Lisa Su displays an ADM Instinct M1300 chip as she delivers a keynote address at CES 2023 at The Venetian Las Vegas on January 04, 2023 in Las Vegas, Nevada.
David Becker | Getty Images
AMD said on Tuesday its most-advanced GPU for artificial intelligence, the MI300X, will start shipping to some customers later this year.
AMD’s announcement represents the strongest challenge to Nvidia, which currently dominates the market for AI chips with over 80% market share, according to analysts.
related investing news
GPUs are chips used by firms like OpenAI to build cutting-edge AI programs such as ChatGPT.
If AMD’s AI chips, which it calls “accelerators,” are embraced by developers and server makers as substitutes for Nvidia’s products, it could represent a big untapped market for the chipmaker, which is best known for its traditional computer processors.
AMD CEO Lisa Su told investors and analysts in San Francisco on Tuesday that AI is the company’s “largest and most strategic long-term growth opportunity.”
“We think about the data center AI accelerator [market] growing from something like $30 billion this year, at over 50% compound annual growth rate, to over $150 billion in 2027,” Su said.
While AMD didn’t disclose a price, the move could put price pressure on Nvidia’s GPUs, such as the H100, which can cost $30,000 or more. Lower GPU prices may help drive down the high cost of serving generative AI applications.
AI chips are one of the bright spots in the semiconductor industry, while PC sales, a traditional driver of semiconductor processor sales, slump.
Last month, AMD CEO Lisa Su said on an earnings call that while the MI300X will be available for sampling this fall, it would start shipping in greater volumes next year. Su shared more details on the chip during her presentation on Tuesday.
“I love this chip,” Su said.
The MI300X
AMD said that its new MI300X chip and its CDNA architecture were designed for large language models and other cutting-edge AI models.
“At the center of this are GPUs. GPUs are enabling generative AI,” Su said.
The MI300X can use up to 192GB of memory, which means it can fit even bigger AI models than other chips. Nvidia’s rival H100 only supports 120GB of memory, for example.
Large language models for generative AI applications use lots of memory because they run an increasing number of calculations. AMD demoed the MI300x running a 40 billion parameter model called Falcon. OpenAI’s GPT-3 model has 175 billion parameters.
“Model sizes are getting much larger, and you actually need multiple GPUs to run the latest large language models,” Su said, noting that with the added memory on AMD chips developers wouldn’t need as many GPUs.
AMD also said it would offer an Infinity Architecture that combines eight of its M1300X accelerators in one system. Nvidia and Google have developed similar systems that combine eight or more GPUs in a single box for AI applications.
One reason why AI developers have historically preferred Nvidia chips is that it has a well-developed software package called CUDA that enables them to access the chip’s core hardware features.
AMD said on Tuesday that it has its own software for its AI chips that it calls ROCm.
“Now while this is a journey, we’ve made really great progress in building a powerful software stack that works with the open ecosystem of models, libraries, frameworks and tools,” AMD president Victor Peng said.
China is one of Nvidia’s largest markets, particularly for data centers, gaming and artificial intelligence applications.
Avishek Das | Lightrocket | Getty Images
Two Chinese nationals in California have been arrested and charged with the illegal shipment of tens of millions of dollars‘ worth of AI chips, including from Nvidia, the Department of Justice said Tuesday.
Chuan Geng, 28, and Shiwei Yang, 28, exported the sensitive chips and other technology to China from October 2022 through July 2025 without obtaining the required licenses, the DOJ said.
The illicit shipments included Nvidia’s H100 general processing units, according to a criminal complaint provided to CNBC. The H100 is amongst the U.S. chipmaker’s most cutting-edge chips used in artificial intelligence applications.
The Department of Commerce has placed such chips under export controls since 2022 as part of broader efforts by the U.S. to restrict China’s access to the most advanced semiconductor technology.
This case demonstrates that smuggling is a “nonstarter,” Nvidia told CNBC. “We primarily sell our products to well-known partners, including OEMs, who help us ensure that all sales comply with U.S. export control rules.”
“Even relatively small exporters and shipments are subject to thorough review and scrutiny, and any diverted products would have no service, support, or updates,” the chipmaker added.
Geng and Yang’s California-based company, ALX Solutions, had been founded shortly after the U.S. chip controls first came into place.
According to the DOJ, law enforcement searched ALX Solutions’ office and seized phones belonging to Geng and Yang, which revealed incriminating communications between the defendants, including those about evading U.S. export laws by shipping sensitive chips to China through Malaysia.
The review also showed that in December 2024, ALX Solutions made over 20 shipments from the U.S. to shipping and freight-forwarding companies in Singapore and Malaysia, which the DOJ said are commonly used as transshipment points to conceal illicit shipments to China.
