All the latest news and updates from San Jose

We go from hardware to software, but this is “really” [what] It completely removes the possibility of AI. She leaves the stage and takes a phorical microphone to talk about ROCM, AMD's AI SVP Vamsi Boppana.

These are the instinct MI350 series partners with products that will be on the market in the third quarter of 2025, including Oracle, Dell Technologies, HPE, Cisco, and Asus.

Yu Jin Song explains that the MI350 will become an important part of the company's AI infrastructure (already using the MI300X accelerator).

“I'm also very excited about the MI350X's capabilities,” he says. “It brings all computing power and accelerated memory and support for FP4 and FP6 while maintaining the safety pull factor on the MI 300, and quickly deploys it.”

“We see significant advances that started from[the inference][andstillextendingtoallAIproducts”saysYuJinSong”AMDandMetaarealwaysstronglyconsistentwiththeneuralnetsolution”[inferencing[andnowextendingtoallofyourAIoffering”saysYuJinSong”IthinkAMDandMetahavealwaysbeenstronglyalignedonneuralnetsolutions”[推論から[そして現在もAIのすべての製品に拡張されていることから始まった重要な進歩が見られます」とYuJinSong氏は言います。「AMDとメタは常に強く整合していると思います神経ネットソリューション。」[inferencing[andnowextendingtoallofyourAIoffering”saysYuJinSong”IthinkAMDandMetahavealwaysbeenstronglyalignedonneuralnetsolutions”

“We can try to film problems and deploy an optimized system in schools, so our team considers AMD a strategic, responsive partner and someone we really like.

And now another guest: Yu Jin Song, Meta VP of Engineering – of course, the force behind Lama LLM.

Su claims that the MI350 series “provides a massive 4x jump in AI computing to accelerate both training and reasoning.” Some of the claimed specifications are as follows:

• 288GB memory
• Run up to 20 billion parameters on a single GPU
• Double FP throughput compared to competitors
• 1.6x more memory than competitors

(These are AMD claims that have not been independently verified. Itpro))

“The Mi 350 series offers the largest generational performance lead in the history of instincts, and we're digging deep into the development of the Mi 400 in 2026,” says Su.

“Today, I am extremely excited to launch the Mi 350 series, our most advanced AI platform, providing leadership performance with the most demanding model in this series,” she says.

Sun leaves the stage to have its first product launch, the AMD Instinct MI350 series.

Sun focuses on how his small team is being enhanced by AMD hardware like MI300X and, of course, open

Su is currently talking about how her first guest, Xa's Xia Sun, is on stage and the company behind Grok is working with AMD.

“Linux surpassed UNIX as the system of choice as the data center operating system when global collaboration was unlocked,” she continues. “Android's open platform helped scale mobile computing to billions of users. In each case, openness has resulted in more competition, faster innovation, and ultimately better results for users. That's why, for us at AMD and, frankly, for us as an industry, openness should not be just a buzzword.

SU hits with another theme that has gained popularity among many enterprise hardware providers: open source and openness.

“AMD is the only company dedicated to the openness of hardware, software and solutions in general,” Su claims. “Our industry history shows, again and again, that innovation really starts when things are open.”

AI is now not only a cloud or data center question, but an endpoint question, especially a PC. “We hope that AI will be deployed on all devices,” she says.

Now, like many companies, including partner Dell and rival Nvidia, AMD is discussing (positive) impacts on hardware with AI agents.

Su says Agent AI represents “a user of a new class.”

“What we're actually seeing is adding to our global computing infrastructure, which is equivalent to billions of new virtual users. All of these agents are here to help us.

And this is the woman from time for Dr. Lisa Su, CEO of AMD.

“We are extremely proud that billions of people use AMD technology every day, whether they're talking about services like Microsoft Office 365, Facebook, Zoom, Netflix, Uber, Salesforce or SAP. “At AI, the biggest cloud and AI companies use their instincts to enhance their latest models and new production workloads. There are also a lot of new innovations being made in new AI startups.”

And we're ongoing in the opening video. It seems that trust is the key to the future of AI.

The auditorium is beginning to fill up at the San Jose McKennaly Convention Center as Lisa Su waits for her to take the stage this morning for her keynote address.

With just under two hours before Lisa Su's keynote begins, it's worth considering what happened at the company's final advance AI conference, which took place eight months ago in October 2024.

At that event, Su launched an enterprise-centric Ryzen AI Pro 300 series CPU range, unveiling the Pensando Selina data processing unit, and announced the arrival of the instinctive MI325X accelerator. It is unlikely that the company is inviting customers, partners and presses to hear the latest updates on these products, but it may indicate the direction of your trip.

At Computex in June 2024, Su announced that the company has moved to the GPU annual release cadence and has been able to see updates in this area with the release of the Instinct MI325X Accelerator Series in its final progression AI.