It took me until just five seconds ago to realize something: a plethora of people who speculate here don't actually have fun - they speak every word as though its truth or falsehood might be some slight against the purest core of their existence. Either that, or they just don't use enough emoticons.
Most of a processor's cost is in its development. And even then, that cost is likely minimal in a Digic - if only because its a lightweight processor - on a par with an ARM driven package. If they've designed it correctly and are using an appropriate process, the cost per chip is negligible. Each package equates to about $100-$200 of the consumer price, assuming a top of the line brand new design. Thus a dual CPU package is about $200-$400. Which is piddily compared to the other components in a camera in the X series.
Power - again, arguments can be made, but the draw is probably somewhere in the 1500-3000 milliwatt range at max draw. So double would be 3w to a maximum of 6w - and that's a peak number. For most processes, the device is going to sip power at somewhere in the 300-700 mW range per package. Its probably the fourth most power hungry system in the camera, following the sensor, lens, and shutter/mirror assembly.
So if the device required it, I see no reason why canon would not use a dual processor system if it gave them the ability to give the 5DmIV an edge over its direct competitors and to make it attractive to those who own its predecessor.
As for Megapixels - this is the one thing Canon is more likely to focus on in this camera. Its been their end game for the last few generations with it and its the one thing they are okay with the 5 series outperforming their 1 series cameras on. And they don't really have to do much to achieve a decent jump from iteration to iteration. Moores law dictates that computational capacity doubles roughly every 18-24 months. The Digics, unlike most high end CPUs are likely nowhere near that threshold and have probably another decade before their designs start pushing the limits of physics. Assuming a spherical cow... You're probably going to see a practical gain of 20%-30%. So taking a 22.4 megapixel sensor and taking it up to 29.12 megapixels or thereabouts should be easily done without any drastic changes in design or capabilities while imparting the same image performance. Canon could also achieve a similar gain by switching to back side illumination on the sensor without changing the performance of the processor. Or they could put in two processors for another 20%-30%.
The latest Digics support C-Fast. You could pump 32 megapixel raws to the c-fast card at 10fps. So if you have a moderate buffer (4gb), you could easily sustain a 10fps burst for 12 seconds before experiencing a mild falloff to about 9.6 fps (due to the transfer delay from file to file) until the card was filled. For UHS-II, the maximum fps rate to card would be about 7fps at 32mp, and about 8fps at 29.12mp but there is no sign the digic system supports UHS-II SpeedClass 3. ISO 100-400. More noise = more megabytes after lossless compression so... *shrug*
Canon is unlikely to put 10fps on the 5DmIV. 8fps might happen, but odds are we will see 7fps so that either CFast or UHS-II cards can offer near "limitless" buffer given a fast enough card and keeping the high speed buffer storage down to a minimum. Its this that will take $$$ out of the cost of the electronics as memory is the most expensive component in the system due to how fast it must be in order to transfer data from sensor through to the processor(s), then back to the card. 7fps will require a small push to the AF system, but they can just take what the 7DmII already has and adapt it. They would likely also need to dampen the shutter a bit better.
If they wanted to make a sensor capable of being used for 8k DCI video, they'd need something in the 42-50mp range. Which would be limited to about 3-6fps at buffer exhaustion depending on storage types. They'd also need an 8gb buffer to store a significant amount of raws. And if they wanted to improve noise performance, they'd need to switch to BSI, as there's a law of diminishing processor returns once you start taking the pixel density to where the 7DmII and 5DS/r are unless you have appropriately increased processing capacity. I'd say that's a 2020 or so thing.
Again, that's assuming that canon isn't in its usual conservative "Thou shall not cannabilze thy sales" mode. But I really have a hard time not seeing canon bump to 29-32 mp, as it would cost them little to do so. And because of the 4k video (assuming its there and uses mjpeg) they really have nothing to lose. You're going to need CFast media for that so you're already going to have to make the investment in the cards. They wouldn't lose sales or performance with the megapixel bump. If they forego the 4k video... or use a highly compressed format... then UHS-II and CF would work just as well... and you'd only need a 24mp sensor. Although I think such a iterative upgrade would lose more new customers than bumping up would lose existing customers, and would make them even farther behind the field when it comes to video capability.
My two cents... for what its worth.