The PC is no longer just a device you use. It is becoming a device that works for you. AMD has officially announced the Ryzen AI Halo developer platform and unveiled the new Ryzen AI Max PRO 400 Series processors — designed to power the next generation of Agent Computers: AI-enabled systems capable of understanding prompts, planning actions, and executing tasks with minimal user intervention.
The Ryzen AI Halo system launches at $3,999 for pre-orders starting June 2026, with an upgraded next-generation version powered by Ryzen AI Max PRO 400 Series processors arriving in Q3 2026.
Table of Contents
Two Announcements, One Vision
Ryzen AI Halo — Pre-Orders Open June 2026
The Ryzen AI Halo developer platform launching this June is based on the Ryzen AI Max+ 395 — codenamed Strix Halo — which combines Zen 5 CPU cores, RDNA 3.5 GPU, and an XDNA 2 NPU in a single SoC architecture. It has already seen major adoption in laptops, handhelds, and mini PCs, and is now entering a dedicated developer-focused form factor.
Up to 128GB of unified system memory, AMD ROCm software optimisation, and Windows and Linux support give developers the headroom to run AI models on device. The new AMD Ryzen AI Halo can capably run models of up to 200 billion parameters, allowing developers to build, test, and operate agentic AI applications locally.
Ryzen AI Max PRO 400 Series — The Upgrade Coming in Q3
The next-generation platform powered by Ryzen AI Max PRO 400 Series processors features higher clock speeds, up to 192GB of unified memory, and 160GB of VRAM — with OEM partners launching Agent Computers and workstation-class systems starting this year.
The memory upgrade is the number that matters most. Moving from 128GB to 192GB of unified memory is the specification that decides which models can actually be fitted and run locally — a meaningful jump over both today’s Halo and NVIDIA’s DGX Spark.
Specs at a Glance
| Ryzen AI Halo (June 2026) | Ryzen AI Max PRO 400 (Q3 2026) | |
|---|---|---|
| Processor | Ryzen AI Max+ 395 (Strix Halo) | Ryzen AI Max PRO 400 Series |
| Unified Memory | Up to 128GB | Up to 192GB |
| VRAM | Up to 96GB | Up to 160GB |
| Max Model Size | ~200B parameters | Beyond 200B |
| NPU | XDNA 2 (50 TOPS) | XDNA 2+ |
| Software | ROCm, Windows, Linux | ROCm, Windows, Linux |
| Starting Price | $3,999 | TBA |
| Pre-Orders | June 2026 | Q3 2026 |
Why Agent Computers Change the Economics of AI
Autonomous agents run continuously and consume orders of magnitude more tokens than chat — roughly four to fifteen times more by current estimates, and trending well beyond that. When consumption scales at that rate, per-token cloud pricing becomes a real line item, and owning local inference capacity stops being a hobbyist choice and starts being an economic one.
AMD meets these needs by enabling developers, businesses, and OEMs to deliver advanced AI workloads and intelligent agents to local PCs. For agentic AI to move beyond the cloud, the PC needs to handle real-time tasks locally while responding quickly, protecting sensitive data, and supporting the memory demands of complex agent workflows.
AMD claims the Ryzen AI Halo pays for itself within six months for developers running significant token volumes — positioning the $3,999 entry price as an infrastructure investment rather than a hardware purchase.
Ryzen AI Halo vs NVIDIA DGX Spark
Ryzen AI Halo is designed to compete directly with NVIDIA’s DGX Spark mini PC — a compact AI system targeting similar local AI workloads. Both offer high-performance computing in small form factors with extensive memory capacity for large AI models.
The key differentiator AMD is leading with is value: AMD claims leadership tokens per dollar versus NVIDIA’s Spark, with the ROCm open software stack being a long-term advantage for developers who want to avoid vendor lock-in.
For developers currently running local AI inference on RTX 5060 Ti 16GB, Ryzen AI Halo sits in a completely different tier — not a gaming GPU repurposed for AI, but a dedicated agent computing platform with unified memory architecture that eliminates the VRAM bottleneck entirely.
What Developers Can Build on It
As a demonstration at AMD Developer Day in San Francisco, a single agent generating design concepts scaled into a coordinated team of collaborative agents that produced designs, analysed data, and built applications — all running locally, entirely on AMD hardware. The Ryzen AI Max processors combine CPU, GPU, and AI engine compute so agents can run, reason, and act in real time while keeping data close to the user.
For Indian enterprises and developers building agentic AI products — including those exploring Claude Managed Agents or OpenAI’s enterprise deployment stack — Ryzen AI Halo offers a path to local deployment that eliminates cloud API costs, latency, and data residency concerns simultaneously.
Read the full official announcement at AMD’s blog.
The Bottom Line
AMD Ryzen AI Max PRO 400 Series processors combine AI, graphics, and compute in a single architecture for professional workloads on next-generation Agent Computers, reducing the need for discrete GPUs or cloud-based compute. The $3,999 Ryzen AI Halo is the most credible local AI workstation alternative to NVIDIA’s DGX Spark that has ever been announced — and the 192GB unified memory ceiling on the Max PRO 400 variant arriving in Q3 pushes the addressable model size well beyond anything a discrete GPU can match at this price point.
The Agent Computer era has a start date: June 2026.
Stay tuned to TechnoSports for all AMD AI hardware news, local inference guides, and developer tools coverage.
Word count: ~700
