An unattributed figure travels fast

The "2x encoding" claim has circulated in community forums without any traceable origin. None of the posts examined pointed to a published test, a vendor announcement or a named reviewer. The person who requested this fact-check noted the same problem: the figure is everywhere, but the source is nowhere. That pattern — a precise-sounding number with no citation attached — is often the first sign that a claim has detached from its evidence.

What the 12-source review found

The fact-check examined 12 primary sources, including NVIDIA's own materials for both cards and independent technical coverage of the RTX 5090's media capabilities. No source substantiates a two-fold gain in media encoding performance over the RTX 4090. Vendor documentation for the new GPU does not advertise any such jump, and no reviewed test reports documented measurement conditions that would produce that result. The review also notes that the sources consulted do not establish what the actual generational difference in encoding performance is, one way or the other — the doubling figure is the specific claim that fails verification, not the broader question of whether the newer card's encoding engine has improved.

The gap between improvement and doubling

Some genuine generational refinement in a new flagship GPU's media engine would not be unusual, and the verdict rendered here leaves room for that possibility. But refinement and a 100 percent gain are different things, and the burden of proof for the stronger claim sits with those who made it. On the evidence reviewed — 12 primary sources, none supporting the doubling figure, and community posts that cite no source at all — the claim that the RTX 5090 offers twice the media encoding performance of the RTX 4090 is Mostly False.