Arm Collaborates with Samsung on AI Accelerators but Focuses More on Data Center Ambitions
Arm Holdings has partnered with Samsung to develop a next-generation, power-efficient AI accelerator SoC using Arm's IP and Samsung’s SF2 2nm process. The collaboration is aimed at mobile inference applications rather than the high-margin data center segment where NVIDIA dominates. While this move expands Arm into on-device AI processing for consumer electronics, it does not signal an immediate breakthrough in the lucrative cloud computing market. Arm has positioned its AGI CPU as a key growth driver with demand exceeding $2 billion across fiscal 2027 and 2028. The company's data center royalty revenue more than doubled year-over-year, and management aims to build a $15 billion silicon business in the long term. However, Arm’s gross margins on CPUs are expected to be significantly lower—around high 30s or low 40s—compared to its IP licensing model. NVIDIA remains dominant in data center revenue with Q2 fiscal 2027 earnings reaching $89 billion and a growing demand for AI infrastructure. Arm’s Vera CPU is embedded within NVIDIA's offerings, highlighting the complex coexistence between the two companies rather than direct competition.
Tickers: $ARM $NVDA $QCOM $TSM
Topics: AI accelerators Arm Holdings data center market expectations Samsung