Nvidia CEO Jensen Huang on Blackwell Sales, Vera Rubin, and China Market
Bloomberg PodcastsNovember 19, 202518 min362 views
33 connectionsΒ·40 entities in this videoβNvidia's Strong Sales and Future Products
- π Blackwell GPUs are experiencing "off the charts" sales, with Nvidia's cloud offerings currently sold out.
- π‘ Nvidia has planned its supply chain exceptionally well, working with partners like TSMC and SK Hynix to meet strong demand.
- ποΈ The new Vera Rubin platform is on track for delivery around Q3 of next year, continuing Nvidia's annual product cycle.
- βοΈ The rack scale architecture, including the MVLink 72 switch, is revolutionary and will seamlessly transition from Grace Blackwell to Vera Rubin.
China Market Strategy
- π¨π³ Nvidia's forecast for the China market is currently zero, but the company desires to re-engage with excellent products.
- π The Chinese market is substantial, estimated at around $50 billion this year, and Nvidia believes its participation benefits both the US and China.
- π€ Nvidia is committed to engaging with both the US and Chinese governments to allow participation in the open market.
Export Controls and Middle East Operations
- πΊπΈ Nvidia is permitted to export up to 35,000 Blackwell chips each to Saudi Arabia and the UAE through G42.
- β οΈ A key US requirement is preventing tech transfer to China through the Middle East, an area Nvidia has consistently addressed with rigorous measures.
Energy and Compute Constraints
- β‘ Growing at a rate of $10 billion per quarter presents challenges, requiring world-class supply chain management and collaboration with energy providers.
- π Nvidia's extensive network of cloud service providers, OEMs, and global partners helps find power solutions at various scales.
- π§ Nvidia's architecture runs every AI model, from frontier models like those from OpenAI and Anthropic to open-source and specialized models, ensuring high offtake for any cloud provider.
Chip Longevity and Depreciation
- β³ Nvidia's CUDA ecosystem provides versatility, allowing GPUs to accelerate every phase of AI (pre-training, post-training, inference) and run diverse models.
- π Even older chips like the Ampear A100, shipped six years ago, remain relevant due to continuous software updates and their significant performance advantage over CPUs for many workloads.
- π° Nvidia's contribution to a 1-gigawatt data center is billions of dollars, with Vera Rubin systems representing approximately $35 billion of a $50-55 billion data center cost.
Knowledge graph40 entities Β· 33 connections
How they connect
An interactive map of every person, idea, and reference from this conversation. Hover to trace connections, click to explore.
Hover Β· drag to explore
40 entities
Chapters8 moments
Key Moments
Transcript69 segments
Full Transcript
Topics14 themes
Whatβs Discussed
NvidiaJensen HuangBlackwell GPUVera Rubin PlatformAISupply ChainChina MarketExport ControlsData CenterCUDAGPUCloud ComputingArtificial IntelligenceEnergy Constraints
Smart Objects40 Β· 33 links
CompaniesΒ· 15
ProductsΒ· 8
ConceptsΒ· 11
MediasΒ· 3
PersonΒ· 1
LocationsΒ· 2