Well it went to $6 billion with some really smart and complimentary acquisitions (Interwoven, Meridio, ZANTAZ etc), and I even went on record as saying that potentially Autonomy could be the UK's next CA or Oracle.
Then HP paid (literally) 'top dollar' at $10 billion for the company in 2011, but didn't really appear to have a game plan for the integration of the technology into HP's developing information management stack. It was noticeable that HP didn't want to keep Mike because they didn't appoint him to the board. It was also noticeable that Autonomy became a very independent unit within HP, even to the point of re-hiring people it lost when it was first acquired.
Now HP says it has to take an $8.8 billion charge on the acquisition after "serious accounting improprieties".
Were there "accounting improprieties"? I am not qualified to say. But if I was going to spend $10 billion on a company with a market cap of $6-7 billion, I would want to do a heap of due diligence first. Whilst some things will always be missed in the M&A process, overestimating by 88% a company's value (which is what HP now believes) is a huge error on any accountant's spreadsheet.
So what next? Lots of fees for lawyers and (hopefully better) accountants, and there is a good chance that much of the really good software that Autonomy brought to HP will be lost in the melee.
My final thought is of a sign I saw in a tableware shop in Palo Alto "You broke it - you bought it".