Don't know for certain. Had heard many times that a large majority of top WTA players prefer PP. For ATP players, a smaller majority... somewhat less than 2/3 but greater than 60%.
I recall that pressure (force) plate studies performed in the 1990s revealed that PP, on average, produced a greater ground reaction force. But the GRF difference between PP and Platform stances measured was relatively minor.
While many servers can generate greater racket head speed / power with a PP stance, this is not the case for all. Some pro and many rec players are more stable (balanced) with the Platform and actually serve faster with their implementation of that stance.
In his 20s (and possibly, early 30s), Federer could generate a massive amt of RHS for monster kick serves and often got his flatter serves over 140 mph with his Platform. Certainly, one of the most effective servers on the tour. Sampras, with his modestly-powered 85 sq" racket would hit his 1st serves in the 120-130 mph range. However, Pete was getting an incredible amt of RHS on his serves... 1st serves often around 3000 rpm and 2nd serves sometimes exceeding 5200 rpm... one of the heavier, and most effective, servers of all time.
Andy Roddick employed a narrow platform stance. He would often hit serves in the upper 140's and sometimes exceed 150 mph. He held records for the fastest serves for a good part of the 00s. Dominic Thiem, who switched from PP to Platform, has managed to serve up to 144 mph. Other conventional Platformers include: Milos Raonic (155 mph), Taylor Dent (149.8 mph), Frances Tiafoe (147.3 mph), Taylor Fritz (147.3 mph), and Fernando González (146.6 mph).