VoltgridNews

MSI XpertStation WS300 Review: $99,999 GB300 Tower Idles Above 250 W

Two measured reviews of MSI's GB300 tower, StorageReview's in August and ServeTheHome's in October, find a solid build that idles above 250 W, hits 46-49 dBA at full load and lists from $99,999 at Newegg.

Local AI Desk

8 min read

Follows NVIDIA DGX Station Review: 748 GB, 1.6 kW and a $160,195 Dell List Price

MSI XpertStation WS300 GB300 tower shown from the rear and front beside the NVIDIA baseboard it is built around

MSI's XpertStation WS300 tower, rear and front, beside the NVIDIA GB300 baseboard (MSI).

Image: MSI

Source · Press image from the MSI newsroom, used for news reporting.

MSI's XpertStation WS300, a GB300 deskside tower, never idled below 250 W and reached 46-49 dBA at full load in ServeTheHome's review of Oct. 23, 2026, and Newegg US listed it from $99,999.00 on Nov. 15.

Those are the two numbers a buyer of this machine cannot get from the platform alone. StorageReview published the first measured review of the WS300 on Aug. 24, 2026, testing LLM serving with vLLM; ServeTheHome followed two months later with power and noise readings from its own unit. Both outlets describe MSI's build as solid. This review covers what MSI adds to NVIDIA's design and what it leaves out. The GB300 platform itself, its memory pool and its serving speeds across every partner build, is covered in our NVIDIA DGX Station review, and its case against multi-card RTX PRO 6000 towers in our DGX Station alternatives comparison.

What's new

Every DGX Station partner starts from the same NVIDIA baseboard. StorageReview puts it plainly: NVIDIA supplies the Grace Blackwell Ultra board and the software, while MSI supplies the chassis, the liquid cooling, the power delivery, the storage layout, the external ports and the service model. MSI's product page lists the result as one Blackwell Ultra GPU with up to 252 GB of HBM3e at 7.1 TB/s, a 72-core Grace CPU with up to 496 GB of LPDDR5X at up to 396 GB/s, and 748 GB of coherent memory across the two. High-bandwidth memory (HBM) sits beside the GPU; the slower LPDDR5X pool is reached over NVIDIA's NVLink-C2C chip-to-chip link.

The cooling is MSI's own. StorageReview describes a liquid loop with blocks on the GPU, the Grace CPU, the SOCAMM memory modules and the ConnectX-8 network chip, including its optical cages, feeding two 360 mm radiators, with a separate 120 mm fan for the rest of the case. MSI rates the loop for up to 1,400 W across CPU and GPU, and Newegg's spec table repeats that figure. In StorageReview's testing the GPU peaked at 71C and the CPU at 65C while the GPU drew 1,292 W.

MSI also prepared the case for a graphics card. StorageReview found a pre-wired 12VHPWR lead at the front of the chassis and an anti-sag bracket built into the vertical brace for long RTX PRO cards. That matters because the GB300 has no display output of its own: the only video out of the box is the management controller's Mini DisplayPort, which StorageReview lists at a maximum of 1024 x 768.

The case layout is the other change. ServeTheHome reports that MSI left the whole front open to airflow behind a mesh screen, unlike the closed front of Gigabyte's DGX Station build, and moved the buttons and ports to the top: two 10Gbps USB-C, two 10Gbps USB-A and audio jacks. MSI labels its USB ports with the USB-IF logo, which ServeTheHome counts in its favour, though it would have liked port speeds printed too. The tower measures 245.7 x 529.0 x 570.4 mm on MSI's page, and ServeTheHome puts its weight at nearly 30 kg.

Storage is set at the factory. MSI fills the two CPU-attached M.2 slots with a pair of 2 TB SSDs in RAID 1 for the operating system, StorageReview says, and leaves the two slots behind the ConnectX-8 free. Newegg's listing describes the same split and ships the system with Ubuntu 24.04 LTS and NVIDIA's developer tools.

What it solves

The WS300 solves the same problem as every GB300 tower: a team can load models far beyond a single workstation card on one machine beside the desk. StorageReview served GLM-5.2 in NVIDIA's NVFP4 format, a 433 GB checkpoint with 216 GB of expert weights pushed out to the LPDDR5X pool, at 36 output tokens per second for one stream and 139 at 32 concurrent streams, using vLLM with MTP speculative decoding at a forced acceptance length of 3, so those are best-case figures. A 96 GB RTX PRO 6000 cannot hold that model at all.

