Just to get the terminology correct … "GRID" is the software component that lays over a given set of Tesla (Currently M10, M6, M60) (and previously Quadro (K1 / K2)) GPUs. In its most basic form (if you can call it that), the GRID software is currently for creating FrameBuffer profiles when using the GPUs in "Graphics" mode, which allows users to share a portion of the GPUs FrameBuffer whilst accessing the same physical GPU. GRID is always being enhanced and developed to offer better performance, feature enhancements and functionality, and this is why NVIDIA have opted for a software defined model, as opposed to a hardware model, where there are far more limitations.
No, the M10, M6 and M60 are not specifically suited for AI. However, they will work, just not as efficiently as other GPUs. NVIDIA creates specific GPUs for specific workloads and industry (technological) areas of use, as each area has different requirements.
Slightly off topic, but if you want the best (official) resources for AI, then you’re looking at a DGX-1 (https://www.nvidia.com/en-us/data-center/dgx-1/) or a DGX Station (https://www.nvidia.co.uk/data-center/dgx-station/). The DGX-1 uses P100s in combination with NVLINK. However, the P100s have very recently been replaced with V100s. As it is a brand new offering, the DGX Station starts with V100s and again, uses NVLINK. Note that they use “Compute” focused GPUs, not Graphics focused.
So, in answer to your question, yes, you can use the M6, and if you’re in “Graphics” mode you can share FrameBuffer between VMs and make better use of the GPU. However for the best performance, guide them to a more focused GPU line.