The debut of Intel Arc Alchemist was bumpy and the truth is that this generation came a little late, but it finally showed that the chip giant it can gain ground in the mainstream consumer graphics card marketand the truth is that at least it left me with anticipation and a desire to see what they could do with the Intel Arc Battlemage.

In case anyone gets lost, I remind you that the Intel Arc Battlemage will be there the next generation of graphics cards giant Santa Clara. Before it will arrive the Intel Arch Alchemist Refresh, a revision of the current generation, the specifications of which have not yet been released, but will generally be minor improvements and retain the same architecture.

Intel Arc Battlemage is a generation that is really hot, but unfortunately we don’t have good news. Rumor has it that the folks at Santa Clara are having issues with this new generation of graphics cards, and for that reason they would have to delay their launch until mid 2024.

The release date is very important when we talk about a generation of graphics cards, because it can affect a large part of its success or failure, and in this case, delaying the next generation would not benefit Intel. Think about what happened to Arc Alchemist who was able to hit the market at a time when demand was huge and it eventually ended up being crossed with GeForce RTX 40 and Radeon RX 7000.

According to the source, this new generation of graphics from Intel is “plagued with problems”which is why the most optimistic release date we have now points to the second quarter of 2024. The higher-end models are also rumored they would be cancelledand that this generation will eventually cover only the middle and lower class with a total of three models that would theoretically be configured as follows:

• A mid-range model with a modified G10 GPU, 448 execution units (3,584 shaders), a 256-bit bus and 16 GB of GDDR6X memory.
• Mid-range model with G21 GPU, 320 execution units (2,560 shaders), 192-bit bus and 12 GB of GDDR6X memory.
• Lower mid-range model with modified G21 GPU, 256 execution units (2048 shaders), 128-bit bus and 8 GB GDDR6.

All these GPUs would be manufactured in TSMC 4nm nodeand would theoretically compete with NVIDIA and AMD’s mid-range and lower-mid ranges within the current generation. The G10 core was said to have a higher configuration, probably with 512 active execution units, but that “not feasible”. I understand you mean that it is not possible to get wafer units with all their shaders working (4096).