AMD announced its Radeon Open Compute (ROCm) platform with hardware acceleration support for the Ryzen AI Max 300 "Strix Halo" client processors, and the Radeon RX 9000 series gaming GPUs. For the Ryzen AI Max 300 "Strix Halo," this would unlock the compute power of the 40 RDNA 3.5 compute units, with their 80 AI accelerators, and 2,560 stream processors, besides the AI-specific ISA of the up to 16 "Zen 5" CPU cores, including their full fat 512-bit FPU for executing AVX512 instructions. For the Radeon RX 9000 series, this would mean putting those up to 64 RDNA 4 compute units with up to 128 AI accelerators and up to 4,096 stream processors to use.

AMD also announced that it has updated the ROCm product stack with support for various main distributions of Linux, including OpenSuSE (available now), Ubuntu, and Red Hat EPEL, with the latter two getting ROCm support in the second half of 2025. Lastly, ROCm gets full Windows support, including Pytorch and ONNX-EP. A preview of the Pytorch support can be expected in Q3-2025, while a preview for ONNX-EP could arrive in July 2025.

AMD at Computex 2025 announced the Radeon AI PRO R9700 graphics card. This card is being launched to cover a wide range of use-cases from professional visualization to AI acceleration at the edge. The card is a beefed up variant of the desktop Radeon RX 9070 XT, and maxes out the 4 nm "Navi 48" silicon, enabling all 64 compute units, for a total of 128 AI accelerators. It also gets 32 GB of 20 Gbps GDDR6 memory across the chip's 256-bit wide memory interface for 640 GB/s of memory bandwidth. The card is rated for a total board power of 300 W. At its given specs, AMD claims up to 96 TFLOPs of FP16 throughput, and up to 1,531 AI TOPS (INT4 sparse).

The 32 GB of memory should come in handy for larger AI models, AMD claims, as it illustrates its performance against an NVIDIA GeForce RTX 5080 with its 16 GB of memory—this could also hint at the target pricing of this card. The card supports driver-level scalability for up to 4 such GPUs, for a memory pool of 128 GB, leveraging the interface level features offered by PCIe Gen 5. The 128 GB memory pool should prove sufficient for Mistral Large Instruct (123 billion parameters with GPTQ4), and DeepSeek R1 Distill (Llama 70 billion parameters, FP8), both of which have memory footprints ranging between 112 GB to 116 GB. Much like the Ryzen Threadripper PRO 9000 "Shimada Peak" workstation processors being announced today, the AMD Radeon AI PRO R9700 should be available in July 2025.

AMD at Computex 2025 announced the Ryzen Threadripper PRO 9000 series processors targeting high-end desktops (HEDT) and workstations. These processors are codenamed "Shimada Peak," and are based on the "Zen 5" microarchitecture. "Shimada Peak" is a variation of the EPYC "Turin" MCM, designed for Socket TR5, with slightly modified I/O. The chip puts out up to 128 PCIe Gen 5 lanes but lacks CXL capability on these lanes, and the memory I/O is set to 8-channel DDR5 (16 sub-channels), down from 12-channel DDR5 on "Turin" (24 sub-channels).

The top SKU, the Ryzen Threadripper PRO 9995WX, comes with a 96-core/192-thread configuration, a 128-lane PCIe Gen 5 I/O, and an 8-channel DDR5 memory interface. The chip comes with a maximum boost frequency of 5.40 GHz, and a total of 384 MB of L3 cache (12x 32 MB). AMD claims a significant 10-20% gain in Cinebench 2024 nT performance over the previous generation 7995WX "Zen 4" processor, and nearly 220% gain over the Intel Xeon W9-3595X processor. Across a wide segment of media, design, architecture, science, and LLM inferencing use-cases, the 9995WX is shown posting performance leads ranging anywhere from 44% to 145% over the W9-3595X. The Threadripper PRO 9000 series will be an OEM-exclusive launch through workstation partners such as Lenovo.

