GPUs & AI Accelerators — procurement report
ProductArena · rankings as of 2026-09-15 · evidence as of 2026-09-15 · 8 products · 43 judged requirements · 344 judged cells
Methodology: Every product is judged against a shared taxonomy of user stories using cited evidence — hands-on probes > repository code > independent community sources > vendor claims — never opinion. Full writeup: https://ultrametric.ai/productarena/methodology
Leaderboard
| # | Product | PA Score | Coverage score | Applicable cells | Confidence |
|---|---|---|---|---|---|
| 1 | AMD Instinct MI355X | 28.1 | 26.9 | 30/43 | C |
| 2 | Radeon RX 9070 XT | 27.8 | 23.2 | 27/43 | C |
| 3 | NVIDIA B200 | 17.9 | 16.9 | 28/43 | D |
| 4 | RTX PRO 6000 Blackwell | 15.4 | 16.7 | 27/43 | C |
| 5 | GeForce RTX 5070 Ti | 13.3 | 16.2 | 27/43 | C |
| 6 | NVIDIA H200 (SXM) | 11.7 | 19.3 | 28/43 | D |
| 7 | GeForce RTX 5080 | 11.7 | 15.5 | 28/43 | D |
| 8 | GeForce RTX 5090 | 6.9 | 19.2 | 28/43 | C |
PA Score = agent-readiness blend (see methodology). Coverage score = weighted share of judged requirements met. Confidence = how much of the score rests on tested vs claimed evidence (A–D).
Uncertainty note
The current #1/#2 gap in this arena is not close enough to qualify for the multi-judge uncertainty pass (or the pass has not covered it yet) — no extra caveat applies beyond the per-product confidence grades above.
Buyer checklist (RFP)
The arena's 43 judged user stories as requirements, grouped by theme. Priorities mirror the story weights our scoring uses (3 = must-have, 2 = should-have, 1 = nice-to-have). Interactive version with per-requirement verdicts for the top products: /arena/gpus/checklist
Agenticness — how well agents can access and operate the productAgenticness
How well agents can access and operate the product
- ai-native userPlug MCP servers into this product so it can use their toolsmust-have
- ai-native userConnect an agent via an official MCP servermust-have
- ai-native userDrive the product through a documented public APImust-have
- ai-native userDelegate tasks to a built-in AI assistant inside the productmust-have
- ai-native userPoint an agent at llms.txt or agent-oriented docsshould-have
- ai-native userRun the product headlessly / in CI for automationshould-have
- ai-native userUse an official CLIshould-have
- ai-native userIssue scoped/least-privilege API credentials for an agentshould-have
- ai-native userBuild against official SDKsshould-have
- ai-native userSubscribe to events via webhooksshould-have
- ai-native userGet AI-generated insights and suggestions from my data inside the productshould-have
- ai-native userSet up automations that run autonomously in the backgroundshould-have
- ai-native userOperate the product with natural-language commandsshould-have
- ai-native userExplore an interactive API reference with runnable examplesshould-have
- ai-native userDownload a machine-readable API spec (OpenAPI or equivalent)should-have
- ai-native userRely on versioned APIs with a documented deprecation policyshould-have
- ai-native userTest against a sandbox environment without touching production datanice-to-have
Ai compute — stories about ai compute in this arenaAi compute
Stories about ai compute in this arena
- ai-native userThis GPU has a documented LLM inference story — low-precision formats (FP8/FP4) and supported serving stacks (TensorRT-LLM, vLLM, ROCm, llama.cpp) for this partmust-have
- ml engineerSize training and inference from published tensor throughput — TFLOPS or TOPS with precision and sparsity stated, not a bare marketing numbermust-have
Automation depth — how much of the product can run unattendedAutomation depth
How much of the product can run unattended
- ai-native userDefine rules that trigger actions automatically on eventsmust-have
- ai-native userPerform bulk operations across many items at onceshould-have
- ai-native userSchedule recurring jobs or workflowsshould-have
- ai-native userVersion, review, and roll back my automationsnice-to-have
Creator media — stories about creator media in this arenaCreator media
Stories about creator media in this arena
- creatorHardware media engines and creator-app acceleration are documented — AV1/HEVC encoders, and professional or ISV-certified driver support where the vendor claims itshould-have
Datacenter scale — stories about datacenter scale in this arenaDatacenter scale
Stories about datacenter scale in this arena
- ml engineerTrain and serve at datacenter scale on this part — documented high-bandwidth interconnect (NVLink, Infinity Fabric), multi-GPU systems, and rack-scale deploymentmust-have
