AMD CEO Lisa Su said the $500 million upgrade in the company’s Instinct 2024 GPU sales forecast was based on achieving “certain customer milestones,” such as meeting reliability requirements in data centers and chip optimization on certain AI workloads.

AMD said it now expects to earn more than $5 billion from sales of its Instinct data center GPUs this year due to strong demand from hyperscalers like Meta and Microsoft as well as other customers using the chips to run AI workloads.

This marks an improvement over the $4.5 billion AMD projected for Instinct GPU sales in 2017. July and the $4 billion forecast he released in April as the Santa Clara, Calif.-based chip designer seeks to challenge Nvidia, which saw its revenue more than double year over year to $30 billion in the past half year. only the last quarterlargely due to its continued dominance in the AI ​​computing market.

(Related: AMD Says Instinct MI325X Outperforms Nvidia H200, Promises Huge Improvement With MI350)

AMD revealed the sales dynamics of its Instinct GPUs during Tuesday’s third-quarter earnings call, where the company’s CEO, Lisa Su, said that “significantly higher data center and client CPU sales higher sales more than offset the decline in sales of games and embedded products.

As a result, AMD’s third-quarter revenue was $6.8 billion, 18% higher than the same period last year and 17% higher sequentially. The company’s gross margin increased 3 points year-over-year to 50 percent and its gross profit increased 24 percent year-over-year to $3.4 million, while its earnings per share increased by 161 percent to 47 cents.

It expects to report revenue of about $7.5 billion for the fourth quarter, which would represent an increase of about 22% year-over-year and a sequential increase of 10%.

“This is an incredibly exciting time for AMD, as the breadth of our technology and product portfolios, combined with our deep customer relationships and the diversity of markets we address, provide us with a unique opportunity as we execute our next arc and make AMD the final leader. AI leader all the way,” Su said on the earnings call.

AMD’s third-quarter revenue beat Wall Street expectations by about $100 million, while its adjusted earnings per share were slightly in line with analyst estimates. Its revenue forecast for the fourth quarter, however, is slightly below Wall Street expectations.

The company’s stock price fell more than 7.5 percent after hours Tuesday.

Distribution of business units

AMD said third-quarter data center revenue more than doubled from the same period last year, increasing 122% year-over-year to a record $3.5 billion , thanks to a “strong increase” in Instinct GPU shipments combined with EPYC CPU growth. sales. This is 25 percent higher revenue than the previous quarter, according to the company.

Customer revenue, on the other hand, increased 29% year-over-year and 26% sequentially to $1.9 billion, driven primarily by “strong demand” for its Zen-based Ryzen processors. 5.

The chip designer, however, suffered in integrated revenue, which fell 25% year-over-year to $927 million due to customer work on existing inventory, but increased 8% sequentially due to improving demand. Gaming revenue also fell 69% year-over-year and 29% sequentially, primarily due to lower sales of semi-custom products.

Why AMD boosted GPU Instinct predictions

Su said the $500 million upgrade to AMD’s Instinct GPU sales forecast for 2024 was based on achieving “certain important milestones for customers,” such as meeting reliability requirements in data centers and chip optimization on certain AI workloads. This allowed the company to ramp up Instinct shipments faster than expected.

Two major customers that helped fuel sales in the third quarter were Microsoft and Meta, both of which expanded their use of AMD MI300X GPU for internal workloads, according to Su.

While Microsoft uses the MI300X to run multiple Copilot services powered by OpenAI’s GPT-4 AI models, Meta has “widely deployed the MI300X to power its large-scale inference infrastructure, including using the MI300X exclusively to serving all live traffic for its most demanding Llama (405). frontier model of several billion parameters,” she said.

The company also saw the availability of MI300X public cloud instances grow among Microsoft, Oracle Cloud and smaller cloud providers, while “several startups and industry leaders” such as Databricks and Essential AI led to “strong” adoption of such bodies.

Su said AMD had also received positive feedback from customers regarding its pending acquisition of a cloud computing specialist for $4.9 billion. ZT Systems.

The acquisition, which is expected to close in the first quarter of next year, is expected to enable hyperscaler customers to “rapidly deploy AMD AI infrastructure at scale while providing server vendors with “optimized board and module designs.” for a wide range of differentiated business solutions.” » according to the CEO.

At the same time, Su said, AMD’s plan to divest ZT Systems’ data center infrastructure manufacturing business has “so far attracted significant interest from a number of parts”.

Looking ahead to the next two quarters, Su said AMD customer engagements around Instinct GPUs are “actually expanding quite nicely.”

“Our largest cloud customers are expanding the set of workloads they run on AMD Instinct, and we are also very engaged with a number of large cloud and enterprise customers who are actively working with us and optimizing their workload,” she said. .