Advanced Micro Devices (AMD) has announced the commencement of production ramp-up for its next-generation EPYC processor, codenamed "Venice," utilizing TSMC's advanced 2-nanometer process technology. This marks the industry's first 2nm-class data center CPU to enter this phase, signaling the imminent large-scale commercial production of such processors. This development coincides with a surge in demand for data center CPUs, driven by the anticipated proliferation of autonomous AI agent tools like Claude Cowork and OpenClaw by 2026, which has created significant supply shortages.
Amid this demand explosion, Wall Street's bullish sentiment toward major CPU players—Intel (INTC), AMD (AMD), and Arm (ARM)—has intensified. Their stock prices have surged this year, with Intel's year-to-date gain reaching 222%. While the AI narrative was previously dominated by GPUs, the rise of agentic AI workflows has underscored the critical role of CPUs in orchestrating inference workloads, data management, task scheduling, and system coordination, effectively repositioning CPUs from undervalued infrastructure back to the center stage of the chip industry.
The AI computing investment focus is shifting from a "single-point GPU race" to a "full-stack computing system driven by AI agents." Consequently, the next wave of alpha returns is expected to spread across the entire AI computing infrastructure stack, including data center CPUs, DRAM/NAND/HBM memory, AI PCBs, liquid cooling systems, data center optical interconnects, and advanced packaging. Within this narrative shift, CPUs, optical interconnects, and memory chips are poised to be major beneficiaries.
AMD's "Venice" represents the sixth-generation EPYC data center CPU. It's important to note that "production ramp-up" does not equate to full-scale mass shipment; it indicates the transition from prototyping and validation to preparation for commercial volume production, pending yield optimization, packaging, platform certification, and customer validation. AMD's official announcement on May 21st stated that "Venice" is ramping up in Taiwan and is planned for future ramp-up at TSMC's Arizona fab.
This milestone demonstrates AMD's progress in delivering the performance and efficiency required for next-generation cloud, enterprise, and AI computing infrastructure. AMD CEO Dr. Lisa Su emphasized that advancing "Venice" on TSMC's 2nm process is a crucial step in accelerating next-gen AI infrastructure, highlighting the deep collaboration with TSMC to bring leading computing technology to market at the required scale and speed.
"Venice," based on the Zen 6/Zen 6c architecture and TSMC's 2nm process, is reported to scale up to 256 cores—a roughly 33% increase over the current EPYC Turin's 192 cores. AMD's roadmap claims it can deliver approximately 1.7x performance improvement and 1.3x thread density increase, supporting higher memory bandwidth and PCIe Gen6. Its positioning is significantly above the current generation of data center CPUs from both AMD and Intel (Xeon 6). If its specifications are realized, "Venice" could substantially amplify AMD's advantages in high-density cloud, HPC, database, virtualization, and AI inference orchestration servers.
AMD also plans to extend its 2nm data center CPU portfolio with "Verano," another sixth-generation EPYC processor optimized for performance-per-dollar and performance-per-watt for cloud and massive-scale AI workloads. "Verano" is expected to introduce advanced memory innovations like LPDDR to meet the demands of power-constrained workloads and future AI agent and GPU platform requirements.
Wall Street analysts are expanding the AI computing infrastructure narrative from a "GPU-dominated" view to a "full-stack" reassessment encompassing AI GPUs/ASICs, CPUs, HBM/DRAM/NAND memory, and optical interconnect systems. Reflecting this, Citigroup recently raised its 12-month price targets for both Intel and AMD, citing the data center CPU demand boom and upward revisions to total CPU market size projections.
At JPMorgan's recent technology conference, Intel CEO Pat Gelsinger noted that the Intel 18A process (roughly equivalent to 1.8nm) is supporting Panther Lake production with yields improving about 7% per month, exceeding internal expectations. He also highlighted the growing importance of CPUs in the AI era as focus shifts from training to inference, with CPU-to-GPU ratios moving from 1:8 toward 1:1 or even 4:1. Intel is also actively pursuing custom AI CPU/GPU ASIC business.
The shift from AI training to inference and agentic AI significantly increases the strategic weight of CPUs in data centers. While GPUs handle massive matrix computations, CPUs manage scheduling, I/O, memory, task orchestration, security, and multi-agent workflow execution. As AI applications evolve into persistent "software labor," CPU demand transitions from traditional server refresh cycles to AI infrastructure expansion cycles.
Citigroup's updated model projects the total addressable market (TAM) for data center server CPUs to grow from $29.3 billion in 2025 to $131.5 billion by 2030, representing a compound annual growth rate (CAGR) of approximately 35%. The firm raised its price target for Intel from $95 to $130 and for AMD from $358 to $460.
Analysts at Citigroup, led by Atif Malik, further noted that the market for data center CPUs driven by AI agents like OpenClaw is expected to grow at a staggering 185% CAGR, potentially reaching $59.4 billion by 2030 and accounting for 45% of the total market, making it the fastest-growing segment.
According to TipRanks, the highest Wall Street price target for AMD stands at $625, implying roughly 40% upside potential. AMD's stock has already surged 110% year-to-date, establishing it as a major "AI super-bull" stock globally. This $625 target from Baird analyst Tristan Gerra, who upgraded from $300, is based on the unprecedented CPU/GPU demand boost from agentic AI and a re-rating of AMD's data center platform value.
Analyst Tristan Gerra stated that AMD's opportunities in AI server CPUs, GPUs, and accelerated computing platforms are being repriced. Driven by the AI inference wave, agentic AI, EPYC CPUs, Instinct GPUs, and next-gen infrastructure expansion, the market is increasingly viewing AMD as a key beneficiary in the AI computing chain, offering a critical "CPU+GPU+super-platform" solution.
Comments