news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

Recently, the US media The Information reported that two people involved in Nvidia's chip and server production revealed that due to technical defects, Nvidia's upcoming AI flagship chip GB200 will be delayed by three months or more, which may affect customers such as Meta, Google and Microsoft, which have ordered tens of billions of dollars worth of chips in total.

The delay affects the company's much-anticipated Blackwell product line, which was released in March. In May, Nvidia CEO Jensen Huang said Blackwell was in full production and would be available to cloud computing providers later this year.

Nvidia informed Microsoft and another large cloud computing provider of the Blackwell delay last week, according to Microsoft employees and people with direct knowledge of the matter who spoke to The Information.

People familiar with the matter also said that a product that combines a processor and a graphics chip may not be available in large quantities as quickly as expected due to problems with the supporting technology.

Bloomberg said a version of the chip, called an “AI accelerator,” is being redesigned to work better with the data center infrastructure designed for the Hopper H100 chip.

Nvidia declined to comment on rumors of recent engineering issues. The company said it has begun shipping Blackwell samples to a broad range of customers, production is expected to increase in the second half of the year, and demand for its Hopper product remains strong.

As soon as the foreign media reports came out, coupled with the general decline in technology stocks, Nvidia's stock price fell by more than 6% as of the close of August 6.

Bloomberg also said that Nvidia is launching new chip designs and technologies at a faster pace in an effort to maintain its advantage in AI computing, and the accelerated pace of product releases puts it under more pressure.

At the same time, Nvidia is about to stop producing its popular GeForce RTX 3060 graphics cards. As inventory gradually runs out, Nvidia will shift its focus to the RTX 4060 series products.

According to Board Channels, Nvidia has notified its partners that it will stop producing the RTX 3060. There are only the last batch of GPUs left to order, which means that manufacturers who want to get this cost-effective graphics card need to place an order as soon as possible.

Although the report did not specify whether the 8GB or 12GB version of the RTX 3060 will be discontinued, considering that both use the GA106 GPU core, it is expected that both versions will face discontinuation.

Nvidia has notified several graphics card manufacturers to allocate the last batch of GPUs, and other manufacturers will receive the last batch of supplies in the next few months. But this does not mean that the RTX 3060 will disappear from the market immediately, but will gradually decrease in the next few months.

Since there is still sufficient stock of RTX 3060 at major retailers, there will still be a certain supply on the market after graphics card manufacturers prepare the last batch of RTX 3060 inventory.

This article is an exclusive article of Observer.com and may not be reproduced without authorization.