It shouldn't be a specific length.
Software development costs due to poor priorities, unreasonable scope, scope creep, and generally weak project management practices should be resolved. The addition of power/an expanded resource budget that comes with more powerful hardware ends up skirting problems with engine optimization or similar measures, allowing such things to bloat and result in bottlenecks that wouldn't otherwise have been there. Architecture has been broadly unified and the biggest third party engines adopted today have been bolstered with porting and profile tools, making large multiplatform releases less difficult than ever before. Despite this, even those things haven't dampened the current trajectory of AAA development. A new console generation doesn't end up helping much on that front and can be an excuse of detriment in the long term.
Regarding consumer-side engagement, keeping the player base of an established console is typically the more advantageous direction. A large install base that continues to use the hardware means more room for first party software sales. The singular digital marketplaces and platform certification costs makes software a profit center of pivotal magnitude on the digital and physical front for all consoles. As manufacturing costs diminish, buy in is easier with strategic discounts and the attractive prospect of large, established libraries. Because of this, elongated console generations make sense depending on performance of the current hardware or projected performance/appeal of future hardware.
Because of growing inflation both on a manufacturing cost level and software development (perhaps leaning more in the west on that front), a new generation becomes riskier as time goes on not just for the health of a given console's hardware sales, but the overall industry ecosystem as we continue down the ever more evidently unfeasible trajectory. The 'death of the middle' regarding moderately sized publishers and software projects was offset by a enormous rise of indie software expanding the low end and beyond, but platform holders (excluding Nintendo) and the major publishers that have been staples of the industry have some big issues to deal with as more modest projects continue to dwindle and the risks/lack of guaranteed profit on big games forces reevaluation.
On top of the previous issues with cost bloat, development time increases mean that a given platform sees less flagship software over its lifecycle. The PS5 is not a platform that has been presented as abandoning exclusive games. Indeed, Jim Ryan bolstered the concept of leveraging the new hardware for unique experiences and said that Playstation 'believe in generations'. Eric Lempel, head of Playstation global marketing, went even further in reinforcing the idea where new software and experiences are concerned, but the amount of first party releases have been meager. Even including those titles that made their way to PC, the amount of exclusives to bolster the value proposition of the console is far, far behind the last generation and those prior. With third parties continuing to release software for the previous generation even now so far into current gen, that issue is exacerbated massively, further dwindling the clear value/advantage of new hardware in the present context.
Xbox has left the exclusive space in the middle of the last generation outright and is finally managing to put legs on its alternative software strategy through the Game Pass service, but this is at expense of Xbox hardware being the forefront or center of their gaming ecosystem proposition. This makes their issue deliberate and different as a tradeoff, but it also helps make the effects of that vacuum more notable and worth considering regarding where PS is going.
The reality is that new hardware is no longer facilitating as much evolution in game design and the power increases aren't as impressive on the screen with each leap since about generation 7. Mid generation refreshes have been breaking up generations and adding more SKUs/profiles to the table for developers to establish on singular projects. I think all of the issues mean that new generations are actually becoming bigger risks, especially in the transitionary period. With things moving as they are, a console generation would have to be closer to 12 or so years in length to see a respectable library formed, and that's obviously not feasible as technology continues to move forward even in spite of the less pronounced graphical elevations.
For my part, I think the industry needs another evolution. Whether it be in a new market segment like indies, the distribution system, more lasting game frameworks (Fortnite style support building on a game as a platform), or -especially- a revolution in project management and a way to make more reasonably scaled titles appealing/valuable in the modern context. Until then, new console generations just become less useful all around, mid generation refreshes dilute them, and any attempts to minimize transitionary losses end up undermining the new generation outright.
Unless you're Nintendo.
So far.