This Is The Real Reason Nvidia Abandoned PC Gamers

A few weeks ago now, I attended Comp Computex, one of the largest technology trade shows, and I've seen some pretty boring shows over my past 20 years or so, but this one was possibly the most

boring, which will probably surprise exactly no one. The climate simply wasn't right for a fun show. DRAM

pricing still cooked. Dan pricing is equally cooked, and as a result, GPU pricing is pretty horrible. It's not

exactly a fun time to be a PC gamer right now, especially if you need a new gaming PC or a platform upgrade. Now is

simply not the time for any of that. The

question everyone around here seems to be asking is when will the right time be? When will this nightmare finally be

be? When will this nightmare finally be over? It's a pretty interesting question

over? It's a pretty interesting question that no one can really answer right now, but there does seem to be two distinct opinions about how all of this is going to play out.

There's the you'll own nothing and be happy crowd. And then there are those of

happy crowd. And then there are those of us who are just hoping to ride this out and get back to something that resembles normality. Though to be clear, we don't

normality. Though to be clear, we don't think we will ever be going back to the way things were say 5 years or so ago.

And that is exactly what I want to explore in this video. But before we get into it, today's sponsor spot is brought to you by Thermal Grizzly and the new Juronaut highperformance thermal paste.

Duronaut offers extreme long-term stability combined with outstanding thermal conductivity. And this has been

thermal conductivity. And this has been achieved through a specially engineered silicon oil combined with aluminium micro particles and zinc oxide nano particles of various shapes. Duronaut

isn't electrically conductive and won't harden over time. For best performance, you'll want to apply a very thin layer.

And to help with this, included is the TG Spatula Pro. The spatula's revised design allows for increased pressure during application, making it easy to spread a thin, even layer of Duronaut.

So, for more information, please check the link in the video description.

Okay, so by normal, I mean things like DDR5 memory or just DRAM in general.

That stuff certainly will be cheap once again, as will SSDs. In fact, with production ramping up and increased competition from China, we could run into a situation where DRM and nan

pricing completely collapses once it becomes clear that the AI data centers aren't actually going to be built in time. And I suspect that that day is

time. And I suspect that that day is looming sooner than some realize. But

maybe I'm just coping. Still, you do have to stop and question reality sometimes. And I found myself doing this

sometimes. And I found myself doing this many times while I was at Computex because much of what is going on right now, it simply doesn't make sense. And a

recent report surfaced revealing that Nvidia has sold 5.4 billion worth of GPUs to a Shell company called Valor, which it's using to transfer GB200 GPUs

to Elon Musk's AI company, XAI. And

they're doing this because they're under immense pressure from Wall Street to continuously post Blockbuster revenue numbers. Now, because Valor is

numbers. Now, because Valor is technically an independent entity, Nvidia can book that 5.4 billion as immediate complete sale revenue on its

financial statements, moving the physical inventory entirely off its books. Now, to make this deal happen,

books. Now, to make this deal happen, Nvidia didn't just sell the chips, they helped fund the buy. Nvidia invested

roughly 1.9 billion of its own cash into Valor as a limited partner. This creates

a sort of financial flywheel. Nvidia

injects equity into the Shell company and then uses that cash to buy its own chips. Nvidia gets its cash back along

chips. Nvidia gets its cash back along with the massive record revenue all while securing XAI as a long-term lockedin partner. This sort of scheme

lockedin partner. This sort of scheme effectively distributes the immense financial risk of unproven AI infrastructure away from the primary tech companies. Now, in my opinion, a

tech companies. Now, in my opinion, a healthy company or at the very least a company that truly believes in AI like Nvidia claims, they wouldn't be doing any of this. Instead, this is the sort

of thing you would do if you know you're propping up BS. So, you're just trying to make as much money as you possibly can before it all falls apart. So, if

what I'm saying here is indeed the case, why has Nvidia done this? The obvious

and correct answer is money. But is it just pure greed or has all of this been born out of desperation? I tend to believe it's more the latter. And I'm

