NVIDIA CEO Jensen Huang confirmed the company has certified Samsung, SK Hynix
and Micron to supply HBM4 memory for its US AI accelerator business. He said the
next‑generation Vera Rubin platform is now in full production and scheduled for
initial deliveries in Q3; systems are built around a Vera CPU and Rubin GPU
clusters and will include TB‑level HBM4 per server.