DEC suffered from a problem that you can see throughout their history in different contexts: partners invariably came to be seen as competitors.
A good example late in their evolution can be seen in MIPS/Alpha. In the late 1980s it was clear the VAX was running out of steam - their own sales teams were telling them in no uncertain terms that customers who had VMS platforms were interested in new ones, but absolutely no one else was interested even in the slightest. If they needed "big iron", IBM's AS/400 was highly competitive, while those looking for compute performance for engineering and science could buy off the shelf RISC stations running Unix that were competitive at a fraction of the price and would clearly be faster than VAX within a couple of years.
Within DEC this caused massive confusion. With Olsen's ear, the 9000 team delivered DECs version of the Itanic, completely oblivious to the changing market. No one was going to buy one even if it was cheaper than IBM, because anyone even considering a non-IBM mainframe would just buy a Unix server. So there goes a couple of billion dollars that would have come in handy.
Others within DEC were perfectly aware of the problem and began development of their own RISC designs, only to be repeatedly shut down. This led the California offices to develop their own workstations based on MIPS. There were some real problems with the R4400, but all they needed was some time and money, both of which they eventually received. But that was not to be...
... because DEC decided that they could not, in good conscience, give money to some other company when they could take all of that money for themselves. So instead of helping MIPS they re-started their own efforts and delivered Alpha. Don't get me wrong, Alpha is a fine chip, but there are immutable laws in the building of chips and their resulting price/performance that DEC had absolutely no way of hitting. Had they stayed with MIPS, who was also selling to SGI, maybe it would have worked, but with their own chip in their own products? No way in a million years.
They could have aggressively marketed Alpha and moved to be a supplier, but those around at the time will no doubt remember their half-assed attempts. Other companies at the high-end had their own designs - MIPS, SPARC, POWER, etc., and, like DEC, felt their design was the secret sauce to future success and would never consider someone else's design.
That left only the smaller vendors who were not going to have the volumes required. And those guys were targeting the server market, which is precisely where DEC was trying to go. Once again, potential partners were instead seen as the enemy. And so DEC's big outside marketing effort was to offer one-off development boards and hobby kits. That was not going to sustain them.
So DEC became one of many companies pushing a RISC chip into a market that had room for perhaps one mass-market design. Intel was going gangbusters through this period, so by the late 90s the jig is up; x86 is offering reasonable performance relative to the RISC chips but at a tiny fraction of the price, and free Unix systems are available for download.
And that was that.