What MSI's build adds to that is a system that keeps running when power delivery goes wrong. StorageReview saw the platform's hardware power brake in action when loose power connectors caused the WS300 to lower its power limit instead of shutting down; it stayed online at reduced power until the connection was fixed. ServeTheHome, which had to recover its unit over the serial console and rebuild the operating system and AI stack after a previous user, says the documented process simply worked, and that the WS300 performed flawlessly both alone and networked with other DGX Stations.

Pros and Cons

The strengths rest on the reviewers' measurements and MSI's specifications. StorageReview measured 9,294 output tokens per second on gpt-oss-120b in MXFP4 with vLLM, a 512-token prompt and 512-token reply, at 128 concurrent streams, 2.7 times the 3,384 it measured on NVIDIA's 600 W RTX PRO 6000 workstation card in a Dell Pro Max Tower T2 on the same workload. MSI's loop is rated for 1,400 W and held the GPU at 71C at 1,292 W in StorageReview's run. The 12VHPWR lead and the card bracket are StorageReview's findings, the open front and top ports ServeTheHome's.

The weaknesses come from the same reviews. ServeTheHome measured idle power that never fell below 250 W and was usually well above it: about 148 W for the GPU and about 78 W for the Grace module. Its noise readings were 37-40 dBA at lower loads and 46-49 dBA at full load; it did not state the microphone distance. There are no USB4 ports, because NVIDIA's board provides none. The 1,600 W supply needs a C19 cord and, in practice, its own 20 A circuit in North America, and StorageReview notes that on 100-114 Vac input it delivers at most 1,300 W. MSI has published no list price.

Who it's for

The first buyer is a development team that needs Blackwell Ultra on hand. StorageReview names kernel developers writing CUDA or Triton code for that GPU, inference-engine and framework teams, and robotics groups, and it carved the GPU into three MIG instances to run NVIDIA Dynamo components side by side. For them the 252 GB of HBM3e is the point: models that fit in it, with room left for context, stay in fast memory.

The second is a group that wants one shared, always-on inference node. ServeTheHome argues the ConnectX-8 card makes the WS300 more of a server than a workstation, best run around the clock and reached over the network. It measured 784.3 Gbps of GPU-to-GPU RDMA traffic between two DGX Stations over two 400G links, and it advises buyers of more than one to plan 400 Gbps cabling to each office.

The third is a lab that must run models of 400 GB and more locally and privately. GLM-5.2 at 433 GB loaded and served on StorageReview's unit, at 139 output tokens per second at 32 streams in the best case.

It is not for a single developer whose models fit in 96 GB. StorageReview says an RTX PRO 6000 workstation costs far less for that work. That card was listed at $16,000, out of stock, on NVIDIA's US marketplace on Aug. 10, and our RTX PRO 6000 review covers it. It is also a poor fit for an office without a spare 20 A circuit, or for a desk where a 46-49 dBA machine would sit at arm's length.

Why two GB300 towers measure the same

MSI and ASUS do not design the GB300 board: NVIDIA does, and both partners fit it with the same 252 GB of HBM3e at 7.1 TB/s and 496 GB of LPDDR5X at 396 GB/s. Two towers built on one board share the same GPU, the same memory and the same power limit, so in a serving test like StorageReview's they land within a fraction of a percent of each other.

What the partner controls is everything around the board: the cooling loop, the power supply, the ports and the case. Those show up in noise, idle power, weight and price, not in token counts. That is why ServeTheHome's 250 W idle and 46-49 dBA readings, and the $99,999 to $160,195 spread of US listings, are the figures that separate one GB300 tower from another.

How it compares

Because MSI and ASUS build on the same NVIDIA board, StorageReview's two gpt-oss-120b results match. StorageReview's result of 9,294 output tokens per second on the WS300 sits beside 9,280 on the ASUS ExpertCenter Pro ET900N G3, on the same vLLM 512/512 workload at 128 streams; on 8,192-token prompts the two scored 5,038 and 5,023. That is a gap of 14 tokens per second, under 0.2%, so measured speed gives no reason to pick one brand over the other. StorageReview tested the WS300 remotely on an MSI-hosted system and the ASUS unit in its own lab, both without an add-in GPU, and did not give the vLLM version.

