Perhaps someone could remind me of the terms of the GPT. I thought I remembered that the GPT had to achieve at least a COP>=4 for 250 days continuously. I don't think it required anything like the 1 MW output. Perhaps Rossi created the 1MW array of devices as a "reactor" to give himself the flexibility under computer control to swap out non-functional or marginally-functional units so as to maintain his COP over the reliability period. That would be a shrewd way of maintaining his claim of continuous reliability.
So, not making 1MW is not a problem for the GPT (I think). Even having a COP << than what was reported is not a problem, as long as it is found to have a COP of at least the minimum requirement of the GPT. The very high reported COP could be a ruse to have the real, much lower COP (but greater than the GPT requirement), come out by independent analysis as a surprise twist in the case. Perhaps the validity of the "customer" may be a sticky legal point, but it may be just a semantic and the court, particularly a jury, could overlook it. There is the other sticky point for Rossi - he has not lived up to his end of the bargain in usefully transferring the technology into IH so that they can make a product. That's what they paid $11M for. Technology transfer would seem to be a prerequisite to beginning the GPT. If Rossi is making a case for having a COP of at least 4 (to use that number for the GPT), how can he not have transferred that technology to IH before the GPT? Had he done this, everybody would be happy today. This is certain to weigh in a jury's decision.
