Intel Arc Graphics 1 Xe Mobile vs NVIDIA Switch 2 GPU Comparison

Where Each One Wins

The Intel Arc Graphics 1 Xe Mobile and the NVIDIA Switch 2 GPU occupy entirely different corners of the hardware landscape. The Intel part is an integrated graphics processor, built into a mobile chip, while the NVIDIA GPU is a discrete-class console processor. The recorded data shows no benchmark wins for either side, meaning the database has not yet produced a direct competitive scoring event between them. Both parts sit at the 50th percentile against all GPUs in the database, which places them at the midpoint of the recorded performance distribution.

The use-case split is defined by their physical and architectural roles. The Intel Arc Graphics 1 Xe Mobile is an IGP with no slot width, no power connectors, and a 25 W TDP. It is designed for a portable device where the display output is dependent on the host platform. The NVIDIA Switch 2 GPU, by contrast, is a 40 W console GPU with a 12 GB dedicated memory pool and a 128-bit memory bus. It carries no display outputs of its own, which indicates it is meant to drive a fixed internal display through a dedicated console design.

In terms of raw throughput, the NVIDIA part wins decisively on every compute metric. The Intel GPU delivers 588.8 GFLOPS of FP32 performance, while the NVIDIA GPU delivers 4.301 TFLOPS, a gap of roughly 7.3 times. The texture rate tells a similar story: 18.40 GTexel/s for Intel versus 67.20 GTexel/s for NVIDIA, a 3.65 times advantage. Pixel rate also favors NVIDIA at 22.40 GPixel/s against 9.200 GPixel/s. These are not close contests. The NVIDIA GPU is built for sustained console workloads with a fixed thermal envelope, while the Intel IGP is a low-power companion solution for portable computing.

The architectural split is equally clear. Intel uses the Xe3-LPG architecture on a 3 nm process with 128 shading units, 8 TMUs, and 4 ROPs. NVIDIA uses the Ampere architecture on an 8 nm process with 1536 shading units, 48 TMUs, and 16 ROPs. The NVIDIA part also fields 12 RT cores and 48 tensor cores, while the Intel part has a single RT core and no tensor core field. The Intel GPU relies on system shared memory with system dependent bandwidth, while the NVIDIA GPU has its own 12 GB of LPDDR5X memory at 102.4 GB/s. That memory arrangement alone changes what each part can realistically do.

Architecture Differences

The two GPUs come from different manufacturers, different foundries, and different design philosophies. Intel builds the Arc Graphics 1 Xe Mobile on a 3 nm process at Intel's own foundry, using the Xe3-LPG architecture. The chip is called Wildcat Lake, and it belongs to the Arc Graphics-M (Wildcat Lake) generation. NVIDIA builds the Switch 2 GPU on an 8 nm process at Samsung, using the Ampere architecture with the GA10B chip, and it belongs to the Console GPU (Nintendo) generation. The process node difference is substantial: 3 nm versus 8 nm. That alone explains much of the efficiency and density gap between the two.

The die size is only recorded for the NVIDIA part at 200 mm². The Intel die size is unknown in the database. Transistor counts are unknown for both. The NVIDIA GPU has a 12 GB LPDDR5X memory pool on a 128-bit bus, delivering 102.4 GB/s of bandwidth. The Intel GPU uses system shared memory with a system dependent bandwidth figure, meaning its memory performance is tied to whatever platform it is installed in. That is a fundamental architectural difference: one part owns its memory, the other borrows it.

Compute resources differ by an order of magnitude. Intel provides 128 shading units, 8 TMUs, and 4 ROPs. NVIDIA provides 1536 shading units, 48 TMUs, and 16 ROPs. The NVIDIA GPU has 12 RT cores and 48 tensor cores, while the Intel GPU has only 1 RT core and no tensor cores listed. The FP16 to FP32 ratio is 2:1 on both parts, but the absolute numbers are far apart: Intel reaches 1,177.6 GFLOPS FP16, while NVIDIA reaches 8.602 TFLOPS FP16. The NVIDIA part is clearly designed for accelerated workloads beyond simple rasterization, given the tensor core count.

Clock behavior also differs. Intel runs at a 300 MHz base and a 2300 MHz boost. NVIDIA runs at a 561 MHz base and a 1400 MHz boost. The Intel part boosts much higher, but it has far fewer execution units. The NVIDIA part runs lower clocks but compensates with massive parallelism. The result is that NVIDIA achieves higher throughput despite a lower boost clock. The Intel part has a 25 W TDP, the NVIDIA part a 40 W TDP. Both are active production parts, but their release dates differ: NVIDIA launched on 2025-06-04, Intel on 2026-04-15.

Head-to-Head Benchmarks

The database holds no direct head-to-head benchmark entries between these two GPUs. There are no recorded wins for either side, and no average benchmark scores are listed. The percentileVsAllGpus field shows both at 50, which is a neutral placement in the overall distribution. That means the analysis must rely on the recorded specification data and the compute rate figures rather than measured game or synthetic scores.

The largest recorded advantage for NVIDIA is in FP32 throughput. The Intel part delivers 588.8 GFLOPS, while the NVIDIA part delivers 4.301 TFLOPS. That is a 7.3 times difference in raw single-precision compute. For FP16, the NVIDIA part reaches 8.602 TFLOPS against Intel's 1,177.6 GFLOPS, a 7.3 times gap as well, since both parts use a 2:1 FP16 to FP32 ratio. Texture rate shows NVIDIA at 67.20 GTexel/s versus Intel at 18.40 GTexel/s, a 3.65 times advantage. Pixel rate shows NVIDIA at 22.40 GPixel/s versus Intel at 9.200 GPixel/s, a 2.43 times advantage.

