AI at AMD

We have been busy building and today we’re excited to share a new set of AI playbooks designed to help developers and builders get hands-on with local AI on AMD hardware. These hands-on guides walk through: → LLM fine-tuning and PEFT workflows using Google Gemma-3 on AMD Radeon™ GPUs. → Building a local multimodal AI stack using Lemonade, supporting LLMs, vision, image generation, and speech via Open WebUI → Fast SFT workflows using LoRA on AMD hardware with Unsloth Gemma-4E4B and the FineTome-100k dataset. → Low-latency translation pipelines using AMD ROCm + PyTorch across Ryzen AI and Radeon GPUs. → Running GenAI locally with Lemonade and switch between Vulkan, ROCm™, and NPU execution modes Whether you're just getting started or looking to push the boundaries of on-device AI, there's a playbook for you. More coming soon, explore the full library and start building today: developer.amd.com/playbooks/

Not every AI workload needs to live in the cloud. The future of AI is hybrid. This blog explores the rise of “Agent Computers” and why developers are increasingly looking at local inference for better cost control, lower latency, and continuous AI workflows. amd.com/en/blogs/2026/…

What do people misunderstand about AI right now? We asked #AMDevs to weigh in.

Join us on May 26 in NYC for a panel discussion on what it really takes to deploy agents safely and reliably at enterprise scale. We'll be joined by an incredible group of leaders and experts from @nvidia, @vercel, @AMD, and @alumniventures. You can expect practical insights, candid discussion, and networking with builders shaping the next generation of enterprise AI. Space is limited so reserve your spot, today! 👇 hubs.la/Q04h2z7L0

Check out the AMD Developer Community on Discord: discord.com/invite/amd-dev

Learn what matters. Build what’s next. AMD AI Academy brings together developer-focused learning across agents, LLMs, inference, optimization, deployment, and more. Join the AMD AI Developer Program community and start exploring AMD AI Academy: bit.ly/4dDYGO2

Local AI is changing how developers build and deploy intelligent systems. @jackhuynh shares how the new AMD Ryzen AI Halo is helping redefine what #AMDevs can build and run locally. 👇

@MichaelGannotti @Hi_MINISFORUM @AMDRyzen @openclaw @NousResearch @gmktecdotcom This is serious power!

I have two separate teams of AI running on two MiniPCs. On top is the @Hi_MINISFORUM @AMDRyzen 7 unit with 32 gb ram and a 512 gb ssd. Running Windows 11 and attached to Microsoft 365/Azure it has 1 @openclaw AI and 1 @NousResearch Hermes AI. Underneath it is my @gmktecdotcom MiniPC with an @AIatAMD Ryzen 9 AI+ 395 chipset with 96gb ram and 1 tb ssd. This unit is running Ubuntu Linux attached to Google Workspace and has 3 Hermes and 4 OpenClaw AI. I want to beef up the AI for my Microsoft environment but I need a beefier rig to add to that. I’d really like to set up several full teams attached to @Azure and @microsoft 365 but going to need some beefier hardware for that. For now though things are progressing nicely 👍

This week was also special as we celebrated the 20th anniversary of our Shanghai R&D center and 10 years of partnership with TFME. So proud of our @AMD team and partners for all we’ve built together.

Great being in China this week with customers, partners, developers, and our @AMD teams across Beijing, Shanghai, and Suzhou. Loved the energy at AMD DevDay in Shanghai showcasing the progress we’re making to bring AI Everywhere. Thanks to everyone who joined us.

The future of AI will be built on open ecosystems, practical developer workflows, and infrastructure ready for production. That’s exactly what AMD is bringing to @Microsoft Build this year. From AI PCs to AI development across cloud and client, and the open software ecosystem supporting it, we’re excited to be part of the conversation at #MicrosoftBuild. Explore more and add our sessions to your agenda: bit.ly/3ROTNdI

Missed AMD AI DevDay? The technical sessions are now live on YouTube — watch them on demand today. youtube.com/playlist?list=… And if you missed DevDay, don’t miss what’s next. Register now for AMD Advancing AI this July in San Francisco. Expect even more technical content, deeper developer sessions, and 30+ hands-on workshops available exclusively onsite. Build. Train. Deploy. Advance. 🚀

Lo-fi. EDM. Movie soundtracks. White noise. This is what #AMDevs listen to when they need to lock in. What's in your headphones when you need to reach flow state?

What does your camera reveal to a vision encoder? Join the next AMD AI Research Club session where Pier Luigi Dovesi, Robotics and Autonomous Systems Lead, @AMDSiloAI hosts Giorgos Tolias, Associate Professor at the Czech Technical University in Prague, for a deep dive into hidden acquisition and processing traces in modern visual foundation models. 📅 May 28, 2026 ⏰ 7 AM CT | 2 PM CET We’ll unpack: • Hidden metadata traces inside visual embeddings • Camera pipelines, compression artifacts, and acquisition signals • How shortcut learning can influence downstream AI predictions Built for AI researchers, developers, and engineers exploring robustness, retrieval, synthetic image detection, and trustworthy computer vision systems. Register here: amd.zoom.us/webinar/regist…

We just hit a new milestone: exact 35-qubit quantum simulation on a single AMD Instinct MI355X GPU running on Qibo, the largest single-GPU result on AMD hardware to date. Read the full breakdown: amd.com/en/developer/r…

Great AI starts with deeper understanding. Through the partnership between AMD and @DeepLearningAI, @realSharonZhou helps developers and engineers connect transformer fundamentals to real-world AI decisions in this newly launched course. #ExperienceOpen #AMDevs

What are #AMDevs saying about ROCm? From “one-line install” to “just works," here's what we Overheard at AMD #AIDevDay.

Meet Henri Sirkkavaara. Henri created Vaara, a real-time signal analysis platform that extracts calibrated directional signals from high-noise, multi-source time-series streams. The full research-to-production stack — ingestion, feature engineering, model training, and low-latency inference — runs continuously on AMD Ryzen + Radeon compute under ROCm software, architected to scale across domains, markets, and data modalities. What is he excited to build next? Scaling the platform across additional domains and data modalities, and hardening the developer-facing surface so external teams can integrate the signal layer into their own decision workflows. Join the AMD AI Developer Program here: amd.com/en/developer/a…

What are you building right now? From open models to on-device agents. Overheard with #AMDevs.

An open-source 3D printer made from 92% LEGO bricks that actually works. 18-year-old @hackclub student Adam built Studprint, which he designed and rendered on a PC powered by AMD Ryzen 5 7600X. Full open-source build on @github here: github.com/AdamTuraj/Stud…