How to write a PhD thesis in three months?

Make an outline and a schedule, then follow the schedule. Include time for revision and feedback in your schedule.

If you know all the content already, the writing can be done much faster than that. In my field, it is only necessary to write about 90 generously spaced pages, so three months is less than a page a day. In more verbose fields, your thesis need not be more than 500 pages, so that's an achievable 6 pages a day.

The real problem is when you do not know the content you need to write - that part we cannot help with here.

I wrote my thesis in 10 days. Granted, I was well published before, so 3/4 of it was a cut and paste.

This was 90s so I had to use lab computer for typing. Went in at night and typed. Then did a morning run, ate breakfast and slept. Avoided distractions from lab group.

Avoided complicated LaTex and just did it old school. Even xeroxed in figures versus fancier drafting program. Did individual chapters to correspond to papers. Picked easiest endnote system (by chapter). I think I even did individual files by chapter because Word got complicated trying to do multichapter work.

At the end of the day, it's pass fail. Just get it done.

Having done this myself, one thing I did was work at certain desks in the stacks at my college library that happened to have metal walls. This acted as a Faraday cage, blocking all wifi and cell phone signals. It was very helpful for focus, while a the same time allowing me to easily access the internet if I needed to look up a paper simply by standing up and walking out of the work area. You may not have something similar, but the key issue of removing easy distractions like phones and especially the internet is very important.

Similarly, I would suggest not using your personal laptop to do this--it's another way for distractions to creep in. You'll want something that can download papers, compile TeX, and pretty much nothing else. You may be able to get a loaner laptop from your department for work purposes or get an old used laptop for cheap.

I also heavily relied on copy and paste from my previous papers, and even some of my advisor's papers. Obviously the sections from my advisor's papers had to be rewritten, but it gave me a complete draft, which was a big change psychologically. Having a complete document that needs lots of revisions is way different from having an incomplete document, even if the amount of work left to do is similar.

One last thing--keep a regular sleep schedule that gets 8 hours of sleep and involves waking up before 8. This is something grad students are notoriously bad at, but you'd be amazed at how much better you feel.