AMD Steam Machine GPU vs NVIDIA GeForce RTX 5090 SE Comparison

Head-to-Head Benchmarks

The recorded data shows that the AMD Steam Machine GPU and the NVIDIA GeForce RTX 5090 SE are both listed with an equal percentile ranking of 50 against all GPUs in the database. Both products also carry identical average benchmark scores of zero, indicating that no direct head-to-head measurement data has been recorded for this pairing. The database shows zero wins for each side, with the win count set to 0 for both the AMD Steam Machine GPU and the NVIDIA GeForce RTX 5090 SE.

The absence of benchmark scores means the comparison relies entirely on the specification sheets contained in the database. The NVIDIA GeForce RTX 5090 SE delivers 66.94 TFLOPS of FP32 compute, which is 3.8 times the 17.56 TFLOPS recorded for the AMD Steam Machine GPU. The texture throughput follows a similar pattern: the RTX 5090 SE reaches 1,045.9 GTexel/s against 274.4 GTexel/s for the AMD part, a 3.8x advantage. Pixel fill rate favors the NVIDIA card at 380.3 GPixel/s versus 156.8 GPixel/s, a 2.4x lead.

Memory bandwidth is another decisive separation. The RTX 5090 SE moves data at 1.34 TB/s across a 384-bit bus, while the AMD Steam Machine GPU provides 288.0 GB/s over a 128-bit interface. That is a 4.7x difference in raw memory throughput. The NVIDIA card also carries 24 GB of GDDR7 memory, three times the 8 GB of GDDR6 found on the AMD design.

The compute unit counts reinforce the performance gap. The RTX 5090 SE uses 14,080 shading units, 440 texture mapping units, and 160 ROPs. The AMD Steam Machine GPU contains 1,792 shading units, 112 TMUs, and 64 ROPs. The NVIDIA part has 7.9x more shading units, 3.9x more TMUs, and 2.5x more ROPs. Ray tracing hardware also differs substantially: 110 RT cores on the NVIDIA chip versus 28 RT cores on the AMD chip, a 3.9x difference. Tensor cores are present only on the NVIDIA side, with 440 units recorded; the AMD Steam Machine GPU lists no tensor core count.

Clock speeds provide the one area where the AMD part holds a nominal advantage. The AMD Steam Machine GPU has a boost clock of 2450 MHz and a game clock of 2250 MHz. The RTX 5090 SE boosts to 2377 MHz and lists no game clock. The base clocks are nearly identical, with the NVIDIA card at 1740 MHz and the AMD card at 1720 MHz. The AMD memory runs at 2250 MHz (18 Gbps effective), while the NVIDIA memory operates at 1750 MHz (28 Gbps effective). The higher effective memory speed on NVIDIA is offset by its wider bus, producing the large bandwidth advantage already noted.

Where Each One Wins

The AMD Steam Machine GPU wins only in the narrow category of clock frequency. Its 2450 MHz boost clock is 73 MHz higher than the 2377 MHz boost on the RTX 5090 SE, and its 2250 MHz game clock has no direct counterpart on the NVIDIA specification sheet. The AMD part also draws far less power, with a 110 W TDP against 500 W for the NVIDIA card. That 390 W difference suggests the AMD design operates within a much tighter thermal envelope, which is consistent with its console-oriented generation classification.

The NVIDIA GeForce RTX 5090 SE wins in every measured throughput and capacity category. It has more memory, a wider memory bus, higher bandwidth, more shading units, more TMUs, more ROPs, more RT cores, and the only tensor cores in the comparison. Its FP32 and FP16 performance are both listed at 66.94 TFLOPS with a 1:1 ratio, matching the 1:1 ratio on the AMD side but at a 3.8x higher magnitude. The NVIDIA card also supports PCIe 5.0 x16 as its bus interface, while the AMD Steam Machine GPU lists no bus interface in the database.

The transistor counts reflect different design philosophies. The RTX 5090 SE contains 92,200 million transistors on a 750 mm² die, manufactured on a 5 nm process at TSMC. The AMD Steam Machine GPU packs 13,300 million transistors into 204 mm² on a 6 nm process, also at TSMC. The transistor density is 122.9M per mm² for NVIDIA and 65.2M per mm² for AMD. The NVIDIA chip is 3.7x larger in die area and holds 6.9x more transistors, yet the process node difference is only one generation step, from 6 nm to 5 nm.

The power connector situation also separates the two. The RTX 5090 SE requires a single 16-pin connector and a suggested 900 W power supply. The AMD Steam Machine GPU lists no power connectors and no suggested PSU, consistent with its lower 110 W TDP.

The Verdict

The database shows a clear performance hierarchy. The NVIDIA GeForce RTX 5090 SE is the stronger product in every compute, memory, and feature category that has recorded values. Its 66.94 TFLOPS FP32 output, 1.34 TB/s bandwidth, and 24 GB memory capacity place it in a different performance class from the AMD Steam Machine GPU, which records 17.56 TFLOPS, 288.0 GB/s, and 8 GB respectively.

The AMD Steam Machine GPU should be selected by users who need a compact, low-power part. Its 110 W TDP, 156 mm length, and lack of external power connectors make it suitable for integrated console-style systems. The RTX 5090 SE, at 500 W TDP, 267 mm length, and a required 16-pin connector, demands a substantial power supply and chassis space.