Driver openness — stories about driver openness in this arenaDriver openness
Stories about driver openness in this arena
- developerLinux is a first-class citizen for this GPU — documented Linux driver releases and independent Linux testing of this partshould-have
- developerRun this GPU on an open driver — open-source kernel modules or upstream Linux support documented by the vendorshould-have
Gaming performance — stories about gaming performance in this arenaGaming performance
Stories about gaming performance in this arena
- gamerThis card drives high-refresh 4K gaming — vendor performance claims corroborated by independent game benchmarksmust-have
- gamerAI upscaling and frame generation are supported on this card — the DLSS or FSR generation is documented for this part, with broad game supportshould-have
Memory vram — stories about memory vram in this arenaMemory vram
Stories about memory vram in this arena
- ai-native userRun a 70B-class quantized LLM on this GPU — published VRAM capacity and memory bandwidth that make local or single-node inference practicalmust-have
- ml engineerMemory specs are published in full for this exact part — capacity, memory type, bus width, and bandwidthshould-have
Openness — open source, data portability, and self-hosting storiesOpenness
Open source, data portability, and self-hosting stories
- ai-native userExport all of my data in open formats and leavemust-have
- ai-native userSelf-host the core productmust-have
- ai-native userDo everything through the API that I can do in the UIshould-have
- ai-native userRead the product's source under an open licenseshould-have
Power cooling — stories about power cooling in this arenaPower cooling
Stories about power cooling in this arena
- ml engineerSustained workloads are power-efficient on this part — documented power envelopes with independent performance-per-watt testingshould-have
- gamerSpec a build around published board power — TDP/TGP, connector requirements, and cooling guidance for this exact cardshould-have
Privacy posture — data-handling and privacy storiesPrivacy posture
Data-handling and privacy stories
- ai-native userPrevent my data from being used to train AI modelsmust-have
- ai-native userChoose where my data is stored (region/residency)should-have
- ai-native userControl data retention and deletionshould-have
- ai-native userOpt out of telemetry and usage trackingshould-have
Software toolchain — stories about software toolchain in this arenaSoftware toolchain
Stories about software toolchain in this arena
- developerShip GPU-compute workloads on the vendor's toolchain — CUDA or ROCm/HIP documentation lists this part as a supported targetmust-have
- ml engineerPyTorch and mainstream ML frameworks run on this GPU through officially documented builds and support matricesshould-have
Appendix: recorded probes
Hands-on probe recordings — transcripts/videos a human can replay, the strongest evidence tier. Watch them at https://ultrametric.ai/productarena/proofs
- AMD Instinct MI355X
curl -sL 'https://rocm.docs.amd.com/en/latest/' | grep -o 'ROCm' | head -1 # the ROCm toolchain docs, liveterminal · recorded 2026-09-15 · exit 0 - NVIDIA H200 (SXM)
curl -sL 'https://www.nvidia.com/en-us/data-center/h200/' | grep -o 'H200' | head -1 # vendor spec page, live and keylessterminal · recorded 2026-09-15 · exit 0 - GeForce RTX 5090
curl -sL 'https://www.nvidia.com/en-us/geforce/graphics-cards/50-series/rtx-5090/' | grep -o 'RTX 5090' | head -1 # vendor spec page, live and keylessterminal · recorded 2026-09-15 · exit 0 - RTX PRO 6000 Blackwell
curl -sL 'https://developer.nvidia.com/cuda-toolkit' | grep -o 'CUDA Toolkit' | head -1 # the toolchain docs the compute stories lean on, liveterminal · recorded 2026-09-15 · exit 0
Cite as: ProductArena by Ultrametric Inc, GPUs & AI Accelerators arena, rankings as of 2026-09-15 — https://ultrametric.ai/productarena/arena/gpus
License: © 2026 Ultrametric Inc. Brief quotation of individual verdicts, scores, or evidence excerpts is permitted with attribution to "ProductArena by Ultrametric Inc (ultrametric.ai/productarena)", as is use of the data to evaluate, contest, or contribute corrections. Bulk copying, redistribution, or use to build competing datasets requires prior written permission (see DATA-LICENSE in the repository).
No liability: rankings, verdicts, and scores are research outputs derived from the cited evidence at a point in time, provided "as is", without warranties. Ultrametric Inc accepts no responsibility for procurement, purchasing, or other decisions made in reliance on them — verify against the cited evidence before acting (https://ultrametric.ai/productarena/terms).