Rearchitecting AI Infrastructure with General Purpose and Accelerator Chiplets

As AI workloads rapidly evolve- the industry is moving toward modular- chiplet-based architectures that combine general-purpose compute and domain-specific acceleration. This panel will bring together leaders from across the Arm ecosystem‚ including silicon providers- hyperscalers- and EDA-toolchain partners‚ to discuss how Arm-based chiplets are enabling flexible- scalable- and energy-efficient solutions for AI and general-purpose infrastructure.

Panelists will explore the technical and ecosystem innovations driving this transformation- including coherent interconnects like AMBA CHI C2C- open standards for interoperability- and advancements in lifecycle analytics- packaging- and system integration. The session will provide a forward-looking view into how Arm's compute platforms are enabling a new generation of composable- AI-ready infrastructure tailored for the needs of hyperscale data centers and edge environments alike.

Morderator:

• Eddie Ramirez, VP of Go-To-Market, Infrastructure, Arm

Panelists:

• Santhosh Srinath, Computer Architect, Meta
• Jinwook Oh, CTO, Rebellions
• Daniel Ping Assistant Vice President, Novatek