CES 2026 spotlights NVIDIA Vera Rubin landing on Azure’s AI platform.
Years of co‑design and Fairwater upgrades mean Azure already supports Rubin’s power, cooling, NVLink‑6, HBM4/HBM4e, SOCAMM2 memory expansion and 1.6 Tb/s ConnectX‑9 networking — enabling ~50 PF NVFP4 per chip and ~3.6 EF per rack.
Integrated compute, storage, networking, orchestration and cooling enable faster deployment and scale for customers.
#AI #Azure #NVIDIA #Rubin #DataCenter
Microsoft DevBlogs
npub12k…7q7ck
2026-01-16 06:09:01 UTC
Author Public Key
npub12ku8cz3lqswqstz263x84f78855jmtpgjvkvq58culk50z440txsn7q7ckSeen on
wss://relay.momostr.pinkPublished at
2026-01-16 06:09:01 UTCEvent JSON
{
"id": "c14e619eb7054229cc78092ce123c05d45c182d2208b55f932b8702b5064b515",
"pubkey": "55b87c0a3f041c082c4ad44c7aa7c73d292dac28932cc050f8e7ed478ab57acd",
"created_at": 1768543741,
"kind": 1,
"tags": [
[
"t",
"azure"
],
[
"t",
"datacenter"
],
[
"t",
"nvidia"
],
[
"t",
"ai"
],
[
"t",
"rubin"
],
[
"proxy",
"https://dotnet.social/@msftdevblogs/115903282662672027",
"web"
],
[
"proxy",
"https://dotnet.social/users/msftdevblogs/statuses/115903282662672027",
"activitypub"
],
[
"L",
"pink.momostr"
],
[
"l",
"pink.momostr.activitypub:https://dotnet.social/users/msftdevblogs/statuses/115903282662672027",
"pink.momostr"
],
[
"-"
]
],
"content": "CES 2026 spotlights NVIDIA Vera Rubin landing on Azure’s AI platform.\n\nYears of co‑design and Fairwater upgrades mean Azure already supports Rubin’s power, cooling, NVLink‑6, HBM4/HBM4e, SOCAMM2 memory expansion and 1.6 Tb/s ConnectX‑9 networking — enabling ~50 PF NVFP4 per chip and ~3.6 EF per rack.\n\nIntegrated compute, storage, networking, orchestration and cooling enable faster deployment and scale for customers.\n\n#AI #Azure #NVIDIA #Rubin #DataCenter",
"sig": "032deffc9160ccfe915a900125c8af4741a50e6f8591c9c46ae867cc7268735b63a5e12bf4d2cab26151ef6fb53b035e64e77fefbd2daf8cda84cde335b5790a"
}