Introduction: With six consecutive years of losses, what does UCloud have to win the battle of 'cloud-based lobster farming'?

From a super AI Agent in the geek circle to a nationwide 'lobster farming' craze, OpenClaw achieved a remarkable triple jump in just 10 days—from 'going viral on GitHub → nationwide lobster farming → the strongest main theme in the stock market.' However, after a brief frenzy, this lobster began to leave many users feeling 'betrayed.'

Numerous individual users suffered from the 'lobster' raiding their personal computer data, with their payment accounts being misused for massive spending. More users were deterred by the high costs of 'lobster farming.' This 'Token-gobbling beast' incurred daily expenses of several hundred or even over a thousand yuan, making it unaffordable for engineers, programmers, and other groups requiring substantial Token usage.

Amid this 'lobster craze,' cloud providers undoubtedly emerged as the most benefited group. Relying on the logic of 'computing power + scenarios + commercialization,' 'cloud-based lobster farming' has proven to be the most convenient, secure, and profitable 'lobster farming model.'

Under the 'lobster concept,' cloud providers have seen a surge in both computing power prices and stock prices. For instance, reports from 'Cailian Press' indicate that Tencent Cloud announced a price hike for some models on March 11, with the adjusted prices reaching up to five times the original rates.

In the capital market, the most closely watched 'lobster concept stock,' UCloud-W (688158.SH), saw its stock price rise for five consecutive trading days starting March 4. On March 6, March 9, and March 10, it accumulated a gain of over 42.55%, with its total market value surpassing 20 billion yuan.

After consecutive surges, the 'lobster concept stocks' experienced a pullback on March 11. Kunlun Tech, Shunwang Technology, Meig Smart, and MINIMAX-WP all declined to varying degrees, with some listed companies issuing risk warnings. UCloud admitted in its risk warning that its 'lightweight cloud host product equipped with the OpenClaw image has not yet formed a scalable product system.'

However, UCloud's stock price did not plummet but only slightly decreased by 0.76%, with the day's trading volume still reaching a high of 5.546 billion yuan and a turnover rate of 26.95%.

Yet, behind the capital's frenzy for the 'lobster concept,' it is essential to see the true face of this 'neutral cloud provider.' Facing major players like Alibaba Cloud, Tencent Cloud, and Volcano Engine, UCloud's market appears very 'niche.'

However, it is worth noting that UCloud still has a trump card—'computing-power-electricity synergy.' However, this trump card has not yet fully taken shape.

UCloud's Shanghai 'Qingpu Data Center Project (Phase I),' positioned as an 'eastern data' intelligent computing center, was originally planned to be completed on January 24, 2026, but recently announced a two-year delay until January 24, 2028. Amid the AI boom, whether the project, upon completion in the next two years, will catch the industry's peak or 'peak at the wrong time' may be another yardstick to verify UCloud's 'mettle.'

01 Six Consecutive Years of Losses: This 'Lobster Stock' Hits the Brakes

UCloud's rise as a 'lobster concept stock' is related to its interaction with investors on March 2.

On that day, UCloud replied to investors, stating that it was among the first in the industry to deploy OpenClaw-related images in late January, with relevant services now available in multiple overseas nodes, including the United States, Singapore, and Japan, helping users quickly build a '7×24-hour personal AI super assistant.'

Subsequently, the 'lobster farming' craze quickly went viral. Since UCloud released information about OpenClaw being 'ready to use' and being among the first to achieve 'cloud-based lobster farming' before the craze, its stock price rose continuously, with investors reaping substantial gains.

On March 9, UCloud's stock price hit a 20cm daily limit, and on March 10, it rose another 13.68% from the previous day, accumulating a gain of over 36% in just two days.

The capital inflows and outflows during these two days were frequent. On March 9, the stock's daily trading volume was 3.108 billion yuan, and on March 10, it reached a record high of 7.625 billion yuan since its listing, with a turnover rate as high as 36.8%.

As the stock price deviation increased, UCloud quickly stepped in to 'cool down,' but this statement slightly differed from its reply to investors on March 2.

