Morgan Stanley, in a March 31, 2026 update following its “China Summit 2026,” argues that Alibaba Group Holding is setting up a near-term cloud re-acceleration via price increases—while insisting its longer-term profitability framework remains intact. The note matters because it draws a clean line between what’s mechanical (pricing) and what’s harder (supply, AI productization, and consumer monetization), at a time when investors want proof that China’s platform giants can turn AI ambition into durable cash flow.

Morgan Stanley reiterates its Overweight rating and a US$180 price target, with the shares at US$121.98 as of March 30, 2026.

A Cloud Price Hike: Growth First, Margins “Still 20%”

Morgan Stanley’s core claim is blunt: “Cloud price hike will drive near-term cloud growth while long-term cloud margins of 20% remain intact.” The bank expects the hike to show up more clearly in results from the June quarter onward, since the increase “will be effective in April.”

Crucially, the report frames the move less as opportunistic pricing power and more as cost pass-through. The price hike, it says, “was mainly driven by component cost increases.” That detail may disappoint anyone looking for a triumphant demand narrative—but it also reduces the risk that the higher prices collapse under competitive pressure, because the alternative would be selling below rising input costs.

Morgan Stanley also sketches an aggressive scale ambition—one that’s easy to tout and hard to deliver. It points to “Cloud 5-year rev US$100bn implied >44% CAGR,” then flags the constraint: “key hurdle from growing faster is supply,” specifically “production capacity from fab.”

In other words, the bull case isn’t merely “cloud rebounds.” It’s: cloud rebounds and Alibaba can physically build enough compute to meet demand.

AI Strategy: MaaS as the Higher-Margin Endgame

The report repeatedly returns to AI as the center of gravity for both investment and monetization. “AI remains the highest priority for investment,” Morgan Stanley writes, including both capex and opex. It also offers a broader framing: “AI total addressable market (TAM) could reach US$6 trillion, from roughly US$3 trillion today, within a US$20 trillion GDP context.”

But the more actionable datapoint is internal mix ambition. Morgan Stanley notes Alibaba has a “very aggressive internal target of MaaS, which could be more than 50% of cloud revenue.” That matters because “MaaS has higher margin than IaaS in the longer term,” supporting the bank’s assertion that the “long-term cloud margin 20% target” remains intact.

This is where the narrative tightens: pricing may lift reported growth in the near term, but mix shift toward higher-value AI services is what defends margins when competition inevitably responds.

The Hardware Angle: T-Head, Integration, and a Supply-Led Ceiling

Morgan Stanley’s discussion of Alibaba’s chip and server stack reads like a reminder that AI in 2026 is still a hardware business pretending to be software.

It highlights “T-head” serving “60% of external cloud customers,” while emphasizing Alibaba is “not selling PPU directly to external customers,” instead embedding compute “into server and software (integrated services) sold to external customers through public and hybrid cloud.”

The bank adds comparative performance color: the chips are “mainly used for inference but can also do some training,” “better than H20 but not as good as H200.” And on cost: “Cost is much lower than NVDA chips and even lower than other local chips.”

The implication is double-edged. Integration can create stickiness and protect economics—yet the same note repeats that “the key hurdle is production capacity.” Cheap compute doesn’t matter if you can’t ship it.

Qwen Goes Consumer: DAU Spikes, Monetization Still a Question

Alibaba’s AI story is not just enterprise. Morgan Stanley says it “will focus on ramping up Qwen App as building a 2C agent is strategically important for future monetization.” The bank contrasts Alibaba’s strength in “2B” with a peer that has “much higher 2C MAU/DAU,” implying Alibaba is playing catch-up in consumer mindshare.

Still, the Qwen data points are eye-catching: “Achieved 44mn DAU two months (peak 77mn DAU) after CNY promotions,” and “Qwen now becomes a widely recognized brand and is connected to BABA ecosystem.”

What’s missing—because it’s still missing in the business—is a stable revenue model. Morgan Stanley notes Alibaba “still want[s] a similar take rate… i.e. currently combined take rate between ads and commissions is between mid to high single digit.” That’s not a promise; it’s a direction of travel.

Quick Commerce: Losses Narrow, But the Subsidy Trap Looms

On quick commerce, Morgan Stanley says “QC loss is narrowing QoQ for both BABA and Meituan.” It expects “Mar Q 26 < Dec Q 25,” with March-quarter losses “lower than Dec Q.” The plan: “continue to focus on driving GMV growth and market share via increasing AOV.”

Yet the report also signals a potential easing in promotional intensity: while “all others loss will be wider in Mar Q vs Dec Q,” it “probably won’t have similar level of CNY promotions going forward.” Translation: the DAU spikes and momentum may have been bought, and management appears wary of paying that bill repeatedly.

Overseas Cloud: Small Base, Bigger Narrative

Morgan Stanley notes overseas revenue is “<20% of Alicloud revenue” but “will increase going forward.” It’s a modest line, but strategically loaded: international growth can diversify regulatory and macro exposure, while also testing whether Alibaba’s AI and cloud stack travels beyond its home market.

Bottom Line

Morgan Stanley’s thesis effectively separates optics from execution. The price hike may deliver near-term cloud growth “mechanically,” but the long-term debate is about capacity, mix shift toward MaaS, and whether Qwen can become a consumer agent with monetization that’s more than “mid to high single digit” take rates. In 2026, the market isn’t short of AI stories. It’s short of AI businesses that can scale without hitting the wall of supply—or the trap of subsidies.

Related Coverage:

Alibaba Cloud Raises Prices by 34% to Address AI Infrastructure Shortage and Free Up Resources for MaaS Expansion

Alibaba’s Damo Academy Unveils XuanTie C950, Pushing RISC-V Into Server-Class AI Computing