share_log

英伟达财报四大亮点:主权AI起来了、客户没有等待、Blackwell二季度就发货、新芯片“一年一代”

Four highlights of Nvidia's earnings report: Sovereign AI launched, customers didn't wait, Blackwell shipped in the second quarter, and the new chip “one generation a year”

wallstreetcn ·  May 22 21:15

The Blackwell platform achieved revenue this year, the Blackwell compatible Hopper system, and the sovereign AI business revenue may reach several billion dollars, and R100 is expected to be launched next year.

Nvidia's Q1 earnings report was “crushed” in every way! In conjunction with Hwang In-hoon's remarks during the earnings call, there are four highlights worth paying attention to.

The Blackwell platform achieved revenue this year

According to Nvidia's earnings report, the Blackwell platform is fully in production and is expected to be delivered to customers this year.

Hwang In-hoon said:

“We're ready for the next wave of growth. The Blackwell platform is fully in production, laying the foundation for trillion parameter-level generative AI.”

At the same time, Huang Renxun said that Blackwell architecture GPUs will be shipped in the second quarter and production will increase in the third quarter. It is expected that they can be deployed in customer data centers in the fourth quarter, and Blackwell is expected to bring in a large amount of revenue this year.

Eliminate concerns about “gaps” in demand

As Blackwell enters production, some investors are worried that Nvidia may face a sales “gap”: some customers will choose to wait for Blackwell and abandon Hopper.

In response to this concern, Hwang In-hoon emphasized during the conference call that Blackwell can be “backwards compatible” with Hopper's architectural system.

This means that customers don't need to “wait patiently” for Blackwell to ship, buy Hopper early and enjoy it early, and it will also be “silky” if they want to transition to Blackwell later.

Nvidia Chief Financial Officer (CFO) Colette Kress also pointed out that “demand for H2000 and Blackwell far exceeds supply,” and that this situation is expected to continue “until next year.”

The rise of sovereign AI and multi-modal businesses

Nvidia said that sovereign AI and new multi-modal capabilities are the driving factors for the company's further development.

Sovereign AI refers to the ability of national entities to use their own infrastructure and data to produce AI. Nvidia sees this as a new field that is rapidly emerging, and believes that as data privacy issues come into focus, sovereign AI may become more attractive.

Kress shows:

“The importance of artificial intelligence has drawn the attention of every country.”

“As countries around the world deploy sovereign AI, data center revenue continues to diversify... We believe this year's sovereign AI revenue will rise from zero revenue last year to a high of several billion dollars.”

Multi-modal capability means that with AI's ability in multiple modes such as audio and vision, the demand for computing power will continue to grow as AI products rapidly iterate in this field, which will also boost demand for Nvidia chips.

New chip “one generation per year”

Hwang In-hoon revealed that Blackwell will launch a new chip later, and will also speed up the chip architecture update speed, from a two-year update to an annual update:

“I can announce that Blackwell will launch another chip later. Our pace is one generation at a time.”

“We are going to fully advance at a very fast pace.”

“A new CPU, a new GPU, a new network card, a new switch... lots of chips are on the way.”

According to Tianfeng Securities analyst Guo Mingyi's previous forecast, Nvidia's next-generation architecture “Rubin” will be launched in the fourth quarter of 2025, and a new R100 AI chip will be introduced at that time.

Disclaimer: This content is for informational and educational purposes only and does not constitute a recommendation or endorsement of any specific investment or investment strategy. Read more
    Write a comment