Hell has frozen over, Fudzilla Both Microsoft and Intel trying to acquire AMD

[ShintaiDK]

Living off selling high end gaming GPUs only is not sustainable. And remember R&D and cost goes up constantly. So higher volume/revenue is needed every time.

Discrete GPU sales are going down fast at a rate of ~40% a year. And while nVidia is still doing ok, its on the pure expense of AMD. And that cant continue.

EUV however can give a bit more life into discrete GPUs. But that wont last.

Unless something revolutionizing happens the path is set.

 

[Idontcare]

Living off selling high end gaming GPUs only is not sustainable. And remember R&D and cost goes up constantly. So higher volume/revenue is needed every time.

Discrete GPU sales are going down fast at a rate of ~40% a year. And while nVidia is still doing ok, its on the pure expense of AMD. And that cant continue.

EUV however can give a bit more life into discrete GPUs. But that wont last.

Unless something revolutionizing happens the path is set.

Ironically enough, AMD was absolutely right. The future was (is) Fusion.

RIP discrete GPU market/suppliers/AIBs.

 

[Dresdenboy]

Ironically enough, AMD was absolutely right. The future was (is) Fusion.

RIP discrete GPU market/suppliers/AIBs.

Yep. One thing is the idea. Another is execution.

 

[Dresdenboy]

2.7x over the average volume and account to 60M$ in total trade value (Sold and bought). Same as in march for example with the Samsung buyout rumour.

Not exactly a buyout frency. I guess techsite readers are easy to fool financially. I wonder if SA and Fudzilla got financial interest in this.

Keep an eye on it next week. But this is money flowing into an - according to many world-influential forum members - doomed company.

 

[Arachnotronic]

Ironically enough, AMD was absolutely right. The future was (is) Fusion.

RIP discrete GPU market/suppliers/AIBs.

I'm not sure that the data really supports this, though.

The low-end of the dGPU market is obviously going away/pretty much gone, but I think the dGPU for gamers and professionals will continue to thrive for a good long while.

What this will probably translate into, though, is higher prices for mid-range and high-end discrete chips to offset the loss of the low-end volumes. Fortunately, this is a market where people have shown that they are willing to pay for performance, and hardware requirements for games are continually rising.

 

[SPBHM]

better than Mcafee and Nokia phones I guess?
the rumor is kind of interesting, and makes sense in many regards...

 

[dark zero]

Living off selling high end gaming GPUs only is not sustainable. And remember R&D and cost goes up constantly. So higher volume/revenue is needed every time.

Discrete GPU sales are going down fast at a rate of ~40% a year. And while nVidia is still doing ok, its on the pure expense of AMD. And that cant continue.

EUV however can give a bit more life into discrete GPUs. But that wont last.

Unless something revolutionizing happens the path is set.

Actually that was a job from MS.. They could made the Pro Edition much heavier and renders any tablet chip like Celeron N or Amd E1 totally useless and make the true minimun the Phenom II and the Nehalem Core chip...

 

[mrmt]

The low-end of the dGPU market is obviously going away/pretty much gone, but I think the dGPU for gamers and professionals will continue to thrive for a good long while.

The fastest growing IHV on Steam is Intel, with just above 20% of the respondents of the Steam Survey using any kind of HD graphics. So while the enthusiasts gamer won't quit their dGPUs for the foreseeable future, I think the market will keep shrinking, even for gamers, especially if Intel continue the trend of putting more iGPU on the die. Cannon lake and Ice lake might be a game changer for the market.

 

[ultimatebob]

Living off selling high end gaming GPUs only is not sustainable. And remember R&D and cost goes up constantly. So higher volume/revenue is needed every time.

Discrete GPU sales are going down fast at a rate of ~40% a year. And while nVidia is still doing ok, its on the pure expense of AMD. And that cant continue.

EUV however can give a bit more life into discrete GPUs. But that wont last.

Unless something revolutionizing happens the path is set.

Yeah... discrete graphics is a shrinking market. That probably explains why AMD is rebranding a lot of their existing cards instead of spending R&D money to develop new ones.

 

[Fjodor2001]

They got less performance, no matter how you twist it. MBA is at least twice as fast in regular workloads.

You can use the Atom one as a somewhat compare. Fastest against slowest.

An MBA from 2013 (same year PS4/XBONE was released) is twice as fast as the PS4/XBONE CPU?

None of the devices in your benchmark charts have 8 cores which PS4/XBONE does.

Also, you need to show benchmarks of sustained MT loads.

 

[mrmt]

Keep an eye on it next week. But this is money flowing into an - according to many world-influential forum members - doomed company.

Do you think AMD is in a sustainable trajectory?

 

[DrMrLordX]

AMD problem isn't the debt load. The debt load would be a problem if servicing this debt were draining the resources if operational results were fine but the financial expenses were hindering the application of the positive cash flows on the operation, but that's not what we find with AMD. The debt service is fairly cheap, especially for the pre-bankruptcy state of the company and there isn't any amortization until