TLDR
- Alibaba launched the XuanTie C950, a 5nm AI chip built on RISC-V architecture that runs over 3x faster than its predecessor
- Morgan Stanley reiterated an Overweight rating and $180 price target, keeping BABA as a “Top Pick”
- The chip natively supports large AI models including Qwen3 and DeepSeek V3
- Morgan Stanley values Alibaba’s T-Head chip unit at $28B–$86B, or ~$22 per share
- Wall Street consensus is Strong Buy with an average price target of $188.38
Alibaba Group Holding unveiled a new in-house AI chip on Monday, and Morgan Stanley wasted no time reaffirming its bullish stance on the stock.
Alibaba Unveils New Chip Design to Meet Surging Demand for AI
Alibaba Group’s Damo Academy unveiled the XuanTie C950, a RISC-V CPU for agentic AI and inference computing, optimized for cloud use and customizable for clients. The launch advances Alibaba’s all-stack AI ambitions,… pic.twitter.com/3OfkL2TeNx— CN Wire (@Sino_Market) March 24, 2026
The chip is called the XuanTie C950. It uses a 5-nanometer process and is built on open-source RISC-V architecture. It runs more than three times faster than Alibaba’s previous chip and natively supports large AI models including Qwen3 and DeepSeek V3.
The C950 comes out of Alibaba’s T-Head semiconductor unit, which the company has been quietly building out for years. That in-house capability is now a key part of Morgan Stanley’s investment case.
Analyst Gary Yu reiterated an Overweight rating and kept his $180 price target. He said the chip reinforces his view that Alibaba has built strong control over the key layers of AI infrastructure.
BABA was trading at $126.06 at the time of the note, down 7.7% over the prior week.
Alibaba Group Holding Limited, BABA
What the Chip Means for Alibaba’s AI Stack
Morgan Stanley’s argument is straightforward: having its own chips means Alibaba depends less on third-party suppliers. That lowers costs, allows faster scaling when demand rises, and reduces exposure to U.S. export controls.
The chip sits at the base of what Morgan Stanley calls Alibaba’s full AI stack. Above it sits AliCloud for computing power, the open-weight Qwen model family, and consumer and enterprise-facing applications.
Alibaba recently launched Wukong, an AI-native enterprise platform with agentic capabilities. It also introduced the Alibaba Token Hub. Morgan Stanley said both products could help convert AI activity into actual revenue over time.
Valuation and Analyst Consensus
The firm runs a sum-of-the-parts valuation on Alibaba that reaches $245 at the midpoint. T-Head alone accounts for $22 per share of that figure, based on a valuation range of $28 billion to $86 billion for the unit.
BABA trades at a P/E ratio of 21.82 with a market cap of around $281.8 billion. The stock holds more cash than debt on its balance sheet.
Across Wall Street, the consensus rating on BABA is Strong Buy. That’s based on eight Buy ratings and one Hold over the past three months.
The average price target among analysts sits at $188.38, implying around 49% upside from where the stock was trading when Morgan Stanley issued its note.
Alibaba’s most recent quarterly results came in below expectations. Revenue for the December quarter was RMB284.8 billion, up 2% year-over-year, or 9% when stripping out divested businesses. That still missed estimates by 2%.
Following those results, Jefferies cut its target to $212 from $225, and Mizuho trimmed its target to $190 from $195. Both kept their Buy-equivalent ratings. US Tiger Securities went the other way, upgrading BABA to Buy from Hold with a $175 target, citing AI and cloud momentum.
BofA Securities maintained its Buy rating with a $180 price target, also pointing to Alibaba’s AI investment track.