going to explain why. For years now, it has become increasingly obvious that achieving the kinds of performance gains that Nvidia were once able to is no longer possible. Back in the early

longer possible. Back in the early 2000s, Nvidia were able to push out a new GeForce generation every 12 months or so. And these updates almost always

or so. And these updates almost always delivered big performance gains. As a

gamer, it was basically impossible to keep up. Now, as exciting as this period

keep up. Now, as exciting as this period was for gamers, it wasn't the best of times for Nvidia. In the decade between 2000 and 2010, Nvidia was a fraction of

the size it is today. Instead of fueling an artificial intelligence gold rush, the company was fighting a fierce war for the desktop PC market, primarily selling graphics cards to PC gamers and

tech enthusiasts. During this era,

tech enthusiasts. During this era, Nvidia's annual revenue grew from under a billion dollars to hovering around 3 to four billion. At the time, this sort of growth would have been really

exciting for Nvidia. They were a major player in an emerging market, and the constant roll out of new products meant gamers were always scrambling to hand over cash for the latest gear. And the

good times would continue for about another decade. But as 2020 approached,

another decade. But as 2020 approached, Nvidia faced an issue, and they knew it before we did. Despite the fact that the gaming industry was as big as it had ever been and it was continuing to grow,

Nvidia also dominated the market. They

had the majority market share. So all of that sounds really good. The problem

though was the product. Making it better had become a serious challenge and an expensive one at that. And as I said in the early days, you could expect a new GeForce generation about every 12

months. And the gains were massive. In

months. And the gains were massive. In

fact, it was pretty common to expect around a 50% genon gain right up to the GeForce 10 series of Pascal. The GTX

1070, for example, was around 60% faster than the GTX 970. Now, although gamers didn't realize it at the time, the GeForce RTX 20 series was actually the

first sign of trouble for Nvidia.

Now, let's start by looking at 10 generations of flagship Nvidia GPUs, ignoring the dual GPU models and the Titan class products. The GTX 285

offered around a 60% performance increase from the previous generation flagship, which would have been the 9800 GTX Plus. It cost just $360 US, and was

GTX Plus. It cost just $360 US, and was rated at a TDP of just 215 W.

Then the following year, Nvidia initially fumbled with the GTX 400 series, but quickly replaced it with the much better 500 series, giving us the

GTX 580, which was another 200 W part, but this time it delivered a massive 70% performance uplift. After that, we had

performance uplift. After that, we had to wait a year and a half for the GTX 680. And although it was only 40% faster

680. And although it was only 40% faster than the 580, it was still one of the most technological significant upgrades Nvidia ever managed to pull off. Moving

from a matured 40nm process down to the cutting edge at the time, 28 nm process, allowing Nvidia to shift their entire engineering focus from raw computing power at all costs to extreme

performance per watt efficiency. And as

a result, the GTX 680 dropped down to just 170 W. Then the following year in 2013, the GeForce 700 series arrived,

and at the helm was the GTX 780 Ti, costing a cool $700 US. That's 40% more expensive than the 680, but it was also

around 60% faster.

In fact, the $330 US GTX 770 was able to match or even slightly beat the 680. The

770 was essentially a tuned 680 offered at a very heavy discount. A few years after that, the GeForce 900 series

arrived, and the 980 Ti offered another 40% increase over the 780 Ti. It also

packed twice as much VRAM, jumping up from 3 GB to 6 GB. And that brings us to most people's favorite GeForce generation, including Nvidia's own CEO,

Jensen Wong. Looking back, Jensen

Jensen Wong. Looking back, Jensen probably views the Pascal architecture in the GeForce 10 series so fondly as it largely sold itself. It was something he could genuinely be proud of, something

he could enjoy like the rest of us without having to spin endless BS just to create some interest. Anyway, for me personally, this generation was the last

time things were genuinely exciting. The

GTX 1080 Ti was a massive 70% faster than the 980 Ti. And this series also offered more VRAM and it was generally more power efficient. Power efficiency

or rather just raw power usage is something we should keep in mind as we move on. The RTX 2080 Ti was released

move on. The RTX 2080 Ti was released just one year after the 1080 Ti, but it only offered a 30% performance increase, the weakest generational uplift we'd

seen from Nvidia in a very long time.

