Insights
Insights
KPMG reports, M&A trends, and owner exit stories — read the market with our specialists.
From AI Factory to Physical AI — How Nvidia GTC 2026 Is Redrawing the Industrial Landscape
Nvidia has unveiled the blueprint for its next-generation AI platform. According to Samjong KPMG Economic Research Institute's Business Focus, GTC 2026 — held March 16–19, 2026 in San Jose, USA — drew over 30,000 in-person attendees and another 30,000 online, with big-tech and industrial firms worldwide unveiling AI-ecosystem strategies.
Vera Rubin and the inference-chip war. The headline announcement at GTC 2026 was the supercomputing platform “Vera Rubin,” which integrates seven new chips, including Groq 3 LPUs. Vera Rubin lowers cost per token by up to 10× compared with Blackwell and lifts throughput per watt by up to 10×. By integrating the LPU (Language Processing Unit) technology from Groq — acquired for approximately USD 20 billion in December 2025 — into Vera Rubin, Nvidia is signaling an expansion of dominance in the inference market. Samsung Electronics and SK hynix supply memory into the Vera Rubin platform, embedding Korean semiconductor firms directly into the ecosystem.
Expansion of AI Factory and Physical AI. Nvidia unveiled the “DSX AI Factory Platform,” a digital-twin-based architecture and simulation platform for designing AI factories. The industry distribution at GTC 2026 — internet software and services (45.1%), IT services (16.3%), robotics (4.4%), and manufacturing (2.3%) — clearly shows the expanding trajectory of physical AI and industrial automation. Hyundai Motor Group's participation as a mobility-sector company is also notable.
What this means for M&A. The AI ecosystem restructuring laid out at GTC 2026 translates into three M&A opportunities. First, strategic bolt-on acquisitions of semiconductor components, materials, and packaging firms aiming to embed into the Vera Rubin ecosystem will expand. Second, with the inference-specialized chip (LPU) market diverging from GPU-centric architectures, competition will intensify to acquire Groq-like inference-specialized AI silicon startups. Third, demand for software, sensor, and actuator company acquisitions will rise — driven mainly by traditional manufacturing and mobility companies — to build AI factories and scale physical AI (robotics, autonomous driving, industrial automation).
---
Full report: Samjong KPMG Economic Research Institute, Business Focus (March 2026)
Contact: Samjong KPMG M&A Center