AMD at Computex 2025 announced the new Radeon RX 9060 XT mid-range graphics card. The card is designed to offer maxed out gaming at 1080p, with ray tracing enabled, and lets you take advantage of new features such as FSR 4 and the upcoming FSR "Project Redstone" feature-set. The card comes in two variants, the RX 9060 XT 16 GB, priced at $350, and the RX 9060 XT 8 GB, priced at $300. Both models are based on the 4 nm "Navi 44" silicon, which they both max out in terms of on-die components. The GPU is based on the RDNA 4 graphics architecture, and comes with 32 CU (compute units), which works out to 2,048 stream processors, 64 AI accelerators, 32 RT accelerators, 128 TMUs, and possibly 64 ROPs. The chip features a 128-bit wide GDDR6 memory interface, the company didn't reveal memory speeds. Both models come with a total board power value of 180 W. The company claims that the RX 9060 XT 16 GB offers up to 821 peak AI TOPS (INT4).

AMD also put out some first party performance claims. The company claims that the RX 9060 XT 16 GB, should beat the GeForce RTX 5060 Ti 8 GB by 6% on average, tested across 40 game titles, at 1440p. The RTX 5060 Ti 8 GB has an MSRP of $380, making the RX 9060 XT 16 GB cheaper by $30, and as a result, have a 15% performance-per-Dollar edge. The company did not put out any performance claims for the RX 9060 XT 8 GB model. Given that NVIDIA is not developing a 16 GB model of the new RTX 5060 (non-Ti), and its $300 price, things could get interesting for AMD, especially if its claim that the RX 9060 XT 16 GB will be the fastest current-gen GPU under $350 holds. Both the 16 GB and 8 GB variants of the Radeon RX 9060 XT should be available on June 5, 2025.

AMD in its Computex 2025 keynote address announced an overview of FSR 4 implementation. FSR 4 introduces a new AI Machine Learning-based super resolution algorithm that's more accurate, and vastly improves image quality at every performance preset. AMD also announced FSR "Project Redstone." This is a future version extension of FSR, and "Project Redstone" is its working title. "Redstone" combines the AI ML super resolution introduced by FSR 4, with three new features—neural radiance caching, an AI ML-based Ray Regeneration, and AI ML-based Frame Generation. The company plans to launch "Redstone" in the second half of 2025.

All three features being introduced by "Project Redstone" aim to achieve technological parity with NVIDIA DLSS 3.5 and DLSS 4, particularly in AAA games with ray tracing enabled. Neural Radiance Caching sees an AI ML model continually learn how light bounces in a scene to predict and store indirect lighting, which reduces the performance cost of ray tracing. AMD FSR "Redstone" Ray Regeneration is functionally similar to NVIDIA DLSS 3.5 Ray Reconstruction. It uses a neural network to regenerate pixels that couldn't be accurately path-traced. This should improve the quality of reflections, particularly when super resolution is used. The next big announcement of course is a new AI ML-based model for Frame Generation.

Here are some close-up shots we took of AMD's latest client and enterprise processors. Starting things off is the Ryzen AI Max "Strix Halo." This is a rather large chip, with a package size resembling some high-end GPUs. This chiplet-based processor is dominated by a large SoC+iGPU tile built on the TSMC N4P process, and two "Zen 5" CCDs. These are the same CCDs as "Granite Ridge," "Fire Range," and "Turin," which come with full 512-bit data paths for the FPU, and large 32 MB L3 caches per CCD. The SoC+graphics tile has the iGPU with 40 compute units. The package is large not just because of the size of the chip, but also its 256-bit wide LPDDR5X memory interface.

Next up, is the Ryzen AI 9 HX Pro mobile processor, which comes in the regular mobile processor package AMD has been using for several generations now. The chip is based on the "Strix Point" monolithic silicon built on TSMC N4P. This particular SKU has the AMD Pro feature-set, and targets commercial notebooks. Lastly, there's the Ryzen Z2 Extreme. This particular chip is built on an optimized variant of "Strix Point," with a few disabled CPU cores, but a maxed out iGPU with all 16 CU enabled. This processor targets gaming handhelds. Winding things up is the EPYC 9006 "Turin" processor with its glorious 12 "Zen 5" CCDs.