Memory bandwidth is another decisive separation. NVIDIA has a fixed 102.4 GB/s from its 12 GB LPDDR5X pool on a 128-bit bus. Intel has system shared memory with a system dependent bandwidth. In a portable device, the Intel part is limited by the host memory controller and the shared memory architecture, which typically introduces latency and bandwidth contention. The NVIDIA part avoids that entirely with dedicated memory.

The shading unit count difference is the most striking specification gap: 1536 versus 128, a 12 times difference. The TMU count is 48 versus 8, a 6 times difference. The ROP count is 16 versus 4, a 4 times difference. The RT core count is 12 versus 1, and the tensor core count is 48 versus zero. These ratios indicate that NVIDIA is not just faster, it is structurally designed for a different workload class. The Intel part is a minimal IGP, the NVIDIA part is a full console GPU.

FAQ

Q: Which GPU has more shading units?

A: The NVIDIA Switch 2 GPU has 1536 shading units, while the Intel Arc Graphics 1 Xe Mobile has 128 shading units.

Q: What memory configuration does each GPU use?

A: The NVIDIA Switch 2 GPU uses 12 GB of LPDDR5X memory on a 128-bit bus with 102.4 GB/s bandwidth. The Intel Arc Graphics 1 Xe Mobile uses system shared memory with system dependent bandwidth.

Q: How do the FP32 performance figures compare?

A: The NVIDIA Switch 2 GPU delivers 4.301 TFLOPS of FP32 performance, while the Intel Arc Graphics 1 Xe Mobile delivers 588.8 GFLOPS. The NVIDIA part is approximately 7.3 times faster in FP32.

Q: What are the TDP ratings for both GPUs?

A: The Intel Arc Graphics 1 Xe Mobile has a 25 W TDP, and the NVIDIA Switch 2 GPU has a 40 W TDP.

Q: Do both GPUs support the same graphics APIs?

A: Yes, both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Q: What process nodes are used for each GPU?

A: The Intel Arc Graphics 1 Xe Mobile is built on a 3 nm process at Intel, while the NVIDIA Switch 2 GPU is built on an 8 nm process at Samsung.

Specification Differences

The two parts differ across nearly every hardware field in the database. The Intel Arc Graphics 1 Xe Mobile uses the Xe3-LPG architecture on a 3 nm process, while the NVIDIA Switch 2 GPU uses the Ampere architecture on an 8 nm process. The Intel chip is called Wildcat Lake, the NVIDIA chip is called GA10B. The NVIDIA die size is 200 mm², the Intel die size is unknown.

Clock speeds differ: Intel runs at 300 MHz base and 2300 MHz boost, NVIDIA runs at 561 MHz base and 1400 MHz boost. Memory differs completely: Intel uses system shared memory with system dependent bandwidth, NVIDIA uses 12 GB of LPDDR5X with a 128-bit bus and 102.4 GB/s bandwidth. The compute unit counts are 128 shading units, 8 TMUs, and 4 ROPs for Intel, versus 1536 shading units, 48 TMUs, and 16 ROPs for NVIDIA. RT cores are 1 versus 12, and tensor cores are not listed for Intel while NVIDIA has 48.

Pixel rate is 9.200 GPixel/s for Intel and 22.40 GPixel/s for NVIDIA. Texture rate is 18.40 GTexel/s for Intel and 67.20 GTexel/s for NVIDIA. FP32 is 588.8 GFLOPS for Intel and 4.301 TFLOPS for NVIDIA. FP16 is 1,177.6 GFLOPS for Intel and 8.602 TFLOPS for NVIDIA. TDP is 25 W for Intel and 40 W for NVIDIA. Intel is an IGP with no slot width and no power connectors, while NVIDIA has no slot width recorded and no power connector data. Intel's display outputs are portable device dependent, NVIDIA has no outputs. Intel's bus interface is IGP, NVIDIA has no bus interface recorded. The release dates are 2026-04-15 for Intel and 2025-06-04 for NVIDIA. The NVIDIA part has a launch MSRP of 449 USD. The Intel part has no recorded launch MSRP. Intel lists HD Graphics-M as its predecessor, while NVIDIA has no predecessor recorded.

The Verdict

The data points to a clear separation of roles. The NVIDIA Switch 2 GPU is the far more capable compute part. It holds a 7.3 times advantage in FP32, a 3.65 times advantage in texture rate, and a 2.43 times advantage in pixel rate. It has 12 times the shading units, 6 times the TMUs, and 4 times the ROPs. It has its own 12 GB memory pool with 102.4 GB/s of bandwidth, while the Intel part relies on system shared memory. The NVIDIA part also includes 12 RT cores and 48 tensor cores, which the Intel part cannot match, since it has only 1 RT core and no tensor core field.

The Intel Arc Graphics 1 Xe Mobile is the lower-power option at 25 W versus 40 W. It uses a smaller 3 nm process and boosts to a higher 2300 MHz clock. It is designed as an integrated solution for a portable device, with display output dependent on the host. The NVIDIA part is a console GPU, built into a fixed system with no display outputs of its own. The 272 mm length, 116 mm height, and 14 mm width dimensions recorded for the NVIDIA part confirm it is a physical console component, not a drop-in mobile IGP.

The choice between the two is not a matter of preference, it is a matter of role. A portable computing device that needs a low-power integrated GPU with system shared memory would use the Intel part. A console system that needs sustained throughput, dedicated memory, and hardware ray tracing and tensor acceleration would use the NVIDIA part. The benchmark data is empty, so there are no measured game scores to weigh. The specification data alone, however, is decisive. The NVIDIA Switch 2 GPU is the stronger compute engine by every recorded metric. The Intel Arc Graphics 1 Xe Mobile is the more efficient, less demanding part for a different class of device.