BofA raises AMD target to $720, sees server CPU TAM tripling to $211B by 2030

Investing.com -- Earlier this week, Advanced Micro Devices (NASDAQ: AMD) became the latest semiconductor company to hit the $1 trillion market cap milestone. One analyst sees the stock continuing higher and sets a new price target, suggesting the market cap could climb to $1.2 trillion over the next 12 months.

BofA Securities analyst Vivek Arya lifted his price target on AMD to $720 from $620 on Friday, reiterating a Buy rating and naming the stock its top CPU pick, as the firm projects the server CPU total addressable market tripling from $61 billion in 2026 to $211 billion by 2030 — with $180 billion of that driven by AI workloads, according to BofA's estimates.

AMD is the most direct equity expression of that TAM expansion, given its competitive x86 server CPU lineup and growing share in AI-optimized data center configurations. BofA's new target is based on 30x its CY2028 earnings estimate, up from a 27x multiple previously, reflecting what the firm describes as higher CPU average selling prices driven by an agentic workload mix. BofA also raised its CY2027 and CY2028 AMD EPS estimates by 2–3% on the same rationale.

The upgrade in multiple and target come as BofA frames agentic AI — systems where software agents autonomously execute multi-step tasks, as a structural demand driver for CPUs, not just GPUs. "Agentic AI requires compute across the system, with CPU latency and task-completion speed increasingly important to GPU utilization and the end-user experience," Arya wrote following a hosted investor call with Ben Bajarin, CEO and Principal Analyst at Creative Strategies. The firm's central thesis is that "CPUs expand the AI system TAM rather than substitute for accelerators", a framing that rejects the zero-sum view of CPU versus GPU competition.

BofA models AI CPU units nearly tripling from 16 million in CY2026 to 53 million in CY2030, reaching roughly 1.5x accelerator unit volumes by then, compared to approximately 0.7x today. The firm also models blended server CPU average selling prices climbing from approximately $1,600 in 2026 to approximately $2,600 in 2030, with the mix shift toward higher-value AI configurations driving the ASP uplift. Bajarin, as cited by BofA, suggested high-end AI CPU package values could reach $4,000–$5,000 or more for NVIDIA's Vera CPU.

BofA's preferred analytical lens for evaluating CPU demand in AI infrastructure moves away from simple CPU-to-GPU ratios. "A better framework assesses how many tasks the GPU initiates, how many the CPU can execute concurrently and how quickly results return to the GPU, with rack-level CPU attach potentially more useful than a universal CPU/GPU ratio," the firm wrote. That framing has direct implications for AMD's addressable opportunity, as agentic architectures rely on CPUs for orchestration tasks including tool calls, code execution, and database interactions.

On architecture, BofA notes that x86 should remain entrenched in legacy cloud and enterprise software environments, while ARM is best positioned for greenfield AI workloads. Bajarin projected ARM-based server CPU unit share approaching 40% by decade-end, "potentially reaching 50% if hyperscaler porting accelerates and NVDA drives broader ARM adoption." He also identified an approximately $80 billion on-premises enterprise CPU opportunity that BofA views as incremental to the cloud-driven TAM.

BofA simultaneously reiterated Buy ratings on both NVIDIA (NASDAQ: NVDA) and Intel (NASDAQ: INTC). NVDA remains BofA's top sector pick, cited for its "broadest portfolio, CPU/GPU fungibility, supply chain investments", attributes that position it to capture value across the compute stack as CPUs and GPUs become increasingly co-dependent. Intel's Buy reiteration was more narrowly grounded in its potential to provide incremental CPU foundry capacity for external customers over the long term.

The Anthropic/Akamai CPU infrastructure deal and Meta's Muse product were both cited by BofA as real-world signals supporting the CPU's expanding role in AI systems, though the firm cautioned that Meta's CPU-rich setup may not reflect GPU-heavy operators more broadly.

For AMD specifically, the key forward catalysts include execution on its server CPU roadmap as agentic AI deployments scale and any updates to hyperscaler capital expenditure plans that could accelerate or delay the TAM ramp BofA projects through 2030. BofA's $720 target implies meaningful upside from current trading levels, and investors will be watching AMD's next earnings report for evidence that CPU ASP trends are tracking toward the firm's CY2027–2028 estimate revisions, the level at which the 30x valuation multiple becomes defensible.

Shares of AMD are up 0.6% to $632.99 in early trading. They are up over 13% this week.

Original Article

You May Also Be Interested In

• Citi cuts Regions Financial price target to $32 from $36
• FICO (FICO) August Breaks UK Credit Card Record as Balances Soar Beyond 2,000GBP
• Jefferies Downgrades Zoetis Inc. (ZTS) to Hold