In an almost uncanny case of recent history repeating itself, AMD has kicked off another RDNA 4 new product teaser campaign. Today's reminder—regarding Jack Huynh's upcoming Computex presentation—included a promotional render of a stubbier dual-fan Radeon RX 9000 Series card design. Casting our memories back to late January (2025), Team Red rolled out an ill-timed advertisement—not long after the official delay of Radeon RX 9070 XT and non-XT cards. Despite denying the existence of "Made-By-AMD" (MBA) reference designs, AMD staffers were likely enraged by Chinese black market channel offerings of alleged "real deal examples." Days after first wave RDNA 4's March 6 global launch, a triple-fan specimen was outed.

Throughout early Q2, insiders and members of the Chiphell forum have played around with Radeon RX 9070 XT and RX 9070 MBA cards. The latter unit (see photo below) seems to utilize a dual-fan configuration in a fairly long enclosure format. Team Red's latest promo post likely points to a forthcoming unveiling of Radeon RX 9060 XT partner models, but curious industry observers will be wondering whether the shorter reference design actually exists in real life. As per usual, a tiny disclaimer claims otherwise: "Artistic Render. Not Available For Purchase." So far, leaks have suggested the presence of Acer, ASUS, GIGABYTE, and XFX custom options during introductory proceedings. Today's refresher outlined upcoming new product categories and partner contributions: "join AMD on May 21 as we reveal what's next in gaming, AI PCs, and more. (Our) SVP and GM of Computing and Graphics Group, along with industry leaders and partners showcase what is built to power the next level."

Insider whispers regarding AMD's plans for not-yet-official Radeon RX 9060 XT graphics cards have floated a possible introduction on May 18, as well as a speculated early June retail launch. These leaked timeframes seem to hinge on something happening at next month's Computex 2025 trade show—similarly, NVIDIA could be readying competing products. BenchLife.info has disclosed inside track information from both camps—their report outlined vaguely refreshed conditions: "the entire 'Blackwell' gaming GPU range is expected to include an (unannounced) GeForce RTX 5050, but this particular graphics card has not yet been confirmed by NVIDIA. After the launch of Team Green's GeForce RTX 5060 8 GB model, AMD is also expected to launch the Radeon RX 9060 XT series graphics cards on May 21, but the finalized stock may not appear for sale until June."

Yesterday, Team Red and involved board partners surprised many industry watchdogs—up until very recently, the Radeon RX 9070 GRE 12 GB model's release status was in flux. VideoCardz believes that misleading information was deliberately sent out to moles and leakers. As correctly theorized by certain outlets, AMD will launch this (RDNA 4) generation's "Great Radeon Edition" card well in advance of lower end options. The aforementioned (alleged) May 21 introduction of Radeon RX 9060 XT cards aligns with AMD's scheduled Tuesday morning press conference. A few days ago, a smattering of news articles indicated the cancelation of Team Red's Radeon RX 9060 XT 8 GB variant—VideoCardz cannot verify this claim. Instead, they reckon that the 16 GB model will be prioritized; in terms of day zero/one media coverage.

AMD today announced that it will be hosting a press conference during Computex 2025. The in-person and livestreamed press conference will take place on Wednesday, May 21, 2025, at 11 a.m. UTC+8, Taipei, at the Grand Hyatt, Taipei. The event will showcase the advancements AMD has driven with AI in gaming, PCs and professional workloads.

AMD senior vice president and general manager of the Computing and Graphics Group Jack Huynh, along with industry partners, will discuss how AMD is expanding its leadership across gaming, workstations, and AI PCs, and highlight the breadth of the company's high-performance computing and AI product portfolio. The livestream will start at 8 p.m. PT/11 p.m. ET on Tuesday, May 20 on AMD.com, with replay available after the conclusion of the livestream event.