On StorageReview's same workload the MSI WS300 and ASUS ET900N G3 are level at 9,294 and 9,280 output tok/s, with the RTX PRO 6000 at 3,384.

Graphic: Voltgrid News

Price does separate them. Newegg's $99,999.00 "Start From" figure, read on Nov. 15, is a request-a-quote listing sold and shipped by Newegg with two 2 TB boot SSDs, not an MSI list price. Against it, Central Computers listed the ASUS build at $111,999.99 on Nov. 8, as Voltgrid reported on the ASUS price; TechRadar reported Supermicro's store price at $92,887.47 on Sep. 1, which could not be re-read since; Dell's Pro Precision with GB300 listed at $160,195.00 on Dell.com on Nov. 8 with an RTX PRO card and more storage; and HP's ZGX Fury listed at $132,000.00 on HP's store on Nov. 11, as Voltgrid reported on the HP price. These configurations differ in storage and add-in cards, so they are not one price series, but MSI's listed starting point is among the lowest in the GB300 group.

Power and noise are harder to compare. MSI, ASUS and Supermicro each list a 1,600 W supply, in different efficiency classes, and ServeTheHome also measured noise on the Supermicro Super AI Station, as Voltgrid reported. Readings from different units and test setups are best read on their own. One spec differs between sources: MSI's product page, StorageReview and ServeTheHome list the WS300's supply as 80 PLUS Platinum, while Newegg's spec table says Titanium; Platinum is the manufacturer's figure. The table below this review sets the WS300 against its rivals on memory, bandwidth, measured speed and price.

What to watch

NVIDIA said on Oct. 21 that DGX Station for Windows would come from partners including MSI in December 2026, and MSI's product page lists a Windows version of the WS300 as coming soon. MSI has still published no list price, so Newegg's quote-only figure remains the only dated US price for this build. StorageReview did not state the vLLM version behind its results, and wall power under full LLM load has not been published by either outlet; a review that measures it at the socket would settle the WS300's real running cost.

