AMD currently has two graphics technologies that serve quite different purposes. The first is RDNA which is used to power graphics cards for gaming purposes, and the second is CDNA which is used to power graphics cards in data centers for purposes such as processing 3D objects, performing complex calculations and so on. The two technologies were separated to allow them to move and be developed separately around 2019, when RDNA was introduced for the first time.
Following in the footsteps of NVIDIA, which uses CUDA core technology to perform all these tasks, AMD has decided to do the same thing, combining RDNA and CDNA technology to perform both of these tasks using the same graphics technology.
It is called UDNA.
This is to enable AMD engineers to develop a better product, where only one offering is needed for ordinary users and data centers to take advantage of the variety of features of a graphics card, from the ability to process graphics for computer games to the ability to process 3D objects or resolutions high and perform a variety of machine learning and artificial intelligence tasks, something that is increasingly common among today's users.
This will fix AMD's issue where they have two teams working on two different products when the workforce of these two groups can be used to build a good product that can be used for different purposes.
For now, AMD hasn't given any confirmation as to when they will start introducing this UDNA graphics technology, but we now know that they will launch their RDNA 4 series of graphics cards soon, and it's also almost certain that they are at least halfway there in developing RDNA 5 technology, so UDNA will probably be introduced after that.
We are sure that AMD will still maintain the difference in graphics card models such as AMD Radeon and AMD Instinct, but in short, both series of graphics cards will use the same UDNA graphics technology inside.