After weeks of rumors, I now have confirmation from the Microsoft website and the man himself. MIT theory prof Madhu Sudan has accepted a permanent research position at Microsoft New England. In standard practice, Madhu is keeping his options open by officially taking a one-year leave from MIT in case the move doesn't work out.
Madhu is one of the best researchers of his generation having done fundamental work in PCPs and coding theory for which he received the 2002 Nevanlinna award. He has had a number of extraordinary Ph.D.s and from what I hear an excellent teacher. Even though Madhu is just moving a few blocks physically, this marks a great gain for Microsoft and a loss for MIT and academics.
I won't comment on why Madhu is leaving MIT but I have a couple of other thoughts:
With all of my non-academic friends worried about their jobs in our current financial situation, I feel quite safe with my tenured position. To give up a tenure in this economy seems risky though someone like Madhu could always find a job.
Nice to see Microsoft willing to make a commitment with a new hire of a pure theorist.
MIT, where I got my Ph.D., has seen the loss of few great theorists over the past couple of years: Dan Spielman, Ronitt Rubinfeld and Santosh Vempala as well as Madhu. Nevertheless MIT has hired some strong junior people such as Scott Aaronson and Jon Kelner and remains an amazingly strong place in theory, though the field is leveling.
If the economy continues to decline, I'm not so sure that having tenure is going to be the security that people think it is. Maybe at really big schools like MIT, but some other schools are just not going to be able to afford their faculty if things stay the way they are. Also, it's going to require a lot more work from faculty if they do not hire next year like they did this year.