The RTX 5090 SE is the choice for workloads that stress raw throughput: high-resolution rendering, ray tracing, and tensor-based compute. The AMD part is the choice for constrained environments where power draw and physical size are the primary limits. The absence of tensor cores on the AMD Steam Machine GPU means any tensor-accelerated workload falls exclusively to the NVIDIA card.

The production status for both parts is listed as Active. The RTX 5090 SE has a launch MSRP of 1,499 USD. The AMD Steam Machine GPU has no launch MSRP recorded. The RTX 5090 SE sits in the GeForce 50-series with a predecessor in GeForce 40 and a successor in GeForce 60, while the AMD Steam Machine GPU has no listed predecessor or successor, consistent with its classification as a console GPU for Valve.

FAQ

Q: Which GPU has higher FP32 compute performance?

A: The NVIDIA GeForce RTX 5090 SE records 66.94 TFLOPS of FP32 performance, which is 3.8 times the 17.56 TFLOPS listed for the AMD Steam Machine GPU.

Q: How much memory does each GPU have?

A: The AMD Steam Machine GPU has 8 GB of GDDR6 memory on a 128-bit bus. The NVIDIA GeForce RTX 5090 SE has 24 GB of GDDR7 memory on a 384-bit bus.

Q: What are the power consumption figures?

A: The AMD Steam Machine GPU has a 110 W TDP and lists no power connectors. The NVIDIA GeForce RTX 5090 SE has a 500 W TDP, requires a 16-pin connector, and suggests a 900 W power supply.

Q: Do both GPUs support the same APIs?

A: Both list DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4 in the database.

Q: Which GPU has tensor cores?

A: Only the NVIDIA GeForce RTX 5090 SE lists tensor cores, with 440 units recorded. The AMD Steam Machine GPU has no tensor core count in the database.

Q: What are the physical dimensions of each card?

A: The AMD Steam Machine GPU measures 156 mm in length, 152 mm in height, and 162 mm in width. The NVIDIA GeForce RTX 5090 SE measures 267 mm in length, 111 mm in height, and 40 mm in width.

Architecture Differences

The two GPUs belong to different architecture generations. The AMD Steam Machine GPU uses the Navi 33 chip built on RDNA 3.0 architecture, with the codename Hotpink Bonefish. The NVIDIA GeForce RTX 5090 SE uses the GB202 chip built on Blackwell 2.0 architecture. The AMD part is classified as a Console GPU for Valve, while the NVIDIA part belongs to the GeForce 50-series generation.

The manufacturing processes differ by one node step. AMD uses a 6 nm process at TSMC, while NVIDIA uses a 5 nm process at TSMC. The transistor counts scale accordingly: 13,300 million for AMD versus 92,200 million for NVIDIA. Die size also differs substantially, with the AMD chip at 204 mm² and the NVIDIA chip at 750 mm². Transistor density is 65.2M per mm² for AMD and 122.9M per mm² for NVIDIA.

The compute architecture differs in unit organization. AMD implements 1,792 shading units, 112 TMUs, 64 ROPs, and 28 RT cores. NVIDIA implements 14,080 shading units, 440 TMUs, 160 ROPs, 110 RT cores, and 440 tensor cores. The tensor cores on the NVIDIA side provide dedicated hardware that the AMD part lacks entirely.

Memory architecture also diverges. AMD uses GDDR6 at 2250 MHz with an 18 Gbps effective rate, producing 288.0 GB/s across a 128-bit bus. NVIDIA uses GDDR7 at 1750 MHz with a 28 Gbps effective rate, producing 1.34 TB/s across a 384-bit bus. The NVIDIA memory subsystem delivers 4.7x the bandwidth of the AMD design.

The clock behavior differs in configuration. AMD lists base, boost, and game clocks (1720 MHz, 2450 MHz, and 2250 MHz respectively). NVIDIA lists base and boost clocks (1740 MHz and 2377 MHz) but no game clock. The AMD part has the higher boost clock by 73 MHz.

Specification Differences

The power delivery systems are fundamentally different. The AMD Steam Machine GPU has a 110 W TDP with no power connectors listed. The NVIDIA GeForce RTX 5090 SE has a 500 W TDP, a single 16-pin connector, and a suggested 900 W power supply. The NVIDIA card is dual-slot, while the AMD card has no slot width recorded.

The bus interfaces differ. The RTX 5090 SE uses PCIe 5.0 x16. The AMD Steam Machine GPU has no bus interface listed in the database.

Display outputs differ in count and version. The AMD card provides 1x HDMI 2.1a and 1x DisplayPort 2.1. The NVIDIA card provides 1x HDMI 2.1b and 3x DisplayPort 2.1b.

Physical dimensions show the size gap. The AMD card is 156 mm long, 152 mm tall, and 162 mm wide. The NVIDIA card is 267 mm long, 111 mm tall, and 40 mm wide. The AMD part is shorter but taller and wider, while the NVIDIA part is much longer and thinner.

The release dates are separated by about six months. The AMD Steam Machine GPU released on 2026-06-28, and the NVIDIA GeForce RTX 5090 SE released on 2025-12-31. The NVIDIA card has a predecessor (GeForce 40) and a successor (GeForce 60) listed, while the AMD card has neither. The NVIDIA card has a launch MSRP of 1,499 USD; the AMD card has no launch MSRP recorded.