Intel is clearly not giving up on making waves in the discrete GPU market. Recent shipping documents have unveiled progress on the new Battlemage GPU and a mysterious new SKU.
Intel is Back at It with the Battlemage “BMG-G31” Edition: Possible New Model for AI and Professional Uses
Last year, Intel caught us all off guard with the introduction of two fresh Battlemage GPUs: the Arc B580 and B570. They stirred up a lot of interest at that time, mainly because NVIDIA and AMD’s next-generation solutions weren’t available yet. Now, as the second quarter wraps up, Intel has remained silent about any plans for their desktop GPU lineup, leading some to believe they have no big ambitions. However, newly revealed shipping documents tell a different story.
Intel’s shipping manifests hint at the existence of “BMG-G31,” a more advanced silicon than what we currently see in the Battlemage GPUs. Rumor has it that this variant could sport 24-32 Xe2 cores, a 256-bit memory bus, and 16 GB of GDDR6 memory. Nevertheless, whispers of its cancellation have surfaced, suggesting that Intel might not bring this to the consumer market. Interestingly, the shipping docs categorize BMG-G31 as being for “R&D purpose,” which suggests it might be angled towards AI and professional workloads instead.
Alongside this, there’s chatter about another intriguing variant, the BMG “C32.” The naming is a bit puzzling since it combines initials from both Battlemage and Celestial. Still, word on the street is that this is part of the Battlemage lineup, possibly a “reworked” version of the “BMG-G31,” according to @mikdt. This points to Intel working on several Battlemage models, but there’s still some mystery about whether these are aimed at desktop users.
How Intel will strategize its discrete GPU offerings remains a point of interest, especially with the company’s current focus on the Xe3 “Celestial” architecture poised for Panther Lake SoCs. With Intel’s new CEO directing attention towards AI, it seems Team Blue might be planning to roll out GPUs geared towards AI-driven tasks.