Surely if dexcel designed the ASIC and the foundry made it and it didn't work wouldn't they sue them?
I brought this up awhile back, I found it odd that a proven chip design company designed a chip which didn't work. My first thought was that perhaps something went wrong in the foundry, afterall AT claims to have paid for silicone which didn't go through QA testing, in order to get it faster.
My other thought was, if in fact they were scamming, then it was a ruse all along: Paying for bum chips, and having something go wrong--all so they could build themselves machines to mine with on your dime (aka, for free), at which point they would ship them out in "September", having already ordered another batch with the new money. Remember, they did say they would build an in-house cloud mining center for the customers--and then stopped talking about it.
The other scam thought was that none of this ever actually happened--there was no design and no foundry. We do know that Drexcel, when contacted, claimed to not know who AT was and denied any relationship with them at one point.
Who knows..
It may not be that the chips "did not work". When ever a wafer is made there are always a number of chips on the wafer that will have low quality. The quality may be so low that they have inherent defects, like they do not perform as fast or overheat more often. Alpha's contract with Drexel was probably to produce a chip that has a minimum specification. Drexel provided that. But Alpha probably wanted to get the chips faster, so rather than complete a full tape out with several spins before masking, Alpha bought a lot of risk chips.
This is where Alpha's incompetence shines. I speculate that Alpha thought, that getting chips early would be a win. They can start assembling the risk lot chips first, and when the money starts rolling using the higher quality chips in later batches which they can call "optimized". What they didn't realize was the significantly poor quality. Which is typical of the first tape out.
So Alpha spent the remaining money (the 30%) to get the risk lot and on what ever else. So they probably couldn't pay off Drexel. So Drexel holds on to the chips. This is when Alpha comes up with the idea in May to announce tape-out and final payment. They figure they can pay Drexel or the Foundry with the remaining funds.
But people want CC and they want the prototype and they hold off paying. This is where everything goes down hill fast.
