Intel is showing no signs of backing down in the competitive dGPU arena. Recent shipping manifests have revealed progress on their next-generation Battlemage models and another intriguingly cryptic SKU.
### Intel Unveils Plans for Battlemage “BMG-G31”; A Move Towards AI and Professional Applications?
Last year, Intel caught us off guard by announcing the Arc B580 and B570, part of their innovative Battlemage GPU series. These GPUs sparked substantial interest, primarily because the latest graphics solutions from NVIDIA and AMD were yet to hit the shelves. As we near the end of the second quarter, Intel’s silence on their desktop GPU strategies seemed to suggest a dwindling ambition from Team Blue. But, fresh insights from shipping manifests paint a rather different picture.
The latest manifests have surfaced a newcomer labeled “BMG-G31”. This silicon is reportedly superior to its Battlemage predecessors, featuring an estimated 24-32 Xe2 cores, a 256-bit memory bus, and 16 GB of GDDR6 memory. Despite rumors flying about the BMG-G31 potentially being axed, further investigation suggests Intel is not gearing this up for the consumer market. Instead, it seems poised for AI and professional use, indicated by its classification under “R&D purpose” in the shipping details.
In addition to this, there’s buzz about a peculiar variant, the BMG “C32”. Its name is a head-scratcher since it melds initials from both Battlemage and Celestial series. Nonetheless, insiders, including @mikdt, speculate that this version is part of the Battlemage lineup, possibly a “reworked” iteration of the BMG-G31. It appears Intel is indeed developing several Battlemage versions, but whether any are destined for the desktop market remains uncertain.
With Intel hard at work on the Xe3 “Celestial” architecture, expected to be integrated into Panther Lake SoCs, the company’s strategy for its discrete GPU lineup certainly piques interest. Under the leadership of Intel’s new CEO, there’s a marked pivot towards AI-focused technologies, hinting that their upcoming GPUs may well be tailored more for AI-driven tasks than traditional consumer markets.
