Alibaba Cloud has issued a price adjustment notice on its official website, drawing significant market attention. Due to the global surge in AI demand and rising supply chain costs, the company is increasing prices for core products such as AI computing power and storage, with the highest increase reaching 34%. Notably, the cost of the in-house developed "Zhenwu" series computing cards by T-Head has risen by 5% to 34%, with the upper limit of the increase even surpassing that of overseas products like NVIDIA's H-series.
An analysis of industry trends, research data, and Alibaba's organizational restructuring reveals that this price adjustment is not an isolated move. Instead, it signifies Alibaba Cloud's strategic shift from being an infrastructure provider to an AI service provider. Morgan Stanley analysts suggest this is not merely a cost pass-through but a crucial trigger to activate its "bull market valuation model."
Global computing power shortages are forcing a restructuring of pricing mechanisms. Since the beginning of this year, cloud providers both internationally and domestically, including Amazon, Google, and Baidu, have successively raised prices for their core cloud products. However, Alibaba Cloud's maximum increase of 34% significantly exceeds that of its peers. This reflects a fundamental reversal in the global supply and demand dynamics for AI computing power. In its latest research report, Morgan Stanley points out that inference demand is replacing training as the primary engine for AI computing power growth. As AI applications shift from training large models to large-scale inference, Token usage is experiencing exponential growth.
According to informed sources, applications involving AI Agents saw rapid growth during the Spring Festival period. Alibaba Cloud's MaaS (Model-as-a-Service) business, "Bailian," recorded its highest-ever growth rate from January to March this year. This explosion in demand has directly led to a shift from a surplus to a shortage of computing resources. More critically, upstream supply chain costs continue to climb. Prices for core components like memory, CPUs, and GPUs are rising across the board, with AWS and Google Cloud having already implemented price increases of 15% to 100% earlier this year. Industry experts analyze that Alibaba Cloud's price adjustment is a reflection of this broader industry trend. However, unlike its overseas counterparts, Alibaba Cloud's more substantial price hikes signal confidence in its own technological advantages.
Amid widespread price increases by global cloud providers, Alibaba Cloud's 34% hike still captures market attention. What underpins this confidence? In the past, cloud providers primarily relied on selling IaaS infrastructure, earning thin margins in an industry plagued by price-cutting competition for market share. Now, with the proliferation of AI applications, the billing unit for cloud services is shifting towards Tokens, allowing computing power consumption to be directly converted into recurring revenue streams.
It is worth noting that organizational changes support this strategy. On March 16, Alibaba established the Alibaba Token Hub (ATH) business group, integrating the Tongyi Lab, MaaS business line, and Qianwen business unit. Alibaba Group CEO Eddie Wu stated in an internal memo that a significant amount of digital work will be supported by AI Agents, which will, in turn, be powered by Tokens generated by models. This signifies that Alibaba Cloud is undergoing a comprehensive upgrade from selling low-margin IaaS infrastructure to selling high-margin MaaS and intelligence.
The financial implications of this model upgrade are clear. Morgan Stanley calculates that assuming a 20% annual contract renewal rate, every 10% increase in overall contract prices would lead to a full 4 percentage point expansion in EBITA margin. Driven by these price increases, Alibaba Cloud's revenue growth is projected to reach 50% by F27, with its EBITA margin soaring to over 12%. Concurrently, valuation multiples are expected to expand, with the EV/Sales ratio for the cloud business increasing from 5.5x to 7x, and the per-share contribution value jumping from $91 to $127. This factor alone could push Alibaba's target price to $260.
Of particular interest is the role of in-house developed chips, which Alibaba Cloud views as a key advantage enabling its pricing power. From an industry perspective, while competitors may be forced to raise prices due to cost pressures, the cost advantages derived from proprietary chips allow Alibaba Cloud to expand its profit margins, creating a widening profit differential. According to Morgan Stanley analysis, Alibaba's self-developed T-Head chips can reduce costs by more than half compared to purchasing third-party GPUs. Furthermore, the deep integration between these proprietary chips and the Qianwen large model can significantly enhance inference efficiency and lower the cost per Token.
Public information indicates that the "Zhenwu" PPU utilizes a self-developed parallel computing architecture and inter-chip interconnect technology, paired with a full-stack proprietary software suite, achieving complete in-house development of both hardware and software. It features 96GB of HBM2e memory and an inter-chip interconnect bandwidth of 700 GB/s, making it applicable for AI training, AI inference, and autonomous driving. According to Alibaba, this chip is already being used extensively for the training and inference of the Qianwen large model and is deeply optimized in conjunction with Alibaba Cloud's comprehensive AI software stack to provide integrated products and services.
Alibaba Cloud's price hike is not an isolated incident. On January 22, AWS announced a 15% price increase for EC2 instances used for large model training. On January 27, Google Cloud announced price increases for data transfer services, AI, and computing infrastructure, with hikes of up to 100%. On March 18, Baidu AI Cloud also adjusted prices for some of its AI computing products, increasing them by approximately 5% to 30%.
Morgan Stanley believes that this price increase could alter the valuation logic for Alibaba, compelling the market to abandon the old habit of using weak e-commerce performance to suppress its valuation. Industry experts suggest that the deeper significance lies in the potential for this round of price increases to initiate a profitability cycle for China's AI cloud industry. The valuation logic is shifting from revenue growth to profit margins, and the transformation of computing power into high-margin Token services may reshape the industry's valuation system.
Comments