Скачать или смотреть #OCPSummit25

Check out full showcase at: https://ngi.fyi/ocpsummit25yt to get the attendee and sponsor perspectives from the show floor of the #OCPGlobalSummit 2025

Aravind Srikumar, SVP of Product at Upscale AI, discusses the critical infrastructure challenges facing AI deployment at scale. As hyperscalers struggle to secure sufficient compute resources, the industry faces a pivotal moment in how AI systems are architected and interconnected. Srikumar explains why the traditional approach dominated by proprietary ecosystems may not sustain the explosive pace of AI innovation, and how his company addresses this through a fundamentally different architectural philosophy. He explores the technical heart of scale-up domains and reveals why the convergence of two traditionally separate technology domains holds the key to meeting AI's unprecedented growth demands.

Why hyperscalers are facing acute compute shortages and how this impacts AI deployment strategies
The role of ASIC technology in interconnecting GPUs within scale-up architectures
How proprietary ecosystems currently dominate the AI infrastructure landscape
Why open standards matter for accelerating innovation in AI compute
The technical and strategic importance of integrating software and hardware layers
How the convergence of compute and networking paradigms enables larger unified clusters
What architectural approaches are needed to match AI's rapid growth trajectory


Want to be involved our video series? Contact [email protected]