Is there another wave in the CPU field recently? Recently, we have seen intensive catalysts in the CPU field: 1. The latest research report from Bank of America proposes to upgrade the server CPU TAM to $210.6 billion by 2030 (up from approximately $170 billion), with a CAGR of approximately 36%. During AMD's financial report in May, Su Mama's expectation was $120 billion, which will be reflected in the Advancing AI event in July. Su Mama raised this number to 220 billion US dollars again. Then Bank of America's recent increase is equivalent to investment banks' expectations being aligned with industry forecasts. 2. The core logic of the CPU still lies in the convergence of the CPU to GPU configuration ratio from 1:8 to 1:1, following the logic chain: the popularization of Agenetic AI → the explosion of inference volume → each agent requires a CPU to orchestrate and process data → the ratio of CPU to GPU is increased from 1:4 to 1:1 or even higher → the CPU TAM is doubled. This call from AMD in early May will be held at https://(x.com)/qinbafrank/status/2051873015458353265? S=46&t=k6rimWSEbo2D2TXolYcM-A and In late May, Chen Liwu delivered a speech at https://(x.com)/qinba frank/status/2057118267681194169? The trend mentioned by s=46&t=k6rimWs Ebo2D2TXolYcM-A seems to be increasingly accepted by the market recently. 3. Intel 1) Chen Liwu subscribed for Intel stocks worth $12 million with personal funds, as part of the company's $19.7 billion stock issuance plan. The management's genuine investment is the greatest guarantee of market confidence and sentiment; 2) The first data center CPU in the 18A process for mass production/ramp up: The Xeon 6+(Clearwater Forest, up to 288 E-cores) adopts Intel 18A (1.8nm level)+Foveros Direct 3D packaging, designed specifically for cloud native, intelligent AI, and network intensive loads; 3) In a recent TechSurge interview, Chen Liwu explicitly explored CPU and memory stacking and new memory architectures (associated with XBM patents, removing traditional HBM silicon intermediaries; Former CEO of SK Hynix has been poached. In fact, the information already mentioned during the previous earnings conference call does not prevent the market from using it to make expectations again. 4 AMD 1) The Advancing AI 2026 event (July 22-23) officially released the 6th generation EPYC Venice (EPYC 9006 series), the world's first high-performance computing chip based on TSMC's 2nm mass production. 2) The Ryzen AI Max 400 series is gaining momentum (up to 192GB of unified memory, the first time in x86 history that it can run a super large model locally); Zen 6 Consumer Grade Expected by 2027; The low-power core has been confirmed to be added. 3) The data center market share continues to increase and is considered one of the biggest beneficiaries of this round of CPU shortage. 5 ARM 1) The self-developed AGI CPU chip (not just an IP) has been sent for customer evaluation and is expected to be mass-produced by the end of 2026, targeting scenarios such as data center head nodes and intelligent agent orchestration. 2) Nvidia's self-developed Arm CPU is accelerating its landing, and the Vera CPU, which works in conjunction with Rubin GPU, is expected to be mass-produced and shipped in the second half of 2026. Early customers are already verifying it. Moreover, it seems that the trend is about to break through the pressure line of this wave of decline. This article is sponsored by @ bitget_zh, titled 'Bitget Buying US Stocks: Instant Entry, Smooth Trading'