Radu,
I agree with your general argument that there is a far better chance of Ricoh-Pentax acquiring a 44x33mm CMOS sensor for its "645D" system than of this coming from the initiative of Hasselblad or Phase One. One key is that Pentax might be willing and able to pursue a substantially higher volume, lower price, lower margin business model than the others, and for all I know, this might be enough to get MF CMOS past the tipping point into commercial viability. It seems that you, Bernard and I are all interested in seeing that happen. But I must repeat the point that Stefan Steib and I have been making: there is a substantial difference specifically between CMOS and CCD sensors in the cost of the initial development of each model of sensor, in the stages from design through to having the mask ready to deliver to the fabrication lines. That is why the Fujifilm DMF back is not relevant to this important CMOS vs CCD commercial viability distinction.
The key is whether any price high enough to cover the unit production costs (plus a bit of profit) can boost unit sales far enough above where they are now, which I believe is about 10,000/year total for all DMF brands and models. Does anyone know what MF camera unit sales were like in the film era? I have heard as high as 100,000/yr. If so, and if one or more less expensive DMF cameras using a single model of 44x33mm sensor could grab sales of 20% of that over a four year lifetime for the sensor model, then the $20,000,000 development cost that Stefan indicated would only contribute about $250 to the component costs of each camera, so that could be viable. For comparison, the sensor for the original Canon 1Ds was probably the lowest volume one in the DSLR world, and if I recall correctly, that was produced at 2,000 or 3,000 a month, so 24,000 - 36,000 per year selling at $8,000. That might be the sort of sales volume that a MF CMOS sensor needs to aim at. However, for Canon, that was only part of a plan for far greater total volume of CMOS sensor sales in later models, so Canon might have been willing to accept losses on that model for the sake of building its market dominance.