How it compares

Full table
MSI XpertStation WS300 Review: $99,999 GB300 Tower Idles Above 250 W: how the machine compares with other local AI hardware
ProductMemoryBandwidthAI computeMeasured speedPrice
ASUS ExpertCenter Pro ET900N G3ASUS (NVIDIA DGX Station GB300 platform)748 GB coherent: 252 GB HBM3e (GPU) + 496 GB LPDDR5X (CPU)[1]7.1 TB/s (GPU HBM3e); 396 GB/s (CPU LPDDR5X)[1]up to 20 PFLOPS AI performance (ASUS gives no precision or sparsity label) (sparse)[2]GPT-OSS-120B MXFP4, vLLM, 512/512, 1-128 streams: 9,280 aggregate output tok/s at 128 streams; 5,023 on 8,192/1,024 (StorageReview, ASUS-supplied unit, no add-in GPU)[3]$111,999.99 at Central Computers, US (ET900N G3-2SSD-V2, read Nov 8, 2026); no ASUS list price[4]
HP ZGX Fury AI StationHP252 GB HBM3e GPU + 496 GB LPDDR5x CPU (4x 128 GB SOCAMM); 748 GB unified[5]7.1 TB/s (GPU HBM3e); 396 GB/s (CPU LPDDR5x)[5]up to 20 petaFLOPS FP4 (HP gives no sparsity label) (sparse)[5]Two-unit cluster with an ASUS ET900N G3 over two 400G DACs: GLM-5.3 NVFP4 5,018 output tok/s at 256 streams, 512/512, vLLM 0.27.1 TP2 with MTP (StorageReview); no single-unit HP result[6]$132,000.00 on HP Store US (ZGX Fury G1n, 3-yr warranty, no graphics card, 4x 2 TB SED SSDs; Add to cart; read Nov 11, 2026)[7]
MSI XpertStation WS300MSI (NVIDIA DGX Station platform)This story252 GB HBM3e GPU + 496 GB LPDDR5X CPU (748 GB coherent)[8]7.1 TB/s (GPU HBM3e); 396 GB/s CPU LPDDR5X[8]—StorageReview vLLM: gpt-oss-120b MXFP4 9,294 output tok/s at 128 streams (512/512), 5,038 on 8,192/1,024; MiniMax M3 NVFP4 1,041 at 32 streams with EAGLE3-GQA at a forced acceptance length of 3 (best-case)[8]$99,999.00 'Start From' at Newegg US (model WS300T60L, 2x 2 TB NVMe RAID1), request-a-quote listing sold by Newegg (read Nov 15, 2026)[9]
NVIDIA DGX StationNVIDIA (systems built by ASUS, Dell, GIGABYTE, HP, MSI, Supermicro and Exxact)748 GB coherent: 252 GB HBM3e (GPU) + 496 GB LPDDR5X (CPU)[10]GPU HBM3e 7.1 TB/s; CPU LPDDR5X 396 GB/s; NVLink-C2C 900 GB/s between them[10]15 PFLOPS FP4 (without sparsity)[10]gpt-oss-120b MXFP4, vLLM, 512/512 chat workload: 9,280 output tok/s at 128 streams (MSI WS300 9,294), single GB300 (StorageReview, ASUS ExpertCenter Pro ET900N G3)[3]No NVIDIA list price; lowest US partner listing read Nov 4, 2026: Exxact Valence VWS-158270643 from $98,450 (Tom's Hardware found $94,930 in Aug); Dell Pro Precision with GB300 lists $160,195[11]
Supermicro Super AI Station (ARS-511GD-NB-LCC)Supermicro (NVIDIA DGX Station platform)748 GB total coherent: 252 GB HBM3e GPU + 496 GB LPDDR5X system[12]7.1 TB/s (GPU HBM3E); 396 GB/s (CPU LPDDR5X)[13]up to 20 PFLOPS FP4 (Supermicro gives no sparsity label; ServeTheHome describes it as sparse FP4)[12]Qwen3.8 Flash Next NVFP4, vLLM, tool-loop 4,096-in/512-out: several scenarios exceeded 3,000 output tok/s at 32 concurrent users, still rising (ServeTheHome; RTX PRO 2000 installed)[14]$92,887.47 before tax and shipping on Supermicro's online store (ARS-511GD-NB-LCC-01-G2); about $90,101.05 with a 3% new-customer discount (TechRadar, Sep 1, 2026)[15]
NVIDIA RTX PRO 6000 Blackwell Workstation EditionNVIDIA (sold through PNY, TD SYNNEX and workstation makers)96 GB[16]1,792 GB/s[16]4,000 AI TOPS (FP4, with sparsity) (sparse)[16]gpt-oss-120b MXFP4, stock vLLM, 512/512 chat workload: 3,064 output tok/s at 128 streams (1,169 on 8,192/1,024 prompts); Server Edition at 600W (StorageReview)[17]$16,000 on NVIDIA's US marketplace (Workstation Edition), listed out of stock[18]
Apple Mac Studio (M5 Ultra)Appleup to 512 GB unified (96 GB on 30-core M5 Ultra; 256 GB or 512 GB on 36-core/80-core M5 Ultra)[19]1.2 TB/s[19]—GLM-5.3-Flash MLX 4-bit with MTP on TensorFold, ISL 2048/OSL 512: 94.0 tok/s at 0 context, 96.5 at 128K, 79.4 at 512K; TTFT 1.1 s at 0, 66.3 s at 128K (Tom's Hardware, 256 GB M5 Ultra)[20]$22,299 minimum with 512 GB ($6,799 36-core CPU/80-core GPU M5 Ultra config plus $15,500 memory upgrade; Oct 28, 2026)[21]
