The dark satanic rumour mill has manufactured a hell-on-earth yarn claiming that Qualcomm is cooking two 2nm Snapdragon 8 Elite Gen 6 chips so phone makers can pay extra for the shiny one.
The claim is that the Snapdragon 8 Elite Gen 6 Pro and the Snapdragon 8 Elite Gen 6 share the same CPU cluster but differ in every other aspect that matters. Tipster Digital Chat Station says both will use a ‘2 + 3 + 3’ layout, which ditches Qualcomm’s recent ‘2 + 6’ habit on the Oryon-based Elite line.
The first obvious split is graphics, with the Pro tipped to get Adreno 850, while the standard part has to make do with Adreno 845.
Nobody has posted proper GPU specs yet, so expect the usual guessing about clocks and disabled blocks.
Cache and memory numbers are where Qualcomm really starts carving it up. The Pro is rumoured to have 18MB of GPU memory, which is 50 per cent more than the regular chip’s 12MB.
Last-level cache is supposedly 8MB on the Pro and 6MB on the standard version. That is not huge on paper, but it is the sort of thing that lets marketing shout about “smoother sustained performance” with a straight face.
The Pro is said to support LPDDR6 and UFS 5.0 out of the gate, which is the real flex. Qualcomm might even let partners drop the Pro to LPDDR5X and UFS 4.1 to shave costs and protect margins.
The standard Snapdragon 8 Elite Gen 6 sounds less forgiving, with LPDDR5X support only. That is the boring option for “premium” phones that still need to hit a price point.
LPDDR6 setups on the Pro are expected to run quad-channel 24-bit, while the regular chip is rumoured to stick with a slower quad-channel 16-bit arrangement.
All of these points to one thing: a Pro chip that benches like a monster and bills like one too. If the rumours hold, Qualcomm will be pushing one of the priciest mass-market mobile SoCs yet, built on TSMC’s 2nm process, with the “standard” version there to make the Pro look irresistible.
