Nvidia's Huang says faster chips are the best way to reduce AI costs

Nvidia's Huang says faster chips are the best way to reduce AI costs

  • 19.03.2025 18:47
  • nbcnewyork.com
  • Keywords: Nvidia

Nvidia CEO Jensen Huang emphasized that faster chips, such as the upcoming Blackwell Ultra systems, will reduce AI costs by improving efficiency and revenue potential. He highlighted that cloud providers have already invested heavily in these advanced GPUs, which are 50 times more effective than previous models, underscoring the importance of prioritizing speed and performance in AI advancements.

Nvidia NewsNvidia ReportsNVDAsentiment_satisfied
open_in_newOriginal article

Estimated market influence

Nvidia

Nvidia

Positivesentiment_satisfied
Analyst rating: Strong buy

Nvidia's CEO Jensen Huang emphasized the importance of faster chips in reducing AI costs and highlighted their Blackwell Ultra systems which are expected to generate significant revenue compared to previous models. The company has secured major cloud provider investments, indicating a strong market position.

Context

Summary: Nvidia's Huang Says Faster Chips Are the Best Way to Reduce AI Costs

  • Key Highlights:

    • Nvidia CEO Jensen Huang emphasized that faster GPUs are the most effective way to reduce AI costs during his keynote at the GTC conference.
    • The company’s Blackwell Ultra systems are expected to generate 50 times more revenue than its Hopper systems due to their superior speed in handling AI tasks.
    • Cloud providers have already purchased 3.6 million Blackwell GPUs, up from 1.3 million Hopper GPUs sold previously.

  • Business Insights:

    • Faster chips will address concerns about cost and return on investment (ROI) for cloud providers and AI companies.
    • Huang demonstrated that improvements in chip performance over the next decade will make speed the most effective way to reduce costs.
    • Nvidia’s focus is on ensuring large projects utilize its latest systems, highlighting the importance of cutting-edge technology.

  • Market Implications:

    • The shift toward faster, more efficient chips underscores the growing demand for high-performance AI hardware in cloud computing and AI research.
    • Nvidia’s roadmap for Blackwell Ultra systems aligns with cloud providers’ need for clarity on future investments in data centers.

  • Competitive Dynamics:

    • Huang dismissed the idea that custom chips from cloud providers could challenge Nvidia’s GPUs, citing their lack of flexibility for rapidly evolving AI algorithms.
    • He expressed skepticism about recently announced custom AI chips (ASICs), suggesting only those offering significant improvements will succeed.

  • Long-Term Effects:

    • The focus on faster chips signals a long-term strategic shift in the AI hardware market toward performance and efficiency.
    • Nvidia’s emphasis on cutting-edge technology positions it as a leader in driving advancements in AI hardware and cloud computing.