Piper Sandler Targets AI Chip Leaders on Compute Demand
Nvidia, Broadcom, and peers face strong growth as AI compute demand outpaces supply, with analysts projecting significant market expansion by 2030.
Piper Sandler analysts have initiated coverage on five major semiconductor firms, assigning overweight ratings to Nvidia, Broadcom, Arm Holdings, Advanced Micro Devices, and Marvell Technology. The investment bank projects price targets ranging from 19% to 37% above current levels, citing a critical shortage in artificial intelligence compute capacity. This assessment is reported by GN stocks/nasdaq, highlighting that demand for GPU usage is driving hourly pricing up significantly, with Nvidia’s H100 units seeing a 30% increase from December prices.
The analysts argue that the market for AI chips is currently sold out for next year, with order books already filled for 2028. They forecast that the total compute market will expand fivefold to reach $2.2 trillion by 2030, of which approximately $2 trillion will be allocated to AI-specific hardware. This structural shift positions the covered companies to capture substantial revenue growth as enterprises and cloud providers race to secure processing power.
Nvidia and Broadcom Lead Custom Chip Markets
Nvidia is identified as the primary beneficiary, holding roughly 80% market share in AI compute. Piper Sandler places a $300 price target on the stock, reflecting a 37% upside, and notes that the company’s CUDA software platform provides a durable advantage in model training. The firm also highlights Nvidia’s strategic move into inference through its acquisition of Groq, strengthening its position against emerging competitors.
Broadcom is viewed as the value play, with a $460 target representing 27% potential gains. The company commands about 75% of the application-specific integrated circuit market for inference, where demand currently exceeds supply by a factor of two. Analysts project Broadcom’s AI revenue will surge 400% over the next two years to reach $230 billion in fiscal 2028, supported by its leadership in custom silicon and networking infrastructure.
CPU Leaders Target Agentic AI Growth
Arm Holdings and Advanced Micro Devices are positioned to benefit from the rise of agentic AI, which requires robust CPU performance. Arm holds a $320 price target, a 26% premium, driven by its expansion into physical server chip manufacturing while continuing to license intellectual property to major players like Nvidia and Amazon. However, the stock carries a high valuation at 114 times forward earnings.
AMD receives a $600 target, implying 19% upside, as it gains server CPU share from Intel. The company’s GPUs are also benefiting from inference workloads. Piper Sandler notes that AMD is well-placed to capture market share in both general-purpose computing and specialized AI processing, leveraging its dual competency in CPUs and GPUs to address the growing needs of enterprise AI applications.
Marvell Focuses On Optical Networking
Marvell Technology is included in the group due to its strength in custom chips and optical networking. While specific price targets for Marvell are part of the broader 19-37% range, the firm’s infrastructure is critical for the high-bandwidth data transfer required by large-scale AI clusters. This positions the company as a key supplier for the physical layer of data centers, complementing the compute-centric strategies of its peers.