At that time, UCloud's reply to investors included information such as 'promoting the widespread adoption of AI technology,' 'lowering the barrier to using cutting-edge AI technology,' and 'continuously optimizing support for AI Agents,' which obviously gave investors a positive impression of UCloud's AI commercialization capabilities.

However, on March 11, UCloud stated in an announcement that its lightweight cloud host product equipped with the OpenClaw image has not yet formed a scalable product system, and technological iteration and commercialization progress may fall short of expectations.

UCloud also warned that currently, the relevant business has not yet formed a stable and sustainable source of income, with an extremely low revenue share. The future revenue scale, profitability, and cash flow contribution of the products are highly uncertain, with limited short-term impact on the company's overall operating performance.

So, what is UCloud's fundamental situation?

UCloud's recently released 2025 annual performance preview indicates that it expects a net loss attributable to shareholders of 76.6832 million yuan for 2025, with a net loss attributable to shareholders after deducting non-recurring items of 146 million yuan.

This marks UCloud's sixth consecutive year of losses since 2020, with cumulative losses nearing 2 billion yuan. From 2021 to 2024, UCloud's revenue also declined continuously, from 2.901 billion yuan to 1.503 billion yuan, a decrease of over 48% in four years.

In 2025, UCloud's revenue was approximately 1.7 billion yuan, a year-on-year increase of 13.11%, finally halting the revenue decline. This is attributed to the rapid development of the AI industry.

UCloud stated that the revenue increase in 2025 was mainly due to continuous investment in the AIGC field, leading to sustained growth in intelligent computing product revenue. Annual AI-related revenue increased by over 40% year-on-year, accounting for more than 40% of total operating revenue.

Before this, UCloud's revenue primarily relied on public cloud, private cloud, and hybrid cloud (the 'three clouds' series), with public cloud being the largest contributor.

In 2021, public cloud revenue once accounted for over 75% of UCloud's total revenue. After 2022, public cloud revenue began to decline, but as of 2024, it still accounted for over 50% of total revenue.

Notably, compared to private cloud, public cloud has lower gross margins. For example, in 2023 and 2024, UCloud's private cloud business gross margins were 36.31% and 46.84%, respectively, while public cloud gross margins were only 12.06% and 21.04%.

02 Cloud Pricing War: Can UCloud Seize the Opportunity?

UCloud's revenue decline and profit losses are also related to the prolonged price war in 'cloud products.'

Since 2018, the prices of UCloud's main products have generally shown a downward trend, with significant price reductions. In 2025, UCloud believed that competition in the future cloud computing market may remain fierce, with major cloud providers likely to continue cutting prices to capture market share. UCloud may also continue to reduce prices in the coming years.

At that time, UCloud also proposed to 'vigorously develop' its private cloud business. From an industry perspective, given the future trend of 'AI Agents,' private cloud's emphasis on data security and privacy is indeed worth pursuing for UCloud, with higher profit margins being its core value.

However, UCloud soon encountered the Agent open-source boom sparked by OpenClaw, which further increased the already tight demand for computing power. Cloud providers collectively raised prices, and UCloud naturally participated.

In late January 2026, Alibaba Cloud and Tencent Cloud announced tiered price hikes, while Amazon AWS also raised EC2 machine learning computing block prices by about 15% in January.

In mid-February, UCloud followed suit, announcing price increases for some cloud servers, object storage, and CDN services starting March 1, 2026, with increases ranging from 5% to 12%, affecting core regional nodes in North China, East China, and South China.

Also in February, Google Cloud announced price hikes starting May 2026, with some products increasing by over 100%.

On March 11, 'Cailian Press' reported that Tencent Cloud announced adjustments to the pricing strategies of some models starting March 13, 2026, ending the free public testing of models like GLM5, MiniMax2.5, and Kimi2.5, and transitioning to a pay-per-use model.

Moreover, Tencent's Hunyuan series model, TencentHY2.0Instruct, saw its input price rise from 0.0008 yuan per thousand tokens to 0.004505 yuan, a 463% increase. The output price also rose from 0.002 yuan to 0.01113 yuan, an increase of over 456%.

