I stumbled across Raja's Xitter post about GTC. Even he acknowledges the industry is a little *too* enamored by Nvidia:
There is good reason why I call it NPC.I stumbled across Raja's Xitter post about GTC. Even he acknowledges the industry is a little *too* enamored by Nvidia:
In thinking about this quote some more, I think those "crushed" are basically those who are partnering with Nvidia, whether it be customers who buy their server GPUs like hyperscalers and cloud service providers or anyone who is hitching their business to Nvidia's success because their own greed necessitates it. However, the "crushing" is literally Nvidia dictating the terms of their business arrangement given their current position as the conductor in this AI hypetrain, even though in many instances Nvidia is developing in-house products or solutions which is currently or eventually will directly compete with those being "crushed". This applies to both the hardware and software side. For example, Google/Microsoft/Amazon/Meta trying to develop their own in-house AI chips while still begging and groveling to buy Nvidia GPUs. On the software side, there's a ton of programmers developing open source tools for Nvidia's AI stack but every GTC Jensen announces their own in-house software stack for an emerging application of AI which obsoletes those who are developing it independently.The full quote:
"The annual green tech cult fest is in town. A tell-tale symptom of a tech cult is when the "crushed" take selfies with the "crusher-in-chief" and post it gleefully on all platforms. You will find some of the "crushed" later in the evenings in bars around San Jose, crying over shoulders of competitors...."
Yeah, I can see the commonality. A lot of those in the physical crowd during Apple keynotes were Apple investors or basically fan boys who would've cheered regardless of what Steve said. The finance news outlets never gave a crap about Nvidia until they saw the money roll in, and then they slurped up the Nvidia Kool-Aid even though they barely understand any of it. Jensen literally just rants for 2 hours and everyone who is on-board just nods their head in blissful ignorance while periodically checking the value of their own Nvidia stock holdings and making sure that the number next to the % symbol is green.There is good reason why I call it NPC.
NV just copy Jobs era Apple, JHH is not nearly as original as most think.
So I asked Jensen what he wants the company to stand for in 3-5 years. His answer was around 14 minutes long. The TLDR: I told everyone a decade ago. I told everybody in the keynote, too. - #1 NVIDIA is an infrastructure business, not an AI product company. Why we laid out 4 years of roadmap “like no other technology company ever has.” - #2 NVIDIA is an AI Factory company. Customer has to do GB and VR because customer has to drive revenue. “And nobody says, Guess what? I think I'm going to save my revenues for three years from now.”- #3 NVIDIA is a Foundational company. “We are a foundational company for every industry and every company, and, therefore, we can’t have surprises.” He made a very interesting parallel to TSMC. My takeaway was that he sees some perceive TSMC as a transistor product company versus a foundational company. (Like NVIDIA).Also an interesting funding conversation that I perceived as trust in execution. Essentially, if you had $100B and two years to build an AI factory, who will you go to? Who do you trust? “My question is, what do you choose? Who do you place the PO on right now as we speak, just try and do it.” I couldn’t help but to think of Stargate…. Maybe a message to hyperscalers with ASICs? On what I perceived as SAM: “We are manufacturing intelligence. And back to the first principle, what are we doing? We manufacture intelligence, and what is, which industry is intelligence for? And I just described every single industry.” Like I said, one of the most insightful conversations with Jensen I’ve been a part of.
Well it all began by using a generic name for a vendor specific marketing conference (WWDC>GTC).Yeah, I can see the commonality. A lot of those in the physical crowd during Apple keynotes were Apple investors or basically fan boys who would've cheered regardless of what Steve said. The finance news outlets never gave a crap about Nvidia until they saw the money roll in, and then they slurped up the Nvidia Kool-Aid even though they barely understand any of it. Jensen literally just rants for 2 hours and everyone who is on-board just nods their head in blissful ignorance while periodically checking the value of their own Nvidia stock holdings and making sure that the number next to the % symbol is green.
Like, wtf does "we are manufacturing intelligence" even mean?
As @igor_kavinski said, It should be $449, $50 less than previous generation. But MRSP doesn't matter when you can't get anything at that price.Still think it's gonna be $500 for the 16.
