Da Mo has completely changed its tune: Qualcomm officially joins the AI winner camp I just finished reading the latest Qualcomm research report by Da Mo today: Investor Day AI presents a more optimistic view of data center; move to EW。 Goldman Sachs just raised its target price for Qualcomm yesterday, and today it's Morgan Stanley's turn. The biggest change for Morgan Stanley is the direct upgrade of Qualcomm's rating from Underweight to Equal weight, with the target price raised from $146 to $231 in one go. Why did it suddenly turn? There is only one core reason: AI data centers Qualcomm gave guidance on Investor Day that far exceeded market expectations: FY26 data center revenue was approximately $300 million, FY27 guidance jumped directly to $5 billion, and FY29 target exceeded $15 billion. Growth has exceeded 16 times in one year. Da Mo admitted that he had been too conservative towards Qualcomm before. There is an interesting sentence in the report: "We suspected wrong before." The reason is that Qualcomm's management has always been conservative, unlike many companies that like to paint big cakes. So the logic of Da Mo is: if the management dares to publicly disclose the figure of FY27 $5 billion, it means that there is a high probability that the order has been placed. They admit that they cannot predict the specific details, but believe that the numbers should be at this level. But Da Mo still had a few doubts, which is why he didn't give Overweight directly The first issue is that the timing of the AI Accelerator is too late. Qualcomm's real data center AI chips are expected to start in the second half of 2027, while Nvidia, Cerebras, and various startups are already shipping them. Da Mo's attitude is "Show me" - get the confirmed customer first before making a decision. Although Microsoft is on the platform, it is currently only expressing a willingness to cooperate in the future and has not yet confirmed the order. The second reason is that the CPU timing of the server is also relatively late. Qualcomm's server CPUs are expected to be available after mid-2028, when AMD, Intel ARM、 Amazon's Graviton, Google's Axion, and Nvidia's CPU may have already been deployed, and the competition will be much more intense. The report also mentioned a detail: currently, due to chip shortages, CPU supply is accelerating and increasing in various places. Meta's attitude is also ambiguous, and they have stated that they are already using other more mature ARM server architectures. The third target is FY29's $15 billion, which Da Mo believes is still unrealistic. They believe in FY27's $5 billion, but are unwilling to price all the growth for the next three years in advance. So this research report is more like a valuation for the next year, not for the next three years. There is a detail that Da Mo himself finds quite surprising, which is worth mentioning separately In FY27, there were almost no CPUs and very few Accelerators in the $5 billion revenue, mainly from customized ASIC business. This shows that Qualcomm's real opportunity is to help cloud manufacturers do ASICs (Qualcomm suggests that it is a Chinese customer and an American customer. The Chinese customer's rate is probably ByteDance, and the American customer is unknown). The speed of realizing this portion of income may be faster than the market originally imagined. This is in line with the logic of OpenAI partnering with Broadcom to develop ASICs mentioned earlier. The AI chip industry is moving towards a dual track architecture where GPUs are responsible for training and ASICs are responsible for inference. Qualcomm's repricing this time is essentially stepping on this line. Besides AI, Da Mo is also optimistic about its automotive business The FY29 automotive revenue target has been increased from $8 billion to $10 billion. Robotics, industrial automation, and physical AI are also considered long-term growth points. In terms of mobile phones, Da Mo is still cautious and has not changed The loss of Apple's baseband, limited growth of Android (assuming a long-term growth rate of only about 5%), rising memory prices that may drag down mobile phone demand, and Samsung's market share returning to normal are all pressures still present. Mobile phones are not the logic behind this rating upgrade, AI data centers and cars are. My own opinion The most noteworthy aspect of this report is the change in narrative. In the past, when Wall Street looked at Qualcomm, the core logic was that it was a smartphone chip company, and its valuation should be relatively low. Now it is becoming an AI infrastructure plus customized ASIC plus automotive plus Edge AI and Physical AI company. If it can really realize $5 billion in data center revenue in the next year, the market is likely to continue to give it a higher AI premium. But I also agree with the caution that Da Mo retains: the $15 billion target after 2027 still lacks customer validation and product realization. Qualcomm is now more like an AI transformation story that has just begun to be re priced by the market, rather than a company that has fully realized its growth. This is consistent with the core judgment of the Qualcomm research report I wrote before: the valuation gap is where the opportunity lies, but whether this gap can truly narrow depends on whether the data center revenue can really land in the next one or two years.