ALX Solutions did not appear to have been paid by entities they purportedly exported goods to, instead receiving numerous payments from companies based in Hong Kong and China.
The U.S. Department of Commerce’s Bureau of Industry and Security and the FBI are continuing to investigate the matter.
The smuggling of advanced microchips has become a growing concern in Washington. According to a report from the Financial Times last month, at least $1 billion worth of Nvidia’s chips entered China after Donald Trump tightened chip export controls earlier this year.
In response to the report, Nvidia had said that data centers built with smuggled chips were a “losing proposition” and that it does not support unauthorized products.
With Opendoor shares up almost fivefold since the beginning of July and trading volumes hitting record levels, CEO Carrie Wheeler thanked investors for their “enthusiasm” on Tuesday’s earnings call.
“I want to acknowledge the great deal of interest in Opendoor lately and that we’re grateful for it,” Wheeler said, even as the stock sank more than 20% after hours. “We appreciate your enthusiasm for what we’re building, and we’re listening intently to your feedback.”
Prior to its recent surge, Opendoor’s stock had been mostly abandoned, falling as low as 51 cents in late June. The situation was so dire that the company was considering a reverse split that could lift the price of each share by as much 50 times as a potential way to keep its Nasdaq listing. Opendoor said last week that it’s back in compliance and canceled the reverse split proposal.
Opendoor’s business is centered around using technology to buy and sell homes, pocketing the gains. The company was founded in 2014 and went public through a special purpose acquisition company (SPAC) during the Covid-era boom of late 2020. But when interest rates began climbing in 2022, higher borrowing costs reduced demand for homes.
Revenue sank by about two-thirds from $15.6 billion in 2022 to $5.2 billion last year.
Much of the stock’s bounce in the past six weeks was spurred by hedge fund manager Eric Jackson, who announced in July that his firm had taken a position in Opendoor. Jackson said he believes Opendoor’s stock could eventually get to $82. It closed on Tuesday at $2.52, before dropping below $2 in extended trading.
Jackson’s bet is that a return to revenue growth and increased market share will lead to profitability, and that investors will start ascribing a reasonable sales multiple to the business.
The turnaround isn’t yet showing much evidence of working. For the second quarter, Opendoor reported a revenue increase of about 4% to $1.57 billion. Its net loss narrowed to $29 million, or 4 cents a share, from $92 million, or 13 cents, a year earlier.
In the current quarter, Opendoor is projecting just $800 million to $875 million in revenue, which would represent a decline of at least 36% from a year earlier. Opendoor said it expects to acquire just 1,200 homes in the the third quarter, down from 1,757 in the second quarter and 3,504 in the third quarter of 2024. It’s also pulling down marketing spending.
“The housing market has further deteriorated over the course of the last quarter,” finance chief Selim Freiha said on Tuesday’s earnings call. “Persistently high mortgage rates continue to suppress buyer demand, leading to lower clearance and record new listings.”
Wheeler highlighted Opendoor’s effort to expand its business beyond so-called iBuying and into more of a referrals business that’s less capital intensive. She called it “the most important strategic shift in our history.”
Investors, who have been bidding up the stock in waves, were less than enthused with what they heard. But at least there are finally people listening.
“This increased visibility is an opportunity to tell our story to a broader audience,” Wheeler said. “We intend to make the most of it.”
Super Micro Computer shares slid 15% in extended trading on Tuesday after the server maker reported disappointing fiscal fourth-quarter results and issued weak quarterly earnings guidance.
Here’s how the company did in comparison with LSEG consensus:
Earnings per share: 41 cents adjusted vs. 44 cents expected
Revenue: $5.76 billion vs. $5.89 billion expected
Super Micro’s revenue increased 7.5% during the quarter, which ended on June 30, according to a statement.
For the current quarter, Super Micro called for 40 cents to 52 cents in adjusted earnings per share on $6 billion to $7 billion in revenue for the fiscal first quarter. Analysts surveyed by LSEG were looking for 59 cents per share and $6.6 billion in revenue.
For the 2026 fiscal year, Super Micro sees at least $33 billion in revenue, above the LSEG consensus of $29.94 billion.
Super Micro saw surging demand starting in 2023 for its data center servers packed with Nvidia for handling artificial intelligence models and workloads. Growth has since slowed.
The company avoided being delisted from the Nasdaq after falling behind on quarterly financial filings and seeing the departure of its auditor.
As of Tuesday’s close, Super Micro shares were up around 88% so far in 2025, while the S&P 500 index has gained 7%.
Executives will discuss the results on a conference call starting at 5 p.m. ET.