The purpose of this product can be seen in the sea of slides, information, and "difficult words". Intel envisioned the Iris Xe MAX as a "comrade" in the new Tiger Lake processor to increase the productivity of ultra-thin and light notebooks.
Intel has released more information about the first "non-integrated" GPU of the new generation called Iris XE Max
The Iris Xe MAX uses a technology that Intel calls Deep Link, which involves the division of labor between integrated and separate graphics, for example for multimedia encoding jobs. This technology also combines the potentials of deep learning and artificial intelligence with integrated and separate graphics. Intel claims that in a specific scenario this combo of integrated and separate graphics has more than twice the encoding performance compared to NVEC in on Nvidia GeForce RTX 2080. Software that already supports Deep Link includes Hand Brake, OBS, XSplit, Topaz Gigapixel AI, Huya, Joyy, while Intel is working to expand support to Blender, Cyberlink, Fluendo, and Magix.
Energy management and performance levels should also be greatly improved over earlier integrated solutions.
The GPU features themselves don't seem overly impressive and this solution on paper crosses the spear with the Nvidia GeForce MX350, but Intel doesn't even intend to treat this GPU as a gaming machine (although gaming in 1080p is quite feasible). It remains to be seen to what extent the target audience, which should be those who "create" their content on light laptops that do not have room for "thirsty" graphics cards, will be.