The AI party is just getting started

From Market Insider:

All eyes are on Nvidia as it is scheduled to report its fourth-quarter earnings results on Wednesday after the market close.

Nvidia has spearheaded the excitement seen in artificial intelligence technologies, and investors will look to the company’s results to see if the hype can continue.

Wall Street analysts are laser focused on the company’s demand outlook for its AI-enabled H100 GPU chips, which can sell for upwards of $40,000, as well as its planned product roadmap over the next year.

. . . .

  • Revenue: $20.41 billion
  • GAAP earnings per share: $4.23
  • Adjusted earnings per share: $4.60
  • Gross margin: 75.4%

While Nvidia has seen incredible demand for its chips from cloud hyperscalers like Microsoft and Amazon, regulatory hurdles have curtailed its ability to sell chips to China, which made up about 20% of its total revenue last year.

Driving much of the strength in Nvidia’s business has been its exposure to data-centers. Investors will be looking to see just how much demand could be left for the data-center market, and whether Nvidia has lost any market share to its competitors like AMD.

. . . .

“The AI revolution starts with Nvidia and in our view the AI party is just getting started,” Ives said.

“While the Street across the board is anticipating another major ‘beat and raise’ special from Jensen & Co. its all about the pace of data center AI driven spending as the only game in town for GPUs to run generative AI applications all go through Nvidia. We believe peak spending is still ahead for the AI market as many enterprises head down the AI use case path over the next few years and we are expecting more good news from the Godfather of AI this week,” Ives said.

Link to the rest at Market Insider

2 thoughts on “The AI party is just getting started”

  1. On NVIDIA: AMD isn’t their biggest threat. Intel is bigger. Microsoft, Google, Meta, and Amazon bigger still. The markup on those $40,000 modules (72% gross margin!) is too big a cost center for the biggest datacenter operators to remain captive to NVIDIA. I see a parallel ADOBE in the early Desktop Publishing era and even more to NETSCAPE in the pre-HTML 5 days.

    Gouging customers in B2B is not long term viable.
    All four of the big AI players are finalizing their own AI chip designs optimized for their specific LLM tech needs to run alongside the NVIDIA processors they are using. They’ll undoubtedly phase over as soon as technically feasible. As in:

    NVIDIA had better be ready with clearly better tech or lower prices. But neither is also an option.
    Way back in the days of the first XBOX MS used an Nvidia Graphics chip to get the box out the door faster. But as production ramped up the volume and chip production costs dropped, NVIDIA refused requests for a discount to lower the price of the console. MS took significant losses in matching PS2 pricing. As a result, the second generation XBOX 360 arrived a year early and used IBM and AMD tech. Sony took note and also licensed its tech from AMD.

    NVIDIA is not a well loved player in the B2B side of semiconductors.
    The moment they lose even a smidge of supperiority the knives will come out for them from all sides.

Comments are closed.