Chipmakers deepen AI hardware collaborations

Major semiconductor companies are announcing new products and collaborations aimed at meeting accelerating demand for artificial intelligence and high-performance computing. Recent disclosures from Nvidia (NVDA), Advanced Micro Devices (AMD) and Applied Materials (AMAT) highlight parallel efforts to improve AI inference, telecom-grade AI, and advanced memory technologies.

A February 27 Wall Street Journal report cited in recent coverage said Nvidia is developing a new processor platform focused on AI inference workloads. The system is being designed for OpenAI and other large customers to support faster and more reliable AI tools that generate responses to user queries.

Nvidia is expected to reveal the new inference platform at its GTC conference in San Jose, scheduled for March 16 to March 19, 2026. The platform will incorporate technology from AI chip startup Groq and is intended to deliver faster and more efficient AI responses for customers.

OpenAI is described as one of the largest prospective users of this new processor, with sources indicating the system could handle about 10% of OpenAI’s inference computing needs. The initiative comes amid growing competition from Amazon (AMZN) and Google, which are building their own AI chips.

Nvidia, a fabless semiconductor and AI computing company, designs GPUs, AI accelerators, APIs and system-on-chip units. Of 69 analysts covering Nvidia as of March 10, 64 rate the stock Buy and four rate it Hold, with an average price target of $265 implying more than 43% upside.

On March 2, Advanced Micro Devices (AMD) used a blog post to outline its role in telecom-focused AI at MWC 2026 in Barcelona. AMD highlighted its participation in the Open Telco AI initiative, led by GSMA with AMD, AT&T (T), TensorWave and other industry players.

The Open Telco AI initiative aims to speed development of telco-grade AI models and systems through open collaboration. It will be supported by the open-telco.ai portal, which is planned to host shared datasets, tools and benchmarking resources.

AMD said its Instinct GPUs, together with the ROCm software platform, will be used to train AI models and support training and inference workloads. The company also introduced EPYC 8005 server CPUs optimized for distributed edge environments and virtualized radio access network workloads.

Additional launches included AMD’s Ryzen AI portfolio, featuring Ryzen AI 400 Series and Ryzen AI PRO 400 Series desktop processors. These chips target OEM systems such as high-performance desktops, laptops and mobile workstations designed to run modern AI workloads.

On March 10, Applied Materials announced a long-term research and development partnership with SK hynix to advance next-generation DRAM and high-bandwidth memory for AI and high-performance computing. Engineers from both companies will work together at Applied’s new EPIC Center in Silicon Valley.

The collaboration at the EPIC (Equipment and Process Innovation and Commercialization) Center will focus on materials engineering, process integration and advanced 3D packaging to improve performance, efficiency and manufacturability of future memory architectures.

Applied Materials said the EPIC Center, expected to open later this year, represents the largest U.S. investment in advanced semiconductor equipment R&D, with capital spending projected to scale to about $5 billion. SK hynix will be a founding partner at the facility and will also leverage Applied’s advanced packaging research in Singapore.

According to Applied, the EPIC Center is designed to give chipmakers earlier access to its research portfolio, shorten development cycles and accelerate the transition of new technologies into high-volume manufacturing for AI. As of March 10, 37 analysts cover Applied, with 29 rating the stock Buy and eight Hold, and an average price target of $427.50 implying upside of more than 23.5%.