Vram | Aillex — Do-It-Yourself AI

Two apparently separate pipes on a workbench that are revealed to join into one feed behind the panel

The Second Source for the 5090 Rumour May Be the First One.

On Saturday we wrote that the RTX 5090 discontinuation story rested on a single leaker’s post and said we would report back either way. Three days on there is still no statement from NVIDIA, the card is still listed on its own product page, and the second leaker being cited may be the same person as the first. Meanwhile people are picking replacement cards today.

A row of shelves where the largest component bays stand empty while the smaller ones remain stocked

One Post Says the 5090 Is Ending. Here Is What Would Actually Change.

NVIDIA is reportedly steering the GB202 die to its professional cards, which would leave the RTX 5080 as the top GeForce. Every outlet carrying it traces back to a single leaker’s post from this morning. The interesting question is not whether the card dies. It is what happens to the amount of VRAM a person can buy new.

Two enormous datacenter towers close together with a small desktop computer far below and apart from both

Open Models Are 4.4 Months Behind the Frontier. That Is Not Your Gap.

A Mozilla analysis put the open to closed capability gap at about 4.4 months, and it went round the local AI world as good news. It is good news. It is also being read as something it does not say, because the open model doing the catching up is 2.78 trillion parameters and will never see the inside of your computer.

A ladder of graphics hardware with an expensive card at the top and a small integrated chip at the bottom, the gap between them stretching

Someone Paid $6,279 for a GPU and Someone Else Ran the Same Job on Integrated Graphics

Both of those happened on the same day, in the same corner of the internet, running comparable work. Prices at the top are climbing, the used enterprise shortcut is failing in public, and the floor keeps quietly dropping. A note on the strangest hardware ladder local AI has had yet.

One graphics card with a model too large spilling off it and streaming to disk, beside another card with two workloads crammed inside colliding

Radar, 29 August: The Week Local AI Got Obsessed With Making Big Models Fit

Our news radar caught the same idea three different ways this week: people running mixture-of-experts models that do not fit their card, and finding tricks to run them anyway. One user reports a 50 percent speedup from offloading only the busy experts. Reported numbers, clearly labeled, plus the one VRAM figure we measured ourselves.

Wan 2.2 Locally: I Benched 5B vs 14B on One Card

Everyone’s arguing about Wan 2.2. So we ran it locally, honestly, the fast-and-fits 5B, the heavy two-expert 14B, the VRAM wall, and the decode that wedges. Here are the real numbers.

Which AI Rig Do You Need? The Honest 2026 Hardware Guide

VRAM first, everything else second: what GPU you actually need for local AI in 2026, the memory-crisis market, the value picks, AI appliances like DGX Spark, four real builds by budget, and when renting beats buying.

Can My PC Run Local AI? Instant Hardware Check

Pick your GPU and RAM, get honest answers: which AI chat models, image generators, voice cloning, talking avatars, video generation, and full AI companions your machine can actually run, from people who run all of it daily.