Corsair FlexPrime V20C Workstation (CSW-9040004-NA)Corsair32 GB GDDR6 on the Arc Pro B70 GPU + 32 GB (2x 16 GB) DDR5 system memory[22]608 GB/s (GPU GDDR6)[22]—Qwen3.8 27B Q4_K_M, llama.cpp server: 19.8 tok/s at 1 request, 33.5 at 4, 35.4 at 8; TTFT 9,005 ms on 8,192-token prompt (StorageReview)[23]$3,999.99 on Corsair US store (32 GB DDR5, 1 TB SSD; read Nov 15, 2026)[22]
Show the other 32 machines
MSI XpertStation WS300 Review: $99,999 GB300 Tower Idles Above 250 W: how the machine compares with other local AI hardware: the other machines
ProductMemoryBandwidthAI computeMeasured speedPrice
NVIDIA RTX Spark (N1X) PCsNVIDIA (MediaTek collaborated on the custom CPU design)Up to 128 GB LPDDR5X (20-core CPU / 6,144-core GPU version); up to 64 GB (18-core / 5,120-core laptop version)[24]Up to 300 GB/s: LPDDR5X-9400 on a 256-bit bus (NVIDIA's product page lists no bandwidth)[25]—gpt-oss:20b via Ollama in a Codex CLI coding agent, 64K context: 58.5 tok/s generation, 3 of 4 tasks passed (Dell XPS 16, 20-core N1X, 64 GB; StorageReview)[26]Laptops from $2,599; pricing and configurations vary by region[27]
AMD Ryzen AI Halo (Ryzen AI Max+ PRO 495, 192 GB)AMD192 GB LPDDR5X (256-bit)[28]273 GB/s[28]——$7,299.99 for the Linux version (released the week of Oct 12, 2026); Windows version coming soon[28]
ASUS Ascent GX10ASUS64 GB or 128 GB LPDDR5X-8533 unified[29]273 GB/s (8533 MT/s), as listed by StorageReview; ASUS's spec page states no bandwidth[30]Up to 1 PFLOP FP4 (GB10; NVIDIA's footnote: theoretical, using the sparsity feature). ASUS's pages say 1 petaFLOP / FP4 with no sparsity label (sparse)[31]gpt-oss-120b, vLLM online serving, equal ISL/OSL: 70 tok/s at batch 1 rising to 680 tok/s aggregate at batch 64 (StorageReview)[30]$5,999 (128 GB, 1 TB, GX10-GG0010BN, ASUS eShop US, Oct 30, 2026)[32]
Dell Pro Precision with GB10 (model FCM1253; launched as Dell Pro Max with GB10)Dell128 GB LPDDR5x up to 8533 MT/s onboard (64 GB option also listed)[33]273 GB/s (128 GB LPDDR5x, 256-bit)[34]—gpt-oss-120b, vLLM online serving, equal ISL/OSL: 68.57 tok/s at batch 1 rising to 670.83 tok/s aggregate at batch 64 (StorageReview)[35]$7,901.00 (default FCM1253, 128 GB, 2 TB Gen4 QLC); 4 TB TLC SED SSD option +$2,450.00; 64 GB option listed at a $2,150.00 difference (Dell.com US, read Nov 10, 2026)[36]
Framework Desktop (Ryzen AI Max+ PRO 495, 192 GB)Framework192 GB LPDDR5X-8533 soldered[37]273 GB/s (256-bit)[37]——$6,799 (Max+ PRO 495 192 GB DIY Edition pre-order, Nov 3, 2026; storage not included)[38]
GIGABYTE AI TOP ATOMGIGABYTE128 GB unified system memory[39]273 GB/s (LPDDR5x-8533, 256-bit)[40]—gpt-oss-120b, vLLM online serving, equal ISL/OSL: 60.71 tok/s at batch 1 rising to 649.62 tok/s aggregate at batch 64 (StorageReview)[40]$7,499.99 (128 GB, 4 TB PCIe 4.0) and $9,405.63 (128 GB, 4 TB PCIe 5.0; more options from $7,999.99); 128 GB from $5,999.99 (1 TB) and $6,999.99 (2 TB) at Newegg US, read Nov 9, 2026[41]
GMKtec EVO-X5 Pro AI Mini PCGMKtec192 GB LPDDR5X-8533 onboard (8 x 24 GB, non-upgradeable); up to 160 GB allocatable as VRAM[42]Up to 273 GB/s[43]—GMKtec Laboratory table: GLM-5.3-Flash UD-IQ4_XS (146.05 GB) decode 17.79 tok/s, prefill 70.18, TTFT 14.71 s; Qwen3.8-Flash-Next ROCmFP4-FAST (93.48 GB) decode 45.08 tok/s[42]$6,799 (192 GB/2 TB) and $7,099 (192 GB/4 TB), all variants available, GMKtec US store, read Nov 9, 2026[42]
