentals. If the IPO breaks down on pricing, fundraising is insufficient, or after listing the stock price plunges, it will not only fail to keep the company alive—it could also trigger shareholder sell-offs, a crisis of supplier confidence, and a direct shock.

Second death risk: industry competition asphyxiation

On the Chinese mainland, Lei Niao Innovation overtook XREAL with a 35.4% market share, while Rokid’s share rapidly rose to 15.3%. Screenless AI glasses and all-in-one AR glasses are growing much faster than XREAL’s main split-type products. Apple’s Vision Pro is broadly rolled out across 2026. With Meta and Microsoft accelerating their setups, XREAL’s technological and pricing advantages are quickly disappearing.

Third death risk: erroneous strategic positioning

XREAL is stuck on “high-end consumer-grade AR,” but consumer-grade AR will only explode at least in 3–5 years—right now it’s still “a geek toy.” Compared with Magic Leap, which burned through $2.6 billion and then pivoted to a 2B survival strategy, and Microsoft HoloLens, which abandoned consumer-grade and focused on industrial use, XREAL insists on the To C route. In essence, it is using the future track to bet on survival today—completely mismatched with the industry stage.

Facing such a dilemma, even XREAL’s豪华 shareholders group is now in “no clear way forward.”

Institutions such as Alibaba, Kuaishou, Sequoia, Hillhouse, and Shunwei have collectively invested more than $2 billion, and their book losses are now over 70%. For financial investors, listing is the only exit channel. For strategic investors, if XREAL goes bankrupt, not only does the investment go to zero, it also means missing opportunities to build out the AR ecosystem.

But capital’s patience has run out. After the Series D round in January 2026, the company promptly filed its prospectus—clearly IPO is being pushed under capital pressure, forced by necessity. If it burns on, the $2.3 billion will be彻底 to zero. Even if the IPO valuation is discounted, it may still recoup part of the costs.

For XREAL’s management, this is “being pushed to list by capital.” It is not because it meets listing conditions, but because if it doesn’t list now, it will die.

6. REAL is not a one-off case—it’s a snapshot of the “money-burning bubble” across the global AR industry

Over the past decade, total AR-track financing has exceeded $124.9k. Magic Leap burned through $2.6 billion before ultimately pivoting to 2B and laying off 90% of staff; Microsoft HoloLens put in over $10 billion, abandoned consumer-grade and focused on defense and industrial; Meta’s Ray-Ban investment exceeded $5 billion and is still in a loss-making state.

The technology is ahead of the market, costs are high, and the business model has not been established. Consumer-grade AR is still in the “early exploration stage.” User demand is not yet formed, there is no content ecosystem, and prices remain high. Everyone is “burning money to pave the way,” and none are truly profitable.

But XREAL’s tragedy is even worse: other players pivoted in time, contracted to stop bleeding, while only XREAL stayed immersed in the illusion of “global #1,” continuously burning money and refusing to compromise, until it reached the brink of a broken capital chain.

This shows a brutal commercial truth: technology ideals can’t be eaten. “Global #1” can’t be spent like money. Without cash flow, without profitability, without self-sustaining cash generation, no matter how glamorous the halo is, it is still a bubble.

Ten years ago, Xu Chi started a business with the dream that “AR is the next-generation computing platform.” Ten years later, XREAL’s $2.2 billion investment, $2 billion losses, and $64 million cash prove the real cost of that dream coming true.

Now, XREAL has no way out. If the IPO succeeds, it may obtain 3–5 years of lifeline capital, with a chance to adjust strategy and achieve profitability. If the IPO fails, the capital chain breaks immediately: 361 people are laid off, patent assets are auctioned, the $2.3 billion investment is wiped out, and a decade of hard work turns to dust.

For XREAL, this is a “do-or-die battle.” For the entire AR industry, it is the final judgment of the “money-burning model”—capital will no longer blindly pay for “technical ideals,” and the market will no longer buy “false prosperity.” Only companies that truly have commercialization capability and can generate profits will be able to survive.

We do not deny XREAL’s technological contributions to the AR industry, but we must ask: is the $2.2 billion price really worth it? Who will pay for management’s arrogance and loss of control?

Time will give the answer. But for XREAL, time is already running out. Either it lists, or it goes to zero—there is no third option.

大量资讯、精准解读，尽在新浪财经APP