Refresh

Nvidia’s autonomous automobile chip would get annihilated in NYC site visitors, I can assure that, although it would work in numerous different cities.

Nvidia RTX 5090, RTX 5080, and RTX 5070 Ti laptops will probably be accessible beginning in March, whereas the RTX 5070 laptops will probably be accessible beginning in April.

(Picture credit score: Nvidia)

OK, we simply received phrase on the Blackwell GPU availability.

The RTX 5090 and RTX 5080 will go on sale on January 30, 2025, for $1,999 and $999, respectively.

The RTX 5070 Ti and RTX 5070 will probably be accessible in February for $749 and $549 respectively. UK and Australia pricing wasn’t given, however we have reached out to Nvidia for clarification.

The fear with artificial information is that you will find yourself with a Habsburg AI, one which’s successfully inbred by itself information to the purpose the place it turns into a ineffective abomination. Google the Habsburg monarchs of Europe if you wish to see why that is such an apt description of the issue.

OK, so the tip of knowledge for coaching fashions is one other main bottleneck for AI, and what Jensen is speaking about right here with Cosmos is producing new information that subsequent fashions could be educated on (artificial information), since these fashions have already consumed all the prevailing information it could possibly be educated on.

Nevertheless.

I’m wondering how Cosmos will keep away from mannequin collapse.

OK, sorry about that people, we had been coping with some technical difficulties, however we’re again to the motion, and that motion is all about tokens. It is tokens all the best way down.

OK, the digital human factor remains to be giving main uncanny valley, however it’s much less extreme than it was once.

Coding assistants are the demise knell for the junior software program developer. A lot for ‘study to code’.

Oh man, I simply had a darkish thought. Are you able to think about coaching the AI agent that is taking your job? That is grim.

I need to say, I discover all of this AI dialogue attention-grabbing from an educational perspective, however I believe there’s numerous expectations for these information facilities, and nobody is mentioning that the ability necessities for these are just about going to place a cap on what they’ll do, since we solely have a lot electrical energy accessible on a grid at anybody time.

So we’ve got absolutely entered into the information heart section of the keynote. Whereas GeForce graphics playing cards received not less than a bit extra time and a spotlight than Lovelace received, it is clear that these playing cards aren’t as essential to Nvidia as the information heart enterprise.

And yeah, that protect bit was a bit…nicely, it was one thing.

Nvidia GeForce RTX 5000 collection cellular GPUs are additionally coming, with the RTX 5070 cellular that includes RTX 4090 efficiency, although I am guessing Jensen means RTX 4090 cellular efficiency.

Additionally, the RTX 5000 collection will probably be accessible beginning in January, although we do not know which will probably be coming first.

If the shader cores may also carry the burden of AI workloads, as Jensen said, then we’re getting manner higher DLSS on these playing cards.

OK, so little or no on specs, however I need to find out about this AI administration processor.

(Picture credit score: Nvidia)

OK, RTX 5090 beginning at $1,999. RTX 5080 for $999. RTX 5070 Ti for $749. Sure, completely. That is what I need to see.

Do not get me mistaken, these are nonetheless costly graphics playing cards, however given the fears of a $1600 RTX 5080, it is a very nice shock.

OK, RTX 5070 at $549, RTX 4090 efficiency. Whoa.

OK, so Jensen is holding the RTX 5090.

(Picture credit score: Nvidia)

Nvidia RTX Blackwell is official, and that may be a very fairly wanting graphics card.

That was a reasonably spectacular demo.

OK, the primary point out of GeForce, so right here we go.

That was a really brief Virtua Fighter demo.

(Picture credit score: Nvidia)

LOL, Jensen’s jacket is bedazzled.

Tokens, tokens, tokens. It is no shock that we’re simply leaping proper into AI, however yeah, it is exceptional how a lot Nvidia has reworked nearly in a single day.

OK, NOW we’re kicking issues off.

OK, the Nvidia section of the keynote is about to start, however it certain is taking some time. Nvidia is generally faster to launch than this.

OK, I unironically love ‘By no means gonna offer you up.’ I used to rollerskate to that track as a child.

(Picture credit score: Nvidia)

CTA President Gary Shapiro is introducing Jensen Huang.

OK, I DO need an exoskeleton. These issues look cool as hell.

OK, right here we go.

The largest factor I am anticipating tonight is Nvidia’s GeForce RTX 5090, RTX 5080, and presumably the RTX 5070 Ti.

Following up Nvidia’s Lovelace GPUs, the Blackwell-based RTX 5000 collection is anticipated to be considerably extra highly effective, with the rumor mill placing the RTX 5080 round 10% quicker than the RTX 4090, at present the perfect graphics card on the patron market.

That, in fact, would put the RTX 5090 in a category totally by itself, and there is not any telling the place its efficiency will finally find yourself. That stated, if hypothesis is on the mark, it ought to characteristic 32GB GDDR7 VRAM with a reminiscence bandwidth of 1.52TB/s on a 512-bit reminiscence bus, making it really the world’s first 8K gaming graphics card.

OK of us, we’re arising on the quarter-hour from the beginning of Nvidia’s CES 2025 keynote, the place CEO Jensen Huang will take the state at Madalay Bay’s Michelob enviornment. We’re anticipating some main information tonight, so for many who’ve been ready to listen to about Nvidia’s next-gen client graphics playing cards, you do not for much longer to attend.

And for those who actually need to hear all about information heart AI and Omniverse stuff, I am certain Jensen will get round to these as nicely.

Newest

AMD CES 2025 Keynote reside weblog: because it occurred

See extra newest ►