Keep pulling the thread on Jensen Huang.
The size of AI models is increasing by an order of magnitude (10x) every year.
The computer industry is undergoing two simultaneous platform shifts: the move to building applications on top of AI, and a fundamental change in how software is developed and run.
An estimated $10 trillion of computing infrastructure from the last decade is being modernized for AI.
Approximately $200 billion in venture capital funding is being invested annually to modernize and invent new AI-native technologies.
Open source AI models have reached the frontier of capabilities, but still lag the most advanced proprietary models by approximately six months.
NVIDIA operates billions of dollars worth of its own DGX supercomputers to develop its open models.
NVIDIA has developed several open, frontier AI models including La Proteina for protein synthesis, OpenFold3 for protein structure, EVO2 for multi-protein systems, ForecastNet and CoreDiff for weather prediction, NemoTron for hybrid Transformer-SSM, Cosmos as a world model, and Groot for humanoid robotics.
Palantir, ServiceNow, Snowflake, CrowdStrike, and NetApp are all integrating NVIDIA's AI and agentic systems into their core platforms.
NVIDIA's Cosmos is an open, frontier world foundation model for physical AI, pre-trained on internet-scale video, driving data, robotics data, and 3D simulations.
NVIDIA has announced Alpamayo, which it claims is the world's first end-to-end trained, reasoning autonomous vehicle AI.
NVIDIA's Alpamayo autonomous vehicle model is open sourced.
The first vehicle featuring NVIDIA's autonomous driving technology, developed in partnership with Mercedes-Benz, will launch in the United States in Q1, followed by Europe in Q2, and Asia in Q3/Q4.