Jim Carlton in his book has documented Apple's first foolish attempt between 1986 and 1989 that ended up flushing down the drain $20 million as follows:
...an Advanced Technology Group engineer named Sam Holland, who in 1986 convinced Gassee that Apple should design its own microprocessor chips.The italics emphasis is mine.
Believing that Gassee and the engineers knew best, Sculley gave Aquarius his unequivocal support, to the extend that he even bought the engineers a $15 million Cray supercomputer they could use to perform their design work on.
The underlying motivation behind the Aquarius project was Apple's fear that the Motorola line of microprocessors that it was dependent on would soon run out of steam and be surpassed by the rival line from Intel.
Holland proposed that Apple design not only a new chip but one that would incorporate four microprocessors onto a single piece of silicon.
The only problem was that this had never been attempted before in a computer as small as a desktop.
"Aquarius was an example of the willingness of the top management of Apple to embark on grand projects that skilled technical management would have known was unfeasible," Alcorn [Al Alcorn was an Apple Fellow] says.
The project was shut down after a chip expert called Hugh Martin was recruited to help with the Aquarius project and this exchange between him and Sculley led to the cancellation of this bold & foolish attempt:
"He said, 'What do you think about Aquarius?", Martin recalls. "I said, John, that's ridiculous. Apple has no fab or chip experience. How do you compete with Intel and Motorola?"Within 6 month after that conversation, the project was canceled.
Let's hope Apple management can still remember this secret project and learn from the history. Now Motorola is effectively out of the contention and Martin's question to Apple in 1989 was as appropriate to Apple as it is now, except that you can substitute Motorola with AMD.