The claimed TDP remained at 250 W and that's despite moving from TSMC's 16 nm process for the previous generation to 12 nmters for the first RTX generation.

Now, the reason for the lack of efficiency improvements here is down to the fact the die size exploded by 60%.

Making the 2080 Ti a much larger product. So, any power savings achieved

product. So, any power savings achieved with the newer node was spent on the much larger die. But at the end of the day, for consumers, it was the poor

value that really hurt this generation.

The RTX 2080 Ti cost 40% more than the 1080 Ti, and that meant for the first time, the price increase was larger than the performance increase. And Nvidia

tried to sweep this issue under the rug with the promise of technology such as real-time ray tracing and upscaling. And

we know how well that played out for the GeForce 20 series. not nearly as well as advertised. Let's just say that.

advertised. Let's just say that.

Resulting, in my opinion, in Nvidia's most disappointing GeForce generation since the initial Fermy release. Looking

at this graph, it appears that Nvidia managed to recover with the Amper RTX30 series, as the 3090 Ti offered a 50% increase from the 2080 Ti. Not too bad

after 2 years. That said, if you'll look a little bit closer, the cracks do start to appear. For example, the 3090 Ti

to appear. For example, the 3090 Ti bumped up power usage by about 80% over the 2080 Ti. So, in other words, Nvidia was very much brute forcing that 50%

performance increase. Not only that, but

performance increase. Not only that, but the MSRP also doubled from $1,000 US to $2,000 US. Then, 2 years later in 2022,

$2,000 US. Then, 2 years later in 2022, Nvidia did a bit better with the RTX 4090. It offered another 60% increase

4090. It offered another 60% increase while slightly reducing power usage. And

it even cost a little bit less at $1,600 US. After that, it would take Nvidia a

US. After that, it would take Nvidia a further 3 years to get out the RTX5090, which was just 30% faster, but came with

a staggering TDP of 575 W and consumed 590 W in our testing.

Still, as I said, this picture doesn't look too bad for Nvidia, and that's because the flagship GPUs really do mask what's going on here. But before we move on, here's a look at the MSRP pricing

for these flagship GPUs adjusted for current levels of inflation. Meaning the

$360 US GTX 285 would have cost $558 in today's money.

What we can clearly see here is that in the early years, gamers were regularly receiving big performance gains at fairly stable and predictable pricing.

But as we move from left to right on our graph, we see the opposite starts to happen. The big gains become less

happen. The big gains become less common, but the big prices, they settle in. Now, if we take a look at GPUs

in. Now, if we take a look at GPUs priced around $500 US, or as close to that as we can get when adjusting for inflation, we find a very different

picture. the GTX 570 that was 55% faster

picture. the GTX 570 that was 55% faster than the previous generation flagship while costing slightly less. So that's a great genen uplift that only took a year

to achieve. The GTX 670 was a little

to achieve. The GTX 670 was a little less impressive, but still offered a solid 45% uplift after about 18 months.

The GTX 770 was again a bit of a letdown, only roughly matching the previous generation flagship, making it just 20% faster than the GTX 670. though

it was almost 20% cheaper when adjusting for inflation. So, still a pretty decent

for inflation. So, still a pretty decent improvement there. The GTX 970 was a

improvement there. The GTX 970 was a gem, ignoring that small issue with the partitioned VRAMm. It was 40% faster

partitioned VRAMm. It was 40% faster than the 770 for the same money. Then we

arrive at Pascal. Once again, an amazing uplift here. The GTX 1070 was 60% faster

uplift here. The GTX 1070 was 60% faster than the 970 while costing just 14% more. again when adjusting for

more. again when adjusting for inflation.

