There are plenty of stories on the rumour of AMD acquiring Xilinx. AMD is growing now and its products looks lot more interesting not only for the reason of cost but also performance. While AMD is growing and ambitious but aiming for acquiring large company such as Xilinx can be seen as "too early".
AMD still need to show some continous business success in AI processor products. Programmable silicon strength is required in AI. Can AMD start afresh in that or go for acquisition? It’s a chicken and egg situation. AMD's internal tech-team will be knowing what there real strength in AI. AMD's top management can decide to buy the tech-prowess or/and develop internally. When company decide to buy, there will be this option of Xilinx or some startups in FPGA. Technology-wise Xilinx is a very good option for the huge IP they have in silicon and software (including design automation software) and the market they hold. But the question of cost of investment will be the tough nut to crack for AMD's top management. It’s a high risk decision.
For Xilinx, however strong it is in FPGA specific market, company may find it difficult to expand in mainstream large scale semiconductor market. So for the revenue growth and expansion-sake, there is an option of either buying a main stream semiconductor vendor and allow it to get acquired.
If not immediatey sooner you may see either/both AMD and Xilinx opting for acquire or merge/sell-off. Consolidation is the crystal clear trend in semiconductor industry.