HP ZGX Nano AI Station (G1n)HP128 GB coherent unified system memory[44]273 GB/s (128 GB LPDDR5x)[45]GB10: up to 1 PFLOP FP4, theoretical, using the sparsity feature; HP and reviewers list 1,000 TOPS FP4 (sparse)[31]gpt-oss-120b, vLLM online serving, equal ISL/OSL: 69.6 tok/s at batch 1 rising to 722.9 tok/s aggregate at batch 64 (StorageReview)[46]$7,399.00 at the HP Store for the 4 TB model; $4,759 on Amazon (Notebookcheck, review updated Oct 5, 2026)[47]
Lenovo ThinkStation PGXLenovo128 GB LPDDR5X unified (256-bit)[48]273 GB/s[48]——US from $5,079 (1 TB base model) on Lenovo's site, per TechRadar's review (Jun 9, 2026); UK from GBP 5,200[49]
MSI EdgeMesa N AI+ (128 GB, X-005US)MSI128 GB LPDDR5X unified system memory[50]273 GB/s[50]——$7,999.00 (128 GB, 4 TB, X-005US, MSI US store, read Nov 11, 2026)[50]
MSI EdgeXpert (MS-C931)MSI64 GB or 128 GB LPDDR5x unified (256-bit)[51]273 GB/s[51]1 PFLOPS FP4 (sparse)[51]—No list price; MSI sells via quote (Get a quote)[51]
NVIDIA DGX SparkNVIDIA (partner GB10 systems from Acer, ASUS, Dell, GIGABYTE, HP, Lenovo and MSI)128 GB; a 64 GB configuration is sold only through OEM partners[31]273 GB/s[31]Up to 1 PFLOP FP4 (theoretical, using the sparsity feature) (sparse)[31]Qwen 3.8 27B (dense) with multi-token prediction: decode of about 20 tokens/s, with time-to-first-token staying interactive out to the full 262K context.[52]$6,950.00 (128 GB, 4 TB) on NVIDIA's enterprise marketplace with an Add To Cart button in a copy cached 07:59 UTC on 2026-10-31; stock unverified (listed out of stock on 2026-10-21); 64 GB partner systems from $4,999[53]
NVIDIA DGX Spark 64 GBNVIDIA (sold by Acer, ASUS, Dell, Gigabyte, HP and MSI)64 GB LPDDR5X unified (two units pool 128 GB over ConnectX-7)[54]273 GB/s (256-bit; NVIDIA spec table covers both 64 GB and 128 GB)[31]1 PFLOPS FP4 (sparse; NVIDIA spec table covers both memory configurations)[31]—from $4,999 (partner systems, Oct 2026; no Founders Edition)[54]
Sipeed MaixStationSipeed128 GB LPDDR5X (256-bit)[28]273 GB/s (Sipeed claim)[28]256 TOPS FP4 (native FP4, no sparsity; Axera/Sipeed claim)[28]—about $3,999 expected for 128 GB, with Kickstarter launch discounts (Sipeed to Tom's Hardware, Oct 2026)[28]
AMD Ryzen AI Halo Developer PlatformAMD128 GB LPDDR5x-8000 unified[55]256 GB/s[55]—gpt-oss-120b, vLLM online serving, equal ISL/OSL 256/256: 222 tok/s aggregate at batch 64 vs 701 on DGX Spark (StorageReview)[56]$3,999 with 2 TB SSD (Jul 2026)[56]
Framework Desktop (Ryzen AI Max+ 395)Framework128 GB LPDDR5x-8000 soldered (32 GB and 64 GB options)[57]256 GB/s[58]—gpt-oss-120b MXFP4 in LM Studio (out-of-the-box setup): roughly 45 tok/s (ServeTheHome; Ryzen AI Max+ 395, 128 GB, Windows 11 Pro)[59]$3,449 (DIY Edition, Max+ 395 with 128 GB, system only; out of stock, Nov 3, 2026)[60]
Acer Veriton GN100 AI Mini WorkstationAcer128 GB LPDDR5x coherent unified[61]——gpt-oss-120b, vLLM, total token throughput: prefill heavy 303.61 to 2,777.56 tok/s; decode heavy 38.38 to 292.41 tok/s (batch 1 to 64, StorageReview)[62]from $3,999 in North America (announced Sep 3, 2025; prices vary by region)[61]
ASRock Industrial AI BOX-A395ASRock Industrial128 GB LPDDR5X-8000 soldered[63]——AgentSTH V7, 32 threads: 15% ahead of the next-closest Ryzen AI Max+ 395 box (single-core within 1%); estimated SPEC CPU2026 int rate: top of a 44% spread across four 395 systems; measured by ServeTheHome, Ubuntu 26.04[64]—
Beelink GTR9 Pro (Ryzen AI Max+ 495, 128 GB)Beelink128 GB LPDDR5X-8533 unified (soldered)[65]—132 AI TOPS on Beelink's page; AMD lists the PRO 495 at Overall TOPS up to 132 and NPU TOPS up to 55; no precision stated by either[65]4.24 tok/s, LM Studio, Qwen3.8-35B-A3B (Beelink's own figure on its store page; quantization, context, backend and RAM size not stated)[65]$4,799 (128 GB/2 TB), Beelink store, pre-order (read Nov 13, 2026)[65]
