• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

*** The AMD RDNA 4 Rumour Mill ***

The bad news for PC gamers never stops does it?
AMD's decline to just 6% marketshare is a big concern but they only have themselves to blame though, they delayed the launch and gave Nvidia a head start and even then despite having a compelling product and winning the reviews they couldn't stick to the prices they advertised.

Jack Huynh should be getting fired after presiding over a halving of AMDs marketshare with his strategy despite targeting 40%.

 
Last edited:
Problem is for nearly all of this year, people have been calling real data fake news, and instead clinging onto incorrect statements from AMD which they the backtracked, and some random small surveys on the internet.


Anyone remember the 200K claim AMD made, which they then stated they never said after it was splashed everywhere.

Update: AMD reached out to Tom's Hardware to clarify that no claims about sales numbers were given at the event. There was discussion about strong demand for RX 9000 (and X3D), but the 200K claim was not a part of that conversation, according to AMD. The Benchlife page which seems to have been the original source of the claim has also been pulled.

Also what about Frank Azor's now proven lie. Replenishing MSRP stock apparently back in March, in the coming days and weeks.

 
Last edited:

Huynh was less straightforward when discussing AMD’s GPU market share on the desktop. The market-share figures from long-time GPU watcher Jon Peddie indicate that AMD’s desktop share has fallen from 12 percent to 6 percent in a year, leaving AMD with just crumbs to Nvidia’s utter dominance. (Intel’s GPU share is still non-existent.)

Huynh’s position reiterates what AMD has been telling Wall Street: “I’m happy that we’re selling everything that we make,” Huynh said.

“The reception for the [Radeon] 9070 XT has been phenomenal,” Huynh added. “I can say that we’re still sold out.”

Of course, whatever AMD’s share of the desktop GPU market is, it’s softened by the fact that AMD’s gaming revenue climbed 73 percent year-over year, though assisted by strong sales of SoC chips for consoles. AMD can also route its resources to the data-center market instead, whose share of the company’s revenues continue to climb. Either way, AMD is making more money than it has ever made before, achieving a record $7.7 billion in revenue for the second quarter of 2025.
 
Last edited:
AI will make the energy use of crypto mining look like a toaster!

Open AI alone has recently signed a deal to build $1.3 trillion in new data centres, about 60% of this cost is the GPU's itself. To power these data centres though may require the construction of 23 new nuclear power plants, but the US has only built 1 in the last 30 years

And that's just one company, it excludes Microsoft, meta and the dozen others doing similar things
 
Last edited:
Open AI alone has recently signed a deal to build $1.3 trillion in new data centres, about 60% of this cost is the GPU's itself. To power these data centres though may require the construction of 23 new nuclear power plants, but the US has only built 1 in the last 30 years

23?? How can they think this is feasible? And who is going to pay for this?

Save your money, when this bubble will pop there will be economic hell.

The way the economy/currency is heading, I think storing toilet paper will have more economic sense. (Though bullion is preferable. :| )
 
Overinflated yes, but not sure it's a bubble. The dot com bubble was because investors were soo tech illiterate they didn't realise half the companies were just a website that any monkey can put up for 20 dollars.

AI definitely has some tangible products under it. We use Github Copilot at work, and it's frankly amazing at what it can do. I can't imagine how good it will be in another 10 years.
 
Overinflated yes, but not sure it's a bubble. The dot com bubble was because investors were soo tech illiterate they didn't realise half the companies were just a website that any monkey can put up for 20 dollars.

AI definitely has some tangible products under it. We use Github Copilot at work, and it's frankly amazing at what it can do. I can't imagine how good it will be in another 10 years.
Same, been a programmer for over 25 years and using Ai as a tool to help with methods, models etc is incredible. especially when I have to context switch between languages and projects so often
 
Same, been a programmer for over 25 years and using Ai as a tool to help with methods, models etc is incredible. especially when I have to context switch between languages and projects so often
Yup. I have zero React/TS experience, but managed to add a whole load of new functionality in it, just getting copilot to do all the work from detailed prompts.
And it never gets annoyed, tired, grumpy, is available 24/7. Soo much better than even the most knowledgeable team-mate.
 
Yup. I have zero React/TS experience, but managed to add a whole load of new functionality in it, just getting copilot to do all the work from detailed prompts.
And it never gets annoyed, tired, grumpy, is available 24/7. Soo much better than even the most knowledgeable team-mate.
Yeah, I hope you address it with good morning chat to stay on it's good side. Same here with react and ts. Invaluable tool
 
Last edited:
Overinflated yes, but not sure it's a bubble. The dot com bubble was because investors were soo tech illiterate they didn't realise half the companies were just a website that any monkey can put up for 20 dollars.

AI definitely has some tangible products under it. We use Github Copilot at work, and it's frankly amazing at what it can do. I can't imagine how good it will be in another 10 years.

Problem with AI is no one is making money from it

Microsoft for example is estimated to have spent about $20 billion a quarter on its AI build out over the last few quarters, but is pulling in $240 million in co pilot revenue. It's not unique to MS either, Open AI and others have the same problem- to make an ROI they'll probably need to more than quadruple their subscription fees
 
Last edited:
Allow me to reiterate (and then I'll stop OT). AI is a valid tool but it's not likely going to make the untold trillions companies are hoping.

Why?
1) LLM peaked, GPT5 proved that we reached diminishing return land on LLM size
2) Small (<=8b) models that can run on consumer hardware using web search or focusing on a single task are proving competitive with the big ones in many scenarios

This means that there is no real way AI companies can increase revenues much more than now with the current paradigm and MCP will just become another way to do APIs.

The only possible catch is if the so-called "World Models" will shift the paradigm to physical interaction with the world, however nobody yet released one so that's the card that will either save or sink the current AI wave.
Just like dot-com, LLMs are here to stay, however their value right now is definitely overinflated.
 
Back
Top Bottom