Piper Sandler has opened formal coverage of the artificial intelligence chip segment, initiating five semiconductor stocks at Overweight in a report that frames the group as primary beneficiaries of rapidly rising demand for AI compute.
Analyst David O'Connor launched his coverage of NVIDIA with a $300 price target, describing the company as the "outright leader in AI compute with 80% market share." He said the emergence of agentic AI workloads this year has produced a step-change in demand for high-performance compute, and he expects supply to remain constrained for the next two to three years. O'Connor also characterized NVIDIA as "among the cheapest in AI universe" at roughly 14 times fiscal 2028 estimates.
Broadcom was started at a $460 target. O'Connor labeled Broadcom the leader in custom ASIC chips with a roughly 75% share of that market and flagged a supply crunch, writing that demand is currently twice available supply and that securing additional capacity will be critical. He noted Broadcom has line of sight to 12 gigawatts of demand in fiscal 2027.
Advanced Micro Devices began coverage at a $600 target and was described as an "Agentic AI Sweetspot." O'Connor pointed to AMD's share gains in server CPUs and the ramp of its Helios GPUs, which already have anchor clients including OpenAI, Meta and Anthropic. He projected earnings for AMD to grow at a 65% compound annual rate through 2030.
Marvell Technology was initiated at $270, with O'Connor emphasizing Marvell's data center franchises and a $120 billion agreement with Google that he called transformative. He also identified an Oct. 6 analyst day as a potential catalyst for the stock.
Rounding out the list, Arm Holdings was started at a $320 target. O'Connor highlighted Arm's dominance in CPU intellectual property and its expansion into custom silicon and accelerator IP, a strategy he said could result in a doubling of earnings.
Summary
Piper Sandler initiated coverage of five semiconductor names tied to AI compute demand, issuing price targets and Overweight ratings while flagging constrained supply, concentrated market shares, and significant demand visibility tied to major customers and agreements.
Key points
- Piper Sandler started five AI-related chip stocks at Overweight with specific price targets for each company.
- Analyst David O'Connor emphasized a structural step-change in demand from agentic AI workloads and expects supply to remain tight for two to three years, creating near-term scarcity in the market.
- The coverage highlights concentrated market leadership across niches - NVIDIA in AI compute, Broadcom in custom ASICs, AMD in server CPUs and GPUs, Marvell in data center networking with a large Google agreement, and Arm in CPU IP and custom accelerator IP.
Risks and uncertainties
- Supply constraints: O'Connor indicates supply will likely remain constrained for two to three years and that demand currently exceeds supply by about 2x, creating execution risk for companies needing additional capacity - a material factor for the semiconductor and data center hardware sectors.
- Customer concentration and execution: AMD's Helios GPU ramp cites anchor clients such as OpenAI, Meta and Anthropic; outcomes depend on continued adoption and execution, which affects the server CPU and AI accelerator markets.
- Dependence on major agreements: Marvell's $120 billion Google agreement is described as transformative, but realization of expected benefits hinges on successful implementation and sustained demand in data center networking.
These initiations position the five companies as core plays on accelerating AI compute demand, while also flagging supply-side limitations and concentrated exposures that could influence outcomes across semiconductor, data-center hardware and cloud infrastructure markets.