Beelink GTR9 Pro (Ryzen AI Max+ 495, 192 GB)Beelink192 GB LPDDR5X-8533 unified (soldered)[65]—132 AI TOPS on Beelink's page; AMD lists the PRO 495 at Overall TOPS up to 132 and NPU TOPS up to 55; no precision stated by either[65]4.24 tok/s, LM Studio, Qwen3.8-35B-A3B (Beelink's own figure on its store page; quantization, context, backend and RAM size not stated)[65]$7,399 (192 GB/2 TB) or $7,799 (192 GB/4 TB), Beelink store, pre-order (read Nov 13, 2026)[65]
Bosgame M5 MAX AI Mini PCBosgame192 GB unified LPDDR5 (soldered); up to 160 GB allocatable to the GPU[66]———$6,199 introductory first-batch pre-order price (192 GB, 2 TB PCIe 4.0 SSD), from Nov 3, 2026, until the first allocation sells out[66]
Dell Pro Max Micro (FCM2250)Dell24 GB GDDR7 on the RTX PRO 4000 SFF (23.9 GB usable measured) + 96 GB DDR5-6400 CSODIMM system RAM in StorageReview's tested build[67]——Ministral 3 8B, llama.cpp server (CUDA), 4,096-in/1,024-out: 52.0 tok/s at 1 request, 108.9 at 4; TTFT 1,477 ms (StorageReview, RTX PRO 4000 SFF 24 GB, driver 595.99.02, Ubuntu 26.04.1)[67]From $1,449.00 on Dell.com US; RTX PRO 4000 SFF 24 GB option +$3,870.00 (read Nov 16, 2026)[68]
Dell Pro Precision 9 T4DellUp to 2x 96 GB GDDR7 (two RTX PRO 6000 Max-Q) + up to 4 TB DDR5 RDIMM in 16 slots, 8 channels[69]——MiniMax M2.7 Q4_K_M (129.3 GB) split across 2x RTX PRO 6000 Max-Q: 83.7 tok/s one user, 127.3 peak; 30.3 on one card plus system RAM (StorageReview, llama.cpp b11716, Ubuntu 26.04.1)[69]From $5,925.00 on Dell.com US; dual RTX PRO 6000 Max-Q option +$30,820.00 (read Nov 14, 2026). StorageReview's tested build (Xeon 698X, 256 GB, 2x Max-Q) about $66,600[70]
Dell Pro Precision with GB300 (model FCT6263; announced as Dell Pro Max with GB300)Dell252 GB HBM3e GPU + 496 GB LPDDR5X CPU[71]———$160,195.00 (Dell.com US default config FCT6263, read Nov 8, 2026)[71]
Gigabyte AI TOP Aeroup to 128 GB LPDDR5X unified[72]————
GMKtec EVO-X2 AI Mini PCGMKtec64 GB or 128 GB onboard LPDDR5X-8000 (non-upgradeable)[73]——Llama 3.3 70B Q6_K Instruct: 3.7-3.8 tok/s generation fully on the Radeon 8060S (about 2 tok/s CPU-only); Qwen3-32B 10.4-10.5 tok/s; 128 GB unit, short prompt, tool not named (ServeTheHome)[74]$3,499.99 (128 GB/1 TB) and $3,649.99 (128 GB/2 TB), each against a $3,999.99 compare-at price (GMKtec US store, read Nov 11, 2026)[73]
HP Z2 Mini G1a WorkstationHPup to 128 GB unified, up to 96 GB assignable to the GPU, up to 8000 MT/s[75]—up to 50 NPU TOPS[75]gpt-oss-120b GGUF (63.39 GB, full GPU offload) in LM Studio: close to 40 tok/s (StorageReview)[76]$4,781 list for StorageReview's review configuration; about $3,300 street (B&H top spec with 2 TB SSD at $3,342.65, Aug 2025)[76]
Microsoft Surface RTX Spark Dev BoxMicrosoft128 GB unified memory[77]—up to 1 PFLOP of AI compute (Microsoft; precision and sparsity not stated on the page) (sparse)[77]—From $5,999.99 (Microsoft Store US, pre-order, read Nov 1, 2026)[77]
MSI EdgeMesa N AI+ (64 GB)————from $4,699 in the US (64 GB soldered LPDDR5X, 1 TB PCIe Gen 5)[78]
Tenstorrent TT-QuietBox 2 (Blackhole)Tenstorrent128 GB GDDR6 on the accelerators + 256 GB DDR5-5600 system memory[79]———$9,999 (Oct 30, 2026)[79]
WiCi One (RTX 5060 Ti edition)WiCi16 GB GDDR7 VRAM; 1 TB NVMe 4 SSD for TurboStream weight paging (WiCi's claim: NVMe used as working VRAM). System RAM not stated[80]———$2,599 list (WiCi shop, Oct 2026); a $600-off code (WICIWELCOME) is applied at checkout. Early-access price was $1,999 (Notebookcheck, Engadget, Sep 2026)[80]
WiCi One (RTX 5090 edition)WiCi32 GB GDDR7 VRAM; 4 TB NVMe 5 SSD for TurboStream weight paging (WiCi's claim: NVMe used as working VRAM). System RAM not stated[80]———$9,999 list (WiCi shop, Oct 2026); a $600-off code (WICIWELCOME) is applied at checkout[80]