Then 2 years later, we arrive in 2019 and are presented with the worst mid-range offering Nvidia has ever presented, the RTX 2060, which thanks to

a price increase, lands at very near $500 US when adjusting for inflation, resulting in a cost of $465

US for a 10% increase over the GTX 1070.

In fact, 10%'s generous as it was more like 6%. And upon release, it didn't

like 6%. And upon release, it didn't even have any ray tracing titles to hang its hat on. The RTX 3060Ti was a bright spot and it arrived just a year later.

Though the RTX30 series was largely spoiled by the cryptocurrency mining, but that's not really relevant for what we're looking at here. The RTX 3060 Ti was 60% faster than the 2060 and

arguably too good as it really killed Nvidia's next generation. The RTX 4060 Ti was a disgrace. The 16 GB model came

in at an MSRP of $500 US and once we adjust for inflation in today's money, it's more like $546.

And for that, you got a mere 5% performance increase. Then most

performance increase. Then most recently, we got the RTX 5070, which is meant to cost $550 US, and at least at one point it did. But when compared to

the RTX 4060Ti, you're getting a 50% performance uplift. There's a clear

performance uplift. There's a clear pattern emerging here, indicating that it is indeed becoming increasingly difficult for Nvidia to regularly offer noteworthy performance gains. Not only

is the window between product releases increasing, but the window between notable performance gains has become even bigger. Now, for those of you who

even bigger. Now, for those of you who wish to claim I'm cherrypicking the data to make Nvidia look bad or something crazy like that, here is a simple MSRP

comparison. Again, targeting the $500 US

comparison. Again, targeting the $500 US price point. As you can see, despite

price point. As you can see, despite changing some of the GPUs, a very similar pattern emerges. In fact, this is a little more favorable for the older

generation GPUs. This line graph shows

generation GPUs. This line graph shows the pattern. From 2010 to 2017, you were

the pattern. From 2010 to 2017, you were getting fairly frequent GPU releases at similar prices offering similar strong performance gains with a small dip in

the middle, but nothing overly concerning. For example, if you bought a

concerning. For example, if you bought a GTX 680, you would have had to wait just a few years to receive around a 60% performance increase with the GTX 980

without having to spend any more money than you did originally. Then 2 years after that, the GTX 1080 was offering an additional 60% performance once again

for about the same money. But beyond

that, we're seeing massive generational dips in performance, meaning gamers would have to wait much longer for worthwhile gains at similar price points. And this here is my favorite

points. And this here is my favorite graph of all for demonstrating the issue that Nvidia and their customers have run

into. From 2010 to 2015, Nvidia's $500

into. From 2010 to 2015, Nvidia's $500 offerings improved by a whopping 140%. See, when going from the GTX 580

140%. See, when going from the GTX 580 to the GTX 980, this is an entirely different gaming experience, unlocking a new level of graphical fidelity in just

a 5-year period. But they did it again, in fact more so, when going from 2015 to 2020. And while this jump from the GTX

2020. And while this jump from the GTX 980 to the RTX 3070 does make the 3070 look very impressive, as we just saw, a large portion of these gains were actually provided by the Pascal

generation.

Also, please note if we were to avoid VRAM limits, as the GTX 980 only had 4 GB of VRAM, if you can believe that, the performance uplift would probably be

more like 150%. Still, that's a sizable uplift, but given we're going from 4 GB to 8 GB in this comparison, the uplift isn't just a result of increased compute

power. But whichever way you want to

power. But whichever way you want to slice it, the result is going to be much the same. Nvidia was more than doubling

the same. Nvidia was more than doubling performance in this period, which then brings us to 2020 to 2025.