Of course they want to sell It, but It won't be cheap.You have to remember that nVidia doesn't really want to sell the 16 GB version. Only the 8 and the inevitable 12 GB Super.
Yep this is exactly like Jobs era Apple, its the almost the embodiment but much worse, I never seen anything like that. Cooking like that in a tech conference is crazy, the attention must be around Jensen. Also Pat doesn't look happy lol. he's like "wtf is jensen doing"Yeah, I can see the commonality. A lot of those in the physical crowd during Apple keynotes were Apple investors or basically fan boys who would've cheered regardless of what Steve said. The finance news outlets never gave a crap about Nvidia until they saw the money roll in, and then they slurped up the Nvidia Kool-Aid even though they barely understand any of it. Jensen literally just rants for 2 hours and everyone who is on-board just nods their head in blissful ignorance while periodically checking the value of their own Nvidia stock holdings and making sure that the number next to the % symbol is green.
The monitor that will bring the 5090 to its knees.
57" 4k ultra wide.
It's even cheaper than the 5090 ROFL.
The curious case of the premium Alienware OLED gaming monitor actually undercutting the competition on price...
Dual-4K meets Mini-LED all over again.www.pcgamer.com
Intel will be fine if Panther Lake and Nova Lake succeed. Also they got 2 board members that are semi people now.Intel might not even make it past this decade because of idiot board members who don't even hold a degree in Computer science
Of course they want to sell It, but It won't be cheap.
You have to remember that nVidia doesn't really want to sell the 16 GB version. Only the 8 and the inevitable 12 GB Super.
What you wrote doesn't make sense.They want to sell the 8 GB version (and the 12 GB Super). I am thinking the 16 GB version will go away once the 12 GB Super comes out.
What you wrote doesn't make sense.
12GB version has to be sold cheaper than the 16GB one
We know Nvidia is greedy, but they are not stupid.Yes, which is why the 16 GB almost has to be $500. It's a throwaway product for marketing reasons.
I am thinking the 8 GB will be somewhere between $399 and $449 and the 12 GB Super $449.
I question If they will actually release Ti 8GB, that would be a truly useless GPU.
A "throwaway" product that charges $70 to double the ram?Yes, which is why the 16 GB almost has to be $500. It's a throwaway product for marketing reasons.
I am thinking the 8 GB will be somewhere between $399 and $449 and the 12 GB Super $449.
So the MSRPs will look something like this:
5070 $549
5060 Ti 16 GB $499
5060 Ti 8 GB $429
5060 $349
And maybe look like this later...
5070 Super 18 GB $549+
5060 Ti Super 12 GB $449
5060 Super 12 GB $399
5060 Super 8 GB $349
The useless one is actually the 8GB version.The 16 GB one is intentionally meant to be the useless SKU. It's just for PR reasons.
It'll be interesting to see. N44 according to leaks is 1/2 of N48 in pretty much every way. These 5060 Ti specs put it just a hair over half a 5070 Ti, so just looking at the spec sheet you'd expect the 5060 Ti to be the faster card with maybe a little larger margin than the bigger cards.We know Nvidia is greedy, but they are not stupid.
9060XT 16GB should compare to 5060Ti 16GB pretty well and I don't think It will cost more than $399.
P.S.
36CU 5060 Ti 16GB - $449
32-34CU 5060 Super 12GB - $349-379
30CU 5060 8GB - $349
I question If they will actually release Ti 8GB, that would be a truly useless GPU.
It will be faster, the question is by how much. If AMD doesn't care about high TDP(TBP), then It could be only a few % difference in raster.It'll be interesting to see. N44 according to leaks is 1/2 of N48 in pretty much every way. These 5060 Ti specs put it just a hair over half a 5070 Ti, so just looking at the spec sheet you'd expect the 5060 Ti to be the faster card with maybe a little larger margin than the bigger cards.
The useless one is actually the 8GB version.
Seriously, who in their right mind would pay $429 for a 5060Ti 8GB, when the 16GB version costs only $70 more according to you?
I am thinking the 8 GB will be somewhere between $399 and $449 and the 12 GB Super $449.