StorageReview

10/15/2026

The Microsoft Surface Laptop Ultra for Business on our bench, the first RTX Spark N1X laptop to get a full review from us. We ran it through local AI, CPU, gaming, storage, and thermal testing against four x64 notebooks, and it generated 89 tokens per second from a 35B-parameter coding agent that the discrete-GPU laptops could not load at all. One FP64 workload shut it down three times, and the review covers why. Full review at the link in bio.

10/14/2026

Four generations of NVIDIA inference cards on the bench: the L4, the new RTX PRO 4500 Blackwell Server Edition, the L40S, and the RTX PRO 6000. We ran all four through vLLM serving in a Dell PowerEdge R770 to see how close a 165W single-slot card gets to the 350W L40S. At short context it matched it, 1,107 vs 995 tokens/s. Full review at the link in bio. NVIDIA Dell

10/12/2026

Micron 9650 PRO SSDs in hand: PCIe Gen6 x4, E3.S 7.5mm, 7.68TB on these. Micron rates the 9650 at up to 28GB/s sequential reads, double what a Gen5 x4 drive can move. Capacities in the family run to 30.72TB.

Gen6 drives need next-gen servers, more on that soon from the lab.

Micron Technology

10/07/2026

HPE's new ProLiant DL585a Gen13, the first eight-GPU server on AMD EPYC Venice. It's a 10U air-cooled chassis built for the data centers enterprises already have, with up to 8TB of memory, 16 E3.S drives, and PCIe Gen6 throughout, one of four Gen13 launch systems. Full story at the link in bio. Hewlett Packard Enterprise

10/02/2026

10/02/2026

NVIDIA just dropped a smaller $4,999 DGX Spark. The 64GB version keeps the GB10 Grace Blackwell chip and full software stack, runs models up to 100B parameters on its own, and two of them link over a single cable to pool 128GB. Ships October 23 from Acer, ASUS, Dell, Gigabyte, HP, and MSI. Full story at the link in bio.

09/30/2026

Eight DGX Sparks on one shelf, all reading from the same Solidigm QLC array over 100GbE. We built shared storage for deskside AI so models and datasets live in the data center, and any Spark loads them with one mount line. Eight at once pulled 24.9GiB/s. Full build at the link in bio.
NVIDIA AI Infrastructure Solidigm

09/24/2026

The NVIDIA RTX PRO 4500 Blackwell Server Edition delivered 2.7x the L4's output throughput in the same HPE ProLiant DL145 Gen11 edge server, for about 105W more at the wall.

Latency is where the gap is widest. The 4500 generates at 12 to 15ms per token against 35 to 40ms on the L4, and at 256 concurrent requests it kept first-token latency under a second on 256-token prompts while the L4 queued for 9 to 39 seconds. With 32GB and native FP4, it also ran Gemma 4 26B-A4B in NVFP4, a model the L4 can't run natively.

Full review at the link in bio.

Hewlett Packard Enterprise NVIDIA AI Infrastructure

09/23/2026

ASUS ExpertCenter Pro ET900N G3: a GB300 DGX Station with 72 Grace cores, a B300 GPU, 252GB of HBM3e, and 496GB of LPDDR5X, cooled by a closed loop with copper cold plates on the Superchip, SOCAMM, and ConnectX-8, two radiators, and a 1,600W Titanium power supply. ASUS added a small fan over the 400G optics cages and two carry handles on top of the 27kg tower. Full review out now!

ASUS

09/18/2026

The ASUS Ascent QN10 is the first mini PC built on Qualcomm's Snapdragon X2 Elite: 18 Oryon cores, an 80 TOPS NPU, and three USB4 ports in a 0.7-liter box you can hide behind a monitor. Check the teardown, where a v***r chamber and copper thermal pads on the M.2 drives show how ASUS keeps it near silent under load. Why it matters: this is the first Arm mini PC we've tested that beats x86 outright, 43% to 76% ahead of the NUC 14 Pro on our CPU suites, with local Stable Diffusion and LLM inference running on the NPU while the CPU sits idle. The trade-off is 32GB of soldered memory and a $1,349.99 starting price against a $700 NUC barebones. Full review at the link in bio.

ASUS