AMD's Server CPUs Take Center Stage as Wall Street Rethinks the AI Hardware Trade

Bank of America raised its AMD price target from $620 to $720, citing CPU demand in AI buildouts after an $11.6 billion Anthropic-Akamai cloud deal.

For years, the market's AI playbook has been almost reflexively narrow: want exposure to artificial intelligence, buy graphics processors and accelerator chips. That shorthand is starting to look outdated.

On Friday, Bank of America analyst Vivek Arya lifted his price target on AMD from $620 to $720 while keeping a "Buy" rating on the stock. His reasoning centers on something less flashy than the GPU arms race — the persistent, load-bearing role of high-performance CPUs in building out modern AI infrastructure. It is a thesis that cuts against the prevailing narrative, and it landed alongside a broader reassessment now rippling through the semiconductor space.

A $11.6 Billion Contract Puts CPUs in the Frame

The catalyst Arya cites is a sprawling industry deal: Akamai Technologies has struck a seven-year cloud capacity agreement with Anthropic valued at roughly $11.6 billion, with room to expand by as much as $9 billion more. What stands out is the technical emphasis — the infrastructure is designed specifically to support CPU workloads. AMD is not named as a supplier in the arrangement, but the direction of travel is hard to miss. Large models and agentic AI systems generate enormous volumes of pre- and post-processing work, and that burden falls squarely on conventional server CPUs.

The pattern had already surfaced earlier in the week. When Meta unveiled its Muse AI agent, it set off a sharp rally across the sector. Bloomberg reported a 10% jump in AMD shares that Monday, while Reuters noted gains across the chip industry and a brief market capitalization above $1 trillion for AMD.

Agentic AI does not operate in a vacuum of pure compute accelerators; it demands flexible system architectures. That intersection is where AMD looks positioned to capture outsized gains in the quarters ahead.

From the Data Center to the Device in Your Bag

There is a case that the market is only gradually pricing in the full strategic breadth of the company, whose portfolio reaches well beyond data centers alone. Take the Ryzen AI Max processors and the Ryzen AI Halo developer platform: Perplexity's Portable Computer runs on supported Windows machines, executing tasks and workflows locally on the device itself. Separately, security firm F-Secure is working with AMD Silo AI on adaptive AI routing that distributes compute loads across local, private, and cloud environments.

Where rivals often concentrate on specialized niches, AMD is methodically building an unbroken bridge between cloud clusters and end-user devices. Such partnerships may look unremarkable next to billion-dollar contracts, but they reinforce the company's deepening integration into real-world software.

Geopolitics Cuts Both Ways

None of this unfolds in a vacuum. The global chip industry is increasingly exposed to trade-policy uncertainty, and AMD's chief executive has stressed the need for stability across the entire technology ecosystem. Her comments came against the backdrop of the Trump–Xi summit and persistent technology-specific tensions between Washington and Beijing.

Those factors carry real weight for chipmakers. Cutting-edge data center processors are produced through densely interconnected, globe-spanning supply chains, and political shocks or trade restrictions can disrupt them quickly. At the same time, the breakneck pace of AI development demands continuous planning certainty in research, development, and production.

Momentum Meets a Demanding Valuation

The stock's technical picture reflects the fundamental setup. AMD closed Friday's session at €554.30, sitting just below the 52-week high of €566.50 touched the same day — a gap of only 2.2%. Reuters attributed support to the recent record run on the Nasdaq and broad interest in AI semiconductors. The market is betting that demand for modern server architecture outweighs geopolitical hurdles, with large data center operators continuing to pour substantial sums into expanding capacity. For training and running complex AI applications, advanced semiconductor components remain indispensable.

Industry experts point to the same operational strength in the core business, citing accelerating demand for AI compute and the company's strong position in both server CPUs and graphics processors. Two forces will shape what comes next: worldwide investment in server and AI infrastructure filling order books, and the company's continued reliance on stable international exchange in the technology sector despite diplomatic friction.

That leaves little room for operational missteps or product delivery delays at such an ambitious valuation. Even so, the balance of risks still tilts toward opportunity. The assumption that tomorrow's data center will be ruled by graphics processors alone is looking more and more like a misread. When hyperscalers and AI pioneers such as Anthropic set aside multibillion-dollar budgets for pure CPU capacity, AMD stands to benefit directly — and the Bank of America upgrade may prove less an isolated move than the opening chapter of a wider revaluation of the CPU's role in the AI era.