TencentHY2.0Think's input price increased from 0.001 yuan to 0.0053 yuan, and the output price from 0.004 yuan to 0.0212 yuan, with the adjusted prices being five times the original rates.

In the tech circle, the collective entry of domestic and foreign cloud providers into the 'cloud computing price hike era' is not complex logic, mainly driven by AI-driven surges in data center power demand, historically low memory chip inventories, and expected annual memory price increases, leading to a supply-demand imbalance in computing power and inevitable price hikes.

The Agent open-source boom sparked by OpenClaw further increased demand for computing power, driving up prices collectively among cloud providers.

UCloud has been relatively restrained in this round of price hikes, and whether it can attract more users remains to be seen.

03 'Strong Clouds Surround': UCloud's 'Computing-Power-Electricity' Trump Card Remains Unplayed

In the tech circle, compared to major cloud providers like Alibaba Cloud, Tencent Cloud, Volcano Engine, and Baidu Cloud, UCloud is considered a small to medium-sized cloud provider. After the 'lobster concept' hype fades, whether UCloud can secure a sustained and stable 'lobster cloud' market amid 'strong clouds surrounding' it is a question UCloud needs to seriously consider.

Although UCloud told investors that it was among the first in the industry to achieve OpenClaw cloud deployment, many major cloud providers also deployed it simultaneously in January this year.

For example, JD Cloud officially launched OpenClaw cloud in January, Baidu Intelligent Cloud launched OpenClaw one-click deployment service on February 3, followed by Volcano Engine launching ArkClaw cloud SaaS version and Tencent Lighthouse lightweight cloud launching OpenClaw one-click deployment.

Facing market competitors, UCloud warned about stock price risks, stating that 'several cloud service providers have launched similar products, and the company faces significant market competition. The contribution of relevant products to future performance is highly uncertain.'""The Agent open-source boom sparked by OpenClaw is indeed a significant boon for cloud providers. Previously, many ordinary users had no demand for cloud services, but with the arrival of 'lobsters,' 'cloud-based lobster farming' became the most convenient approach, equivalent to a free 'mass popularization' wave for cloud providers, which was previously unimaginable.

Under the market principle of 'the strong get stronger,' UCloud still faces significant pressure from major cloud providers. Although it has a differentiated advantage in the 'neutral cloud provider' niche, this niche itself has limited market share. UCloud needs more trump cards to secure its position at the table, with 'computing-power-electricity synergy' being one of them.

UCloud's 'computing-power-electricity synergy' business focuses on 'green power for computing power, computing power stabilizes the power grid,' centered around two self-built intelligent computing centers + liquid cooling/energy storage/virtual power plants + ecological cooperation.

UCloud's core project in operation is the Ulanqab Intelligent Computing Center (Western Computing Hub), covering 140,000 square meters with a total investment of 4.8 billion yuan, planning 12,000 cabinets, with over 7,300 already in operation (2,300+ delivered in Phase III). Located in a green power-rich area, it adopts full immersion liquid cooling, with a PUE as low as below 1.1, using direct green power supply + self-provided energy storage, undertaking large-scale model training, and serving well-known domestic companies.

In recent years, UCloud's cabinet rental fees have directly improved its gross margin, with future profitability worth noting.

The Shanghai Qingpu Intelligent Computing Center (Eastern Computing Hub) is positioned for low-latency inference, deploying domestic thousand-card clusters, accessing offshore wind PPA + user-side energy storage, participating in virtual power plant (VPP) scheduling, and obtaining electricity price benefits through peak shaving and valley filling.

The 'Qingpu Data Center Project (Phase I)' officially commenced operations in January 2023 and was originally planned to be completed on January 24, 2026, but recently announced a two-year delay until January 24, 2028.

Additionally, UCloud is advancing the second phase of Building B at the Ulanqab Intelligent Computing Center, continuously expanding cabinet and liquid cooling capacity.

'Computing-power-electricity synergy' is UCloud's moat project, but the key lies in whether it can play this trump card at the most opportune moment. In the rapidly advancing AI wave, 'timing' can also be fleeting.