If this doesn't clearly illustrate for you the treacherous waters in which Nvidia is heading into, I don't know what will. just a 50% increase in

what will. just a 50% increase in performance from 2020's RTX 3070 to 2025's RTX 5070. That is a shockingly

poor result. And what this tells me is

poor result. And what this tells me is even if the AI bubble does pop, though I believe at this point it's more of a matter of when it pops. But regardless,

we're never going back to the way things were because we can't. Those days are over and there's nothing anyone can do about it. And that includes Nvidia. If

about it. And that includes Nvidia. If

anything, the situation is likely only going to get worse for both gamers and Nvidia. Nvidia relies on new process

Nvidia. Nvidia relies on new process nodes to improve their products. So, if

new nodes take longer to produce and cost more, we're going to see longer development cycles that will result in more expensive products, making it very difficult for Nvidia to offer gamers the

kinds of improvements that they're used to. So whereas if we go back 10 years,

to. So whereas if we go back 10 years, you could realistically upgrade your graphics card every 12 months, certainly every 18 to 24 months. However, in the

last 5 years, noteworthy upgrades have been pushed out to more like 5 years.

And in another 5 years, we could be looking at maybe 7 to 8 years for upgrade cycles for gamers. So, there's

the time component that'll see Nvidia having to entice gamers to buy their products long after they've freshly come out of the oven. And then there's the economics of semiconductor manufacturing

that they also have to contend with. For

nearly two decades prior to 2019, wafer prices remained relatively flat or they even declined as the manufacturing processes matured. However, the

processes matured. However, the introduction of EUV or extreme ultraviolet lithography coupled with rising supply chain complexities and then TSMC's massive market pricing power

that has caused cuttingedge wafer costs to explode by roughly 400% over the last decade. Nvidia weren't willing to

decade. Nvidia weren't willing to sacrifice their margins and in fact they wanted to increase them. So this meant gamers just got less. It's why we've

gone from a 256-bit wide memory bus with a GTX 1070 and 8 GB of VRAM in 2016 to an RTX 5070 with a 192bit wide memory

bus and just 12 GB of VRAM after nearly 10 years. Of course, Nvidia could reduce

10 years. Of course, Nvidia could reduce their margins, but given the technology slowdown that I've illustrated in this video, even that's not really going to be a viable strategy. At best, it would

just buy them a small amount of time.

And that's why Nvidia has started leaning really heavily on software type features like DLSS upscaling and frame generation as they don't require making

the die any bigger. And they can in a way fake performance gains over previous generation hardware. So that all

generation hardware. So that all explains the GeForce side of things. But

for Nvidia, selling directly to customers, it's not nearly as desirable as businessto business transactions.

They always knew that server and professional markets would pay more for their hardware when compared to gamers.

They just needed a way to exploit that market and AI was therein. How that's

all going to play out is yet to be seen.

But whatever the case, I believe Nvidia knows coming back to PC gaming isn't going to be a viable long-term strategy for their business given the technological challenges. They no longer

technological challenges. They no longer view it as profitable enough. Of course,

there will be future GeForce GPUs, and I'm sure many of them will be really impressive. Hell, the GeForce 60 series

impressive. Hell, the GeForce 60 series could be amazing. But the point is, or rather the problem for Nvidia is they can't pump out a new generation

regularly enough anymore, and offering games that will entice gamers to upgrade every generation has become far less common. Still, PC gaming is a

common. Still, PC gaming is a multi-billion dollar industry. So, if

not Nvidia, someone else will eventually fill the void given a long enough timeline. Of course, Nvidia will

timeline. Of course, Nvidia will continue to service the gaming market.

But when opportunities like cryptocurrency mining and AI data centers come knocking, they're going to capitalize just as they have. And when

those things inevitably crash, they'll once again become a gaming company. So,

I guess see you again soon, Nvidia. And

hopefully you can find legitimate reasons for gamers to upgrade more than once a decade. So fingers crossed that ends up being the case. And that is going to do it for this video. If you

liked it, give it a like. Subscribe for

more content. And if you'd like to become a Harbor and Box member, we have the join button or Patreon. Signing up

to either one of those things will give you access to our exclusive Discord server, monthly live streams, Q&A stuff, and behind the scenes content. So check

that out if you're interested. If not,

that's perfectly fine. And I would like to thank you for watching this video.

I'm your host, Steve. See you again next time.