Archive — US — techcrunch
While OpenAI races to build AI data centers, Nadella reminds us that Microsoftalready has them
Microsoft CEO Satya Nadella showcased the deployment of their first massive AI system, emphasizing their existing data centers for AI workloads, while OpenAI seeks to build its own. Microsoft's AI systems use Nvidia technology, including the Blackwell Ultra GPU, amidst OpenAI's significant data center expansion commitments.
Key points
- Satya Nadella tweeted about the deployment of Microsoft's first AI ‘factory’.
- The AI systems utilize more than 4,600 Nvidia GB300 computers with Blackwell Ultra GPUs.
- Microsoft is positioned to deploy hundreds of thousands of these GPUs globally.
- Nvidia's acquisition of Mellanox in 2019 for $6.9 billion enabled cutting-edge networking technology.
- The timing follows OpenAI's recent high-profile data center agreements.
- Microsoft operates over 300 data centers in 34 countries.
- These AI systems will support models with hundreds of trillions of parameters.
Why it matters
- The competition between Microsoft and OpenAI reflects broader trends in AI development.
- OpenAI's projected $1 trillion commitment to data centers highlights significant investment in AI infrastructure.
- The need for advanced AI capacities aligns with 2025 trends emphasizing fast, scalable data processing and machine learning capabilities.
Microsoft is leveraging its established data center infrastructure to meet growing AI demands ahead of OpenAI's ambitious plans, showcasing the competitive landscape in AI tech development.