Stadia and Luna are dead the moment Xbox Game Pass Streaming is pre-installed in every Windows 10 PC, which should be happening next year. They will also upgrades their blades to Series X hardware, having Game Pass better specs than Stadia and Luna.

Phil Spencer has also said they will launch an Xbox streaming stick and a TV app for Game Pass in 2021. They have partnered up with Facebook to have xCloud integration on Instagram and Facebook Gaming in the form of tags for games that let you have a free trial for games on your device.

And the collection of games is simply much better on Game Pass. Stadia and Luna have to step up their game or they'll be destroyed within 2 or 3 years.

No, those are next year numbers. I did forget 2020 holidays so yeah 115 -120 would be the new figure.

For such high numbers I usually like to go math + comparisons to the highest selling consoles from each region. But first lets compare to your prediction.

As of October 1, 2020 (15 quarters)

Switch-

Japan: 16.1m (43%)

Americas: 26.5m (46%)

Europe: 17.7m (36%)

Other: 7.8m  (30%)

So from your prediction alone, it is lagging pretty far behind in Europe and Others for a chance. But the question becomes, is 15 quarters past its halfway selling point or not? The answer is usually a resounding yes in Japan and America, but less so in Europe. The reason being is late launches in territories not yet launched in, but the Switch has already launched in most territories.

Next lets compare to the top consoles at 15 quarters. Actually I can only find data for 14 quarters, so Switch gets a handicap. And for others I cannot find good data.

Japan (DS, 14 quarters): 22.38m (72%)

America (Wii, 14 quarters): 33.4m (79%)

Europe (PS2, 14 quarters): 24.79m (71%)

So while you predict it to surpass the top consoles, it is currently not on pace to do that.

Also as a side note, while I was searching for these numbers, I happened upon a PS2 thread discussing its final sales. All estimates put it between 159.2m and 161.9m. So I was a little high, my bad.

This methodology doesn't make sense when more granular data is available.  Comparing LTD through x quarters vs. final LTD as a % implies a baseline long term trend.  That point is exactly what is being contended when trying to compare legs for an actively selling console vs one that has completed selling. 

IE:

America NDS through 14 quarters - 22.39M

America Wii through 14 quarters - 33.40M (79%)

If that's all the data I present, and ignore the NDS's actual final total, I should conclude that the DS isn't/wasn't on pace for 60M at that time.  But that's silly, because if I looked at more granular data, I could surmise that the NDS seems to be on an upward trend in America later in its life than the Wii's peak, and guess that FY4 and FY5 might continue that trend or stay flat and help keep the NDS in the battle (which turned into a blowout in the NDS's favor).

You're shutting down the nitty gritty of the argument before it even has a chance to begin with this logic.  Plus launch aligning by quarters ignores the problem of the holiday count.

Here's some more broken down data (I can split it in two if it makes it larger and easier to read).  Obviously PS2 is hard to compare 1:1 due to different regional breakdowns.

This makes it more obvious that the Wii is a poor comparison in any region.  Also makes it evident that the NSW is in the middle of a peak year, and its hard to make any conclusions out of the current LTD because its about to spike with a big holiday quarter (possibly record holiday quarter), and ignoring the impact of that by looking at 15 mo LTD is going to miss the actual signal by a longshot.

What would make more sense is to project the rest of the current FY or wait for another quarter (so final FY4 prediction can become more precise - only one quarter to predict instead of the broad range holiday quarter) and then project out from there.  What caused a peak this year?  Factors that drove it, factors that held it back?  Which of those will be present next year and how strong?  From there we can derive potential for FY5 and consider how that might shape up against the NDS sales curves region by region, and then project forward.

But I thought the Switch could only dream of hitting 120m+ and now it's actually within reach of your super pessimistic prediction?

And where does this leave PS4, which apparently Switch could "only dream of" passing in sales?

We need one year to make a better scenario. 

Honestly don’t think NSW can pass 100mil at this point, it’s only at 68mil while PS4 is at 110mil+ and heading for a 140mil-150mil lifetime number. So NSW can’t possible sell more

Clearly Switch will fall off a cliff in 2021 as they have used up all their big games already and who's gonna buy an underpowered system with no multimedia features and hardly any AAA third party support when they could get a PS5 or an Xbox Series and have 4K and ray tracing?

It's over for Nintendo, time to go third party.