What you forget is that Nintendo takes the money that they got from selling their 20+ million of one game, and uses it to pay for development of a heap of titles that don't sell as well, but round out the system's library to expand its appeal. And it makes financial sense for them, because it increases console sales (and they make money on every console sold) and thus expands the install base, thereby encouraging greater third-party development on the system.

To put it another way, Nintendo spending relatively small amounts of money on NSMB Wii and making heaps off it is what gives them the money and confidence necessary to make games like Xenoblade, Pikmin 3, or Disaster: Day of Crisis. Note that I chose those three titles to demonstrate the various possible outcomes - Pikmin 3 is likely to perform quite well (well over a million), Xenoblade sits in a comfortable but not strong position (650,000), and Disaster was actually quite well named (~50,000). Sometimes, it'll pay off, sometimes it'll flop badly. But it's a sensible approach to the gaming industry.

And of course, a lot of people criticise Nintendo for lack of new IPs... despite the fact that they typically have multiple new IPs each year. Just looking at Wii (asterisks means Japan only), there's Wii ___ (Wii Sports, Wii Play, etc), Disaster, Captain Rainbow*, Flingsmash, Dynamic Slash*, And-Kensaku*, Tact of Magic*, Pandora's Tower, Xenoblade, The Last Story, Kiki Trick*, not to mention the unfinished Project HAMMER, Wii Crush, and Cosmic Walker.

I would say it's a fallacy to equate extra grunt with more work. A lot of the enhancements are dynamic/procedural that the game engine handles;- lighting and tessellation probably being the two key areas that will separate the new generation.

It's not the grunt itself that causes a need for more work.

It's competition. With a system that has more grunt, developers are capable of making games with more detail and more complex visuals, which means more time required to develop code and to develop models, etc. It costs more do to these things, but some developers will be quite happy to spend the sort of money necessary (for instance, Call of Duty developers). And as a result, other developers will effectively be forced to also spend more, in order to avoid significantly losing out on comparisons.

Development of new tools to reduce the difficulty of developing code and models will happen, of course, and these will put some downward pressure on costs... but such development is slower than the increase in power, and thus net cost still increases.

Anyway, enhancements in lighting and tessellation don't require massive boosts of power. The next generation of game engines, you're right about... but the next generation of hardware is the problem, not the engines.

It is a trailer but in their words 

“The trailer consists of all the major visual features we wanted to implement and it was a way to test them,” he explained. “At the same time we wanted to express the mood of the world and the story because it is a big part of the whole. We combined these two so in a way the trailer itself is a byproduct of the actual development. We would be poorly allocating resources if we just made a trailer for its own sake. We’re not worried about the Dead Island implication, but only time will tell if we can make your worries disappear.”

“Everything you see in the trailer is straight from the in-game engine, no up-ressed textures, geometry or effects. What you see is what you will get. Except hopefully a little bit better since we’re not even in alpha yet,”

“The trailer is made entirely of in-game material, and I mean all assets, effects, everything, period,” Oksaharju said, noting that making separate tech for a trailer would be a waste of the team’s limited resources, and that licensed tech wouldn’t grant the same level of control.

The digital revolution has opened plenty of opertunities for devs to make smaller games, and next gen I imagine that XBLA, SEN and whatever Nintendo are calling theirs will offer a range of games from a couple dollers up to full AAA releases like Steam does. There is clearly a huge market for games that are not a AAA set peice filled $40 million production with multiplayer. As proved by the indie revolution that is going on with games like Minecraft selling multiple millions etc. The full priced retail AAA market will likely become even more focused on a few big titles as they all try to one up each other, not because of higher end hardware but just because things like BF3 vs COD:MW3 where EA duel to see who can make a bigger more bombastic game to fight over the same market. 

You're right, I'm just wondering what extra work is now involved - I would assume that the basic 'asset creation' work-flow would remain the same.  Everything is Zbrushed and re-topologised at the moment, it's just that they'll have more poly's to play with.  I guess the extra work will be the little details in the environmental assets that currently are breezed past as super low poly simply because the power isn't there. 

No, that's a lie.

Nintendo said that they're not following the Microsoft and SONY with powerhouse numbers back in Revolution time.

Because 360 even came out before wii, and in early 2005 they said that their competition going for graphics and Nintendo is going for new take on how you play a game.

*sorry for my bad english

So what is he suggesting? Should they not release new consoles with more power? The market has been stagnating for quite some time now, new hardware helps to infuse new life into the gaming industry.

Sure the consoles will be expensive at start, developers probably won't be able to use full power of the consoles at the beginning, I mean the 360/PS3 have only recently been showing what they are truly capable of. Plus if the dev costs are high, it will only mean that less shovelware and better games which are more likely to sell.

They can't be doing graphical leaps, the WiiU will cost 250/300$/€ because Nintendo wants to a market price.

If Sony had sold at launch a console at 600$ what they can do with the PS4 with the graphical leap.

Also i'm here rewatching E3 2006 and listenting to Kaz saying: the PS3 will last 10 years.
And Kaz never is wrong. (sarcasm all over the place)

"if the dev costs are high, it will only mean that less shovelware and better games which are more likely to sell."

or it could mean devs going cheap and releasing shovelware all over the place, or devs. closing because they cant afford making a new game, the better the specs the lesser games youll see

"he thinks the consumers are stupid ?"

dude, the market is too big, not only vg fanatics, most gamers today are casual gamers who dont know a shit about gaming, most users dont care about specs and the tech behind the games
hes absolutely right on that spot...