Italic values are reported by an outlet or measured by an independent lab; the rest come from the maker. Measured speeds use different models, quantizations and tools, so read each with its source rather than against the others. Voltgrid News does not test hardware itself.

Sources (80)
  1. [1] ASUS
  2. [2] ASUS
  3. [3] StorageReview
  4. [4] Central Computers
  5. [5] HP
  6. [6] StorageReview
  7. [7] HP
  8. [8] StorageReview
  9. [9] Newegg
  10. [10] NVIDIA
  11. [11] Exxact
  12. [12] Supermicro
  13. [13] ServeTheHome
  14. [14] ServeTheHome
  15. [15] TechRadar
  16. [16] NVIDIA
  17. [17] StorageReview
  18. [18] NVIDIA
  19. [19] Apple
  20. [20] Tom's Hardware
  21. [21] MacRumors
  22. [22] Corsair
  23. [23] StorageReview
  24. [24] NVIDIA
  25. [25] ServeTheHome
  26. [26] StorageReview
  27. [27] NVIDIA
  28. [28] Tom's Hardware
  29. [29] ASUS
  30. [30] StorageReview
  31. [31] NVIDIA
  32. [32] ASUS eShop
  33. [33] Dell
  34. [34] HotHardware
  35. [35] StorageReview
  36. [36] Dell
  37. [37] Framework Blog
  38. [38] Framework
  39. [39] GIGABYTE (PR Newswire)
  40. [40] StorageReview
  41. [41] Newegg
  42. [42] GMKtec
  43. [43] GMKtec
  44. [44] HP
  45. [45] HotHardware
  46. [46] StorageReview
  47. [47] Notebookcheck
  48. [48] Lenovo PSREF
  49. [49] TechRadar
  50. [50] MSI
  51. [51] MSI
  52. [52] Tom's Hardware
  53. [53] NVIDIA Marketplace
  54. [54] NVIDIA Blog
  55. [55] AMD
  56. [56] StorageReview
  57. [57] Framework
  58. [58] Framework Blog
  59. [59] ServeTheHome
  60. [60] Framework
  61. [61] Acer
  62. [62] StorageReview
  63. [63] ServeTheHome
  64. [64] ServeTheHome
  65. [65] Beelink
  66. [66] Notebookcheck
  67. [67] StorageReview
  68. [68] Dell
  69. [69] StorageReview
  70. [70] Dell
  71. [71] Dell
  72. [72] Notebookcheck
  73. [73] GMKtec
  74. [74] ServeTheHome
  75. [75] HP
  76. [76] StorageReview
  77. [77] Microsoft
  78. [78] Notebookcheck
  79. [79] Tenstorrent
  80. [80] WiCi

Sources

Published
Sources
4 cited