AMD stock rises as analysts lift targets on AI demand

AMD shares climbed 3.5% on Thursday after several Wall Street firms raised price targets and amid reports that Samsung may partner with AMD on chip production as Taiwan Semiconductor Manufacturing Co.’s capacity remains tight. Analysts attributed the moves to stronger demand for AI-related chips and server central processing units.

Bernstein increased its price target to $600 from $525 and maintained an Outperform rating. Baird raised its target to $625 from $300. Citi upgraded AMD to Buy from Neutral and set a target of $575, up from $460. Bank of America moved its target to $560 from $500. Barclays set a $500 objective, and Goldman Sachs raised its price target to $450 and upgraded the shares to Buy.

Bernstein also revised its forecast for the global server CPU market to $223 billion by 2030 from a prior estimate of $137 billion. The firm said a next phase of AI, described as agentic systems that plan and execute multi-step tasks, will increase demand for general-purpose CPUs to handle orchestration and memory management alongside accelerators.

Analysts highlighted that CPU-to-accelerator ratios in some inference-heavy applications may shift from earlier assumptions of one CPU per four to eight GPUs toward ratios closer to one-to-one. Market participants pointed to AMD’s EPYC server processors as a product that fits those requirements.

Reports of a possible Samsung manufacturing partnership followed industry notes that TSMC’s capacity for advanced chips is constrained. A Samsung tie-up would offer AMD an alternative foundry option for chips aimed at AI and data-center customers.

AMD has also expanded technology agreements. The company announced a collaboration with quantum technology firm EigenQ tied to a potential $3 billion SPAC transaction; the work is aimed at post-quantum cryptography, quantum-safe communications and sensing for defense and industrial applications. Separately, AMD signed a multi-year agreement with Rackspace Technology for a phased deployment of 30 megawatts of AMD-powered computing capacity between 2026 and 2028, using Instinct GPUs and EPYC CPUs to support regulated enterprise and healthcare AI workloads.

Analysts noted that GPUs remain central for model training and parallel workloads, while CPUs coordinate tasks, manage memory and support inference pipelines. Several firms pointed to a growing mix of CPU and GPU demand as AI applications add orchestration and multi-step processing requirements.

The cluster of analyst upgrades and the production reports followed AMD’s recent partnerships and product announcements. Financial firms adjusted ratings and targets based on updated forecasts for server CPU demand and the company’s positioning in AI infrastructure markets.