Everything About AI
NextFin News - Alibaba Cloud is adding data centers in Europe and the Middle East as part of a broader push to capture global demand for AI infrastructure, a move that puts the Chinese technology giant on a collision course with Amazon Web Services, Microsoft Azure and Google Cloud in markets where data-sovereignty rules are rewriting the competitive map. The expansion, which follows the company's RMB 380 billion (about US$53 billion) commitment to AI and cloud infrastructure announced in 2025, lifts Alibaba Cloud's global footprint to 105 availability zones across 32 regions, up from 91 availability zones across 29 regions at the time the overseas buildout was unveiled at the Apsara Conference in September 2025.
The new facilities in France and the Middle East are more than a capacity story. They are the leading edge of a strategy to export Alibaba's AI stack - the Qwen family of large language models, the Model Studio development platform, and a growing suite of agentic AI services - to customers who increasingly want their data processed inside their own borders. The question is whether Alibaba can turn infrastructure into revenue fast enough to justify the spending, at a moment when its core e-commerce business is under pressure and its profit margins are already absorbing the cost of the AI buildout. Shares rose in pre-market trading in New York on September 22, with the American depositary receipt changing hands at $120.19, up $4.44, or 3.84 percent, from the prior session close of $115.75, according to pre-market data.
The Buildout: What Alibaba Is Actually Doing
Alibaba Cloud's latest expansion adds availability zones in France, Japan, Malaysia and Mexico, with the Paris region - featuring two availability zones - becoming the company's third European hub after Germany and the United Kingdom. In the Middle East, Alibaba is deepening its Dubai presence, where it has operated since 2016 and now runs two availability zones; a second data center in the emirate was launched to improve cloud resilience and disaster recovery for regional customers. In Malaysia, the new Johor region brings two additional data centers, taking the company's total footprint there to five facilities - its largest infrastructure presence in Southeast Asia.
"The expansion of our cloud infrastructure across continents reinforces our ongoing commitment to empowering global businesses with sovereign, secure, and intelligent solutions," said Dr. Feifei Li, Chief Technology Officer and President of International Business at Alibaba Cloud. "Alongside the introduction of our agentic AI services to various markets, this move aligns with our broader strategy to bring our full-stack AI+Cloud ecosystem to global customers as we enter the agentic era."
The timing is deliberate. Alibaba Cloud plans to launch a suite of agentic AI services in Malaysia and European markets in the second half of 2026, and has already rolled out AI-native database and data analytics products - including Data Agent for Analytics, Meta, DAS and DataBridge - in Europe, Japan and Malaysia. The company says its DataWorks Data Agent can compress traditional data-warehouse development cycles from days to minutes, a pitch aimed squarely at enterprises weighing where to run AI workloads. Model Studio, the company's one-stop enterprise-grade AI development platform, has been made available in Japan as part of the same push.
Underpinning the expansion is the capital commitment Eddie Wu, Alibaba's chief executive and Alibaba Cloud's chairman, unveiled at the Apsara Conference in Hangzhou in September 2025. He described it as the company's biggest-ever overseas investment in AI infrastructure, with new cloud regions planned for Brazil, France and the Netherlands, plus capacity additions in Mexico, Japan, South Korea, Malaysia and Dubai. The blueprint announced then is now being executed region by region, and the pace of delivery - from 91 to 105 availability zones in roughly a year - signals that this is a program with momentum, not a press-release pledge.
Why This Is a Structural Shift, Not Just a Capex Cycle
The surface read of this news is simple: AI demand is booming, so Alibaba is building more data centers. That is true as far as it goes, but it misses the mechanism that makes this expansion structurally different from a normal capacity cycle.
The first driver is sovereignty. European regulators and enterprises have spent the past decade building a wall of rules around data - the GDPR, the EU AI Act, and a growing political insistence that sensitive workloads stay on European soil. Alibaba's France-specific messaging is explicit on this point: "The expansion of our cloud infrastructure into France reinforces our ongoing commitment to empowering European businesses with sovereign, secure, and intelligent solutions," Dr. Li said when the Paris region launched in June 2026. A Chinese hyperscaler offering locally hosted, regulation-aligned infrastructure in Paris is a fundamentally different product than one offering the same compute from Hangzhou. The data center is the compliance wrapper that makes the AI stack sellable.
The second driver is the shift from model competition to infrastructure competition. For the first two years of the generative-AI boom, the battlefield was foundation models. Alibaba's Qwen family has earned a credible place in that race, with open-weight releases that compete directly with Western counterparts. But as the market moves into the "agentic era" that Alibaba's executives describe, the scarce resource is no longer just model quality - it is low-latency, locally available compute that can run enterprise agents at scale. Infrastructure becomes the moat, and every region Alibaba opens is a claim on future agent workloads before AWS, Azure and Google Cloud fully lock them down.
This is why the cyclical-versus-structural question resolves toward structural. A cyclical capex wave would mean building capacity to meet today's demand spike, then watching utilization fall when the hype cools. What Alibaba is doing is different: it is buying geographic positions in markets where regulatory barriers, once established, do not reverse. If Alibaba can embed itself in European and Middle Eastern enterprise stacks now, the switching costs and compliance lock-in will outlast any single AI spending cycle. The buildout is a land grab, and land grabs are not mean-reverting.
There is a third driver worth naming: the Middle East is becoming a genuine alternative pole of AI infrastructure investment, and Alibaba is positioning itself inside it. Regional governments are pouring capital into data-center capacity - Saudi Arabia alone is reported to have 33 data centers operating with 42 more in development, adding roughly 2.2 gigawatts of IT load. For Alibaba, Dubai is not just a sales office for the Gulf; it is a foothold in a region that is actively trying to become a non-Western compute hub. That alignment of interests - a Chinese provider seeking geographic diversification and a Gulf state seeking infrastructure sovereignty - is the structural logic behind the Dubai expansion.
The Money: Growth Is Real, But So Is the Bill
The financial evidence supports both the opportunity and the risk. In the quarter ended June 30, 2026, revenue from Alibaba's AI Cloud and Compute Services segment rose 45 percent year over year to 48.44 billion yuan (about $7.14 billion). AI-related product revenue reached 12.38 billion yuan ($1.82 billion), extending a streak of triple-digit year-over-year growth to 12 consecutive quarters. The segment's adjusted EBITA more than doubled, up 133 percent to 5.63 billion yuan ($830 million).
But the group-level picture is where the tension shows. Total revenue for the quarter rose 9 percent to 268.95 billion yuan ($39.64 billion), while non-GAAP net income fell 38 percent to 20.72 billion yuan ($3.05 billion), and adjusted EBITA declined 30 percent to 27.33 billion yuan ($4.03 billion). The AI engine is accelerating; the profit line is not. Alibaba's leadership has been clear that more spending is coming - CEO Eddie Wu told analysts that strong demand for AI could lead the company to boost investment in AI infrastructure further.
That creates the central asymmetry investors need to watch. Cloud infrastructure is front-loaded capital with back-loaded returns. The data centers get built and depreciated on day one; the enterprise contracts that justify them can take quarters or years to materialize. For a company already absorbing margin pressure from investments in instant commerce and competition with ByteDance's Douyin and Pinduoduo in e-commerce, the next several quarters will show the cost before they show the payoff.
The market context explains why Alibaba is willing to accept that trade-off. The Chinese market for AI cloud services - computing infrastructure optimized for generative AI tasks - was forecast to more than double in 2025 to 51.8 billion yuan ($7.3 billion), up from 20.83 billion yuan in 2024, and to grow at an annual rate of 26.8 percent from 2025 to 2030, according to research firm Omdia. Alibaba held roughly a third of the China cloud market in the first quarter. But a domestic growth story, however strong, is not enough to justify a valuation rerating when the home market is saturated with competitors and geopolitical constraints limit how much of that growth can flow to shareholders. International expansion is the escape route - and the reason the Europe and Middle East moves matter more than their near-term revenue contribution.
The Counter-Thesis: Too Little, Too Late in a Crowded Market
The strongest case against this strategy is straightforward: Alibaba is late to markets where the incumbents are already entrenched, and the geopolitical overhang of a Chinese provider in Europe may prove insurmountable no matter how many availability zones it opens. AWS, Azure and Google Cloud have been building European capacity for well over a decade, and they carry none of the trust deficit that comes with Beijing-headquartered ownership. European enterprises running sensitive AI workloads may simply prefer the American option, sovereignty rules notwithstanding.
There is also a scale question. Even after this expansion, Alibaba Cloud's international footprint remains small relative to its domestic base and relative to the American hyperscalers' global networks. If the overseas regions fail to attract marquee enterprise customers quickly, the company risks being left with expensive, underutilized capacity - precisely the outcome that would turn a structural land grab into a cyclical capex mistake. The profit compression already visible in the latest quarter - a 38 percent decline in non-GAAP net income - is the early warning sign that the spending side of the equation is moving faster than the revenue side.
This counter-thesis is not hypothetical. Alibaba's international cloud revenue remains a modest share of group revenue, and the company has not broken out region-level performance for the new European and Middle Eastern sites. The expansion is a statement of intent; it is not yet a statement of results.
Here is the signal that would prove the bullish structural read wrong: if Alibaba's international cloud revenue share does not rise measurably over the next two to three quarters - if it stays below roughly 10 percent of group revenue - or if the Paris region fails to land named enterprise customers by mid-2027, then the land-grab thesis weakens materially. At that point, the expansion would look less like a structural foothold and more like expensive capacity chasing a demand curve that the incumbents already own.
What Comes Next: Three Time Horizons
Short term (next 1-2 quarters): sentiment and spending. Expect continued margin pressure as the capex hits the income statement ahead of revenue recognition. The stock will trade on whether investors reward the AI growth rate - 45 percent segment growth, 12 quarters of triple-digit AI revenue growth - or punish the 38 percent decline in non-GAAP net income. Volatility is the likely baseline, and the shares remain well below their 52-week high, reflecting the market's skepticism about the payoff timeline.
Medium term (6-18 months): customer proof points. The key data points are named enterprise wins in the Paris and Dubai regions, and the uptake of the agentic AI services launching in Europe and Malaysia in the second half of 2026. This is where the strategy moves from announcement to evidence. Investors should also watch whether the company raises its capital-spending guidance again, as Eddie Wu has signaled it might.
Long term (structural): the sovereignty premium. If data-localization rules tighten further - a plausible direction given the EU's regulatory trajectory - Alibaba's early local presence could become a durable competitive asset. The company that owns compliant infrastructure in a walled garden captures the workloads that cannot leave it.
The base case is that Alibaba carves out a meaningful but minority position in European and Middle Eastern AI infrastructure, growing steadily without threatening the American hyperscalers' dominance. The upside case is that sovereignty rules accelerate faster than expected, turning Alibaba's early local buildout into a scarce, high-margin asset. The downside case is that geopolitical friction and incumbent lock-in leave the new regions underutilized, turning the RMB 380 billion commitment into a drag on returns.
Alibaba's data-center expansion is a bet that the future of AI is local - that the winning infrastructure will be the one sitting inside the regulatory border, not the one with the most raw compute. If he is right, the Paris and Dubai facilities are the first stones of a durable moat. If he is wrong, they are expensive monuments to a land rush that never happened.
Explore more exclusive insights at nextfin.ai.
