AMD Steam Machine GPU

The AMD Steam Machine GPU is a console-class graphics processor designed for Valve's Steam Machine platform, built on the RDNA 3.0 architecture with the Navi 33 chip. It sits at the 50th percentile of all GPUs in the database, placing it exactly at the median of the performance distribution. With a 110 W TDP, no external power connectors, and no display outputs, this is a purpose-built embedded solution rather than a retail graphics card. The data reveals a balanced mid-range compute profile: 17.56 TFLOPS of FP32 throughput, 288.0 GB/s of memory bandwidth, and 8 GB of GDDR6 VRAM. Its production status is Active, with a release date of 2025-12-31.

How It Compares

The dataset includes no specific nearest rivals for this GPU. Instead, the percentile field provides the sole comparative metric: a value of 50 against all GPUs. This indicates that the AMD Steam Machine GPU sits precisely at the median of the database's performance distribution. In practical terms, this means it outperforms half of all GPUs and is outperformed by the other half, a position that typically corresponds to a solid mainstream performer. Without rival names or scores, we cannot offer delta percentages, but the percentile alone establishes its mid-tier standing. The absence of direct comparisons also suggests that the GPU occupies a niche role—console integration—where head-to-head with discrete cards is less relevant than its fit within a fixed hardware ecosystem.

Power and Cooling

The GPU's thermal design is governed by a 110 W TDP, a figure that places it in the low-to-moderate power range. Notably, the power connectors field is listed as "None," meaning the card draws all its power from the motherboard slot and requires no auxiliary PCIe power cables. This simplifies installation in compact Steam Machine enclosures, where cable management is often constrained. The suggested PSU field is null, so no specific wattage recommendation is provided; however, a 110 W TDP without extra connectors implies that a standard ATX or SFX power supply with a 6-pin or 8-pin CPU/motherboard header will suffice, assuming the rest of the system is within typical console power budgets. The physical dimensions—156 mm in length, 152 mm in height, and 162 mm in width—indicate a small, board-level footprint, further reinforcing its suitability for tightly integrated designs. No slot width is listed, but the compact size and lack of outputs suggest a low-profile, passive or blower-style cooling solution is likely, though the data does not specify the cooler type.

Benchmark Performance

While the benchmark array is empty, the compute and rendering metrics provide a clear picture of raw capability. The FP32 throughput of 17.56 TFLOPS is a strong mid-range figure, and the FP16 rate of 35.12 TFLOPS (2:1 ratio) doubles that, indicating support for accelerated half-precision workloads. The pixel rate of 156.8 GPixel/s and texture rate of 274.4 GTexel/s are derived from the 1792 shading units, 112 TMUs, and 64 ROPs. These numbers, combined with the 288.0 GB/s memory bandwidth, suggest that the GPU can handle 1080p and 1440p gaming at high settings, though the data does not include frame rates or resolution-specific results. The 50th percentile ranking aligns with this interpretation: it is neither a low-end part nor a high-end enthusiast GPU, but a dependable mid-range performer. The absence of benchmark scores means we cannot compare it to specific rivals, but the theoretical peak numbers indicate it should comfortably exceed last-generation console GPUs in raw compute, given the same TDP envelope.

FAQ

Q: What is the process node and die size of the AMD Steam Machine GPU?

A: The GPU is manufactured on TSMC's 6 nm process, with a die size of 204 mm² and a transistor count of 13,300 million, yielding a transistor density of 65.2 million per square millimeter.

Q: Does the GPU require any external power connectors?

A: No. The power connectors field is listed as "None," meaning it draws all power from the motherboard slot and does not need auxiliary PCIe power cables.

Q: What are the supported graphics APIs?

A: It supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Q: How much memory does it have and what is the bus width?

A: It has 8 GB of GDDR6 memory on a 128-bit bus, providing a bandwidth of 288.0 GB/s.

Q: Does it have dedicated ray tracing cores?

A: Yes, it includes 28 ray tracing cores. The tensor core field is null, indicating no dedicated tensor hardware.

Q: What are the clock speeds?

A: The base clock is 1720 MHz, boost clock is 2450 MHz, and game clock is 2250 MHz. The memory clock is 2250 MHz with an effective data rate of 18 Gbps.

Ray Tracing and Feature Set

The GPU integrates 28 dedicated ray tracing cores, a key feature for modern real-time ray-traced effects. This, combined with DirectX 12 Ultimate (12_2) support, ensures compatibility with the latest ray tracing and mesh shader features in PC and console titles. Vulkan 1.4 support further broadens its API coverage, while OpenGL 4.6 provides legacy compatibility. Notably, the tensor core field is null, meaning there are no dedicated AI acceleration cores for features like DLSS or FSR-based frame generation that rely on tensor hardware. This does not preclude the use of FSR (which runs on standard shaders), but it limits any hardware-accelerated AI workloads. The RDNA 3.0 architecture's dual-issue FP16 capability (35.12 TFLOPS) may assist in some compute-heavy effects, but the lack of tensor cores is a clear differentiator from NVIDIA's RTX lineup. Overall, the feature set is well-rounded for a console GPU, with ray tracing and modern API support, but without the AI-specific hardware found in some discrete cards.

Who Should Consider It

Given its console-specific design, this GPU is not intended for retail consumers building their own PCs. It is a component for Valve's Steam Machine, a pre-built gaming system. The 50th percentile ranking and the compute metrics suggest it is suited for 1080p and 1440p gaming at high to ultra settings, with the ability to handle ray tracing at reduced performance levels. The 8 GB VRAM and 288 GB/s bandwidth are adequate for current titles at these resolutions, though they may become limiting at 4K with high-resolution textures. The lack of display outputs means it cannot be used as a standalone graphics card; it must be paired with a motherboard that routes video through the system's integrated ports. For gamers who prefer a console-like experience with SteamOS, this GPU offers a balanced performance profile that should handle the majority of the Steam library without issue. It is not a high-end part, but its mid-range positioning aligns with the Steam Machine's goal of delivering a plug-and-play PC gaming experience at a reasonable price point.

Memory Subsystem

The memory configuration consists of 8 GB of GDDR6 VRAM on a 128-bit bus, yielding a bandwidth of 288.0 GB/s. This is a modest bandwidth figure for a mid-range GPU, but it is sufficient for 1080p and 1440p gaming. At higher resolutions, such as 4K, the 8 GB capacity may become a constraint for games with large texture packs or ray tracing, as the bandwidth and capacity are both below what high-end cards offer. The effective memory clock of 18 Gbps is typical for GDDR6, and the 128-bit bus width is a deliberate trade-off to keep power consumption low—the 110 W TDP is a key design goal. For a console GPU, this memory configuration strikes a balance between cost, power, and performance. The 288.0 GB/s bandwidth supports the pixel rate of 156.8 GPixel/s and texture rate of 274.4 GTexel/s without bottlenecking in most scenarios. However, users targeting 4K should be aware that the memory subsystem may become a limiting factor, especially in future titles that demand more than 8 GB of VRAM. The absence of any memory error correction or special features is not noted, so the design remains straightforward.