NVIDIA announces Rubin HPC GPU for science workloads, due in 2028
NVIDIA announced Rubin HPC, a new GPU in its Vera Rubin platform, at SC26 on 16 November 2026, with availability expected in 2028 and no price given.
Newsroom
Chronological coverage of new developments in AI and hardware.
NVIDIA announced Rubin HPC, a new GPU in its Vera Rubin platform, at SC26 on 16 November 2026, with availability expected in 2028 and no price given.
Three PCIe 4.0 x4 drives, two hosts, one dedicated suite. The 990 PRO is fastest at 4K reads and holds 1,450 MB/s once its cache is full. The GM7 has the cleanest latency tail. The P310 falls to 177 MB/s.
TechPowerUp and Tom's Hardware report that an outage knocked out Outlook and other Microsoft 365 services in Japan, with many offices reverting to fax machines and phone calls.
11:09
HP's 16-inch workstation offers up to 192GB of memory, with up to 160GB available as video memory. The 192GB model showed as out of stock on HP's UK store on 14 November.
11:09
The 14-inch M6 and the tandem-OLED MacBook Pro with M5 Pro or M5 Max began shipping on 6 November, with up to 128GB of memory on the top model.
11:09
Nvidia says RTX Spark PCs have been available globally since 16 October. In the UK the Surface Laptop Ultra is in stock, but the 128GB ProArt P16 is a pre-order due on 20 November.
11:08
Nvidia's RTX Spark laptops are on sale, Apple has new MacBook Pros and AMD's 128GB machines are still the memory kings. Here is which laptop suits which AI job, with the numbers the makers publish.
09:16
Tested inside our Minisforum MS-02 Ultra. Sixteen gigabytes runs 8B to 20B models comfortably at about 60 tok/s, draws roughly 49 W flat out, and stops at the 30B class.
09:16
We ran the same benchmark suite on both machines. The Ryzen AI Max+ box holds far bigger models and runs them faster; the Intel box wins at all-thread CPU work and Postgres, and keeps a swappable GPU.
09:01
Nvidia's founder resigned as chief executive on 12 November. Here is what is confirmed, what has been reported and attributed to named outlets, and what remains unknown.
20:58
OpenAI has launched ChatGPT Wallet, which lets AI agents shop and check out on a user's behalf, with a limited initial rollout to US ChatGPT Pro subscribers, The Verge and TechCrunch report.
18:56
AMD has released version 6.0 of its AOCC compiler and AOCL libraries, according to Phoronix, with Zen 6 optimisations in both and a new Fortran compiler called aof.
18:56
The ggml-org team released v1.0.0 of both llama.cpp (libllama) and ggml (libggml) on 13 November 2026, and says major-version bumps will now mark API and ABI breaks.
13:38
Wccftech reports that AMD's optional Adrenalin Edition 26.11.1 driver addresses a problem in which the software kept writing data to the storage drive in the background, though the fix has not been independently verified.
13:38
Google says it is bringing Gemini 4 Hydrogen to Google Search globally, though its announcement gives no detail on timing, language coverage or which features change.
13:38
Windows Central and TechCrunch both reported on 12 November 2026 that Microsoft services including Outlook were down, and TechCrunch says Microsoft has not yet given a cause.
13:38
PCWorld reports that Microsoft has patched an actively exploited Outlook vulnerability, CVE-2026-124194, while a separate outage left some users unable to reach Outlook email on 12 November 2026.
13:38
Windows Insiders in the Beta Channel can reportedly change common system settings from taskbar search without opening the Settings app, according to TechPowerUp and Wccftech.
13:38
Minisforum has announced the MS-S1 MAX-P495, a compact local AI machine built around AMD's Ryzen AI Max+ PRO 495, but has given no price or release date.
13:38
OpenAI has announced GPT-6 Cyber, a cybersecurity model for finding exploitable vulnerabilities and developing fixes, though access, pricing and independent evaluation remain unpublished in the material we have seen.
26 articles