BABA logo

Article Brief

Key Takeaways

5 Points30s Read

  1. The moveAlibaba’s US shares rose about 11% on July 8 to $108.98 — the best day in ten months — with Hong Kong up 12.2%, the strongest session of the year.
  2. The contradictionThe same week, Morgan Stanley, Citi, Daiwa and HSBC all cut price targets on heavy AI spending and soft consumption. All four kept Buy ratings.
  3. The inflectionUBS sees June-quarter cloud revenue up ~45% and total growth accelerating to 9% from 3% — the two-year growth stall ending.
  4. The noise it beatThe rally came through an Anthropic complaint to the Senate alleging a 28.8-million-exchange distillation campaign (which Alibaba denies) and an unresolved Pentagon-list fight.
  5. The dateJune-quarter earnings land August 28. Even the lowest cut target ($170) sits 56% above the current price — the dispute is timing, not destination.

This article is for information only and is not investment advice. Prices reflect the July 8, 2026 close; allegations described are unproven and denied by the company. Verify live data before making any investment decision.

Two things happened to Alibaba this week that are not supposed to happen together. On Wednesday, July 8, its US-listed shares rose about 11% to $108.98 — their best single day in ten months — while the Hong Kong line jumped 12.2%, its strongest session of the year. The same week, four major banks cut their Alibaba price targets. The tape and the sell-side looked at the same company and moved in opposite directions, and the interesting part is that neither is obviously wrong. They are underwriting different years.

The quarter is inflecting — that’s what actually moved

The rally was not a headline pop. It was a re-rating of the June quarter before Alibaba reports it. UBS analyst Kenneth Fong wrote that cloud revenue likely grew about 45% in the quarter, with annual recurring revenue from AI model services reaching roughly 10 billion yuan, about $1.5 billion. UBS and Jefferies both project total revenue growth accelerating to 9% from 3% the prior quarter — and that single statistic explains the violence of the move. Alibaba spent two years as the cheap Chinese megacap whose growth had stalled; a three-fold acceleration in the top line, alongside narrowing food-delivery losses and improving e-commerce margins, is the stall ending.

The cloud engine behind that acceleration is not a marginal player. Alibaba Cloud holds a 40.1% share of China’s full-stack AI cloud market, more than its competitors combined, per Frost & Sullivan data. When the largest AI infrastructure provider in the world’s second-largest economy accelerates, the stock’s starting valuation — it entered the week around $98, down more than 30% for the year — does a lot of the rest.

It rallied through the worst news cycle of its year

What makes Wednesday’s move genuinely informative is what it rallied through. In late June, Anthropic told the US Senate Banking Committee that operators affiliated with Alibaba ran roughly 25,000 fraudulent accounts and 28.8 million exchanges with its Claude models between April 22 and June 5 — an alleged distillation campaign bigger than the ones attributed to DeepSeek, Moonshot and MiniMax combined, which TECHi covered when those first scraping accusations surfaced in February. Alibaba denies the allegations, and the figures have not been independently verified. Senators Bill Hagerty and Andy Kim are nonetheless moving a defense-bill amendment that would sanction Chinese AI firms for exactly this behavior.

Washington’s other front is older. Alibaba sits on the Pentagon’s Section 1260H list of alleged Chinese military companies — a roster that grew to 188 firms in June, from around 130 — and a related rule had stripped the company of its Washington lobbyists until a federal judge paused it this week, with a hearing still ahead. A stock that puts up its best day in ten months against that backdrop is telling you how the market currently scores those risks: as headline volatility, not as impairments to cloud cash flow. That scoring could prove wrong — the sanctions amendment is live legislation, not commentary — but it is the message in the price.

Why the banks cut anyway

The same week, Morgan Stanley trimmed its target to $180 from $190, Citi to $192 from $208, Daiwa to $175 from $200 and HSBC to $170 from $176. Every one of them kept a Buy rating. The stated logic is consistent: Alibaba’s AI infrastructure spending is heavy enough to compress near-term earnings, and Chinese consumer spending remains soft enough to drag the advertising and commerce lines that still pay for everything.

Look at the arithmetic, though. Even the lowest of the reduced targets, HSBC’s $170, sits 56% above Wednesday’s close. The cuts model the next two to four quarters of margin pain; the ratings model the destination. When analysts lower the number and keep the Buy, the disagreement with the market is not about whether the AI build-out works — it is about who has to sit through the spending phase. The buyers who showed up Wednesday, after institutional money had been quietly accumulating near the $92 lows, decided the spending phase is the entry.

Beijing just loosened the compute constraint

There is a supply-side reason to take the cloud acceleration seriously. The Information reported this week that Chinese authorities have signaled Alibaba, ByteDance and DeepSeek will be allowed to buy Nvidia’s H200 chips — fewer than 200,000 units in total, less than half what the companies requested, but movement after months in which Beijing withheld approvals even though Washington had already licensed roughly ten Chinese firms to buy. TECHi readers have watched this door swing before: the stock rallied on an H200 breakthrough in January, then Beijing sat on its hands while pushing domestic silicon.

For a cloud business growing 45% against capacity constraints, chips are the binding input. Alibaba is attacking the constraint from both ends — its in-house T-Head silicon featured prominently in this week’s earnings optimism, and the company has been pouring capital into the build-out since its $53 billion AI commitment last fall. H200 access, even rationed, raises the ceiling on the one business line that has become China’s AI backbone. It also cuts the other way: fewer than half the requested chips is Beijing reminding everyone — including shareholders — that it controls the tap.

What August 28 settles

Alibaba reports June-quarter results on August 28, before the US open. The setup is now unusually clean:

  • The cloud print versus the 45% bar. The sell-side just raised the whisper number while cutting targets. Meeting 45% validates the rally; missing it hands the tape back to the target-cutters.
  • The margin disclosure. How much of the cloud acceleration is being bought with capex and model-price subsidies determines whether the banks’ near-term caution was right.
  • The Washington calendar. The defense-bill amendment on AI distillation and the Pentagon-list hearing both land while the earnings narrative is still forming.
  • H200 logistics. Approval reports are not delivery schedules. Actual chip volumes, and any strings attached, will show up in cloud guidance before they show up in filings.

The banks are pricing what Alibaba’s AI build-out costs. Wednesday’s buyers are pricing what it buys. August 28 is the first date those two prices have to reconcile — and $109 versus a $170 floor on the Street’s own reduced targets says the market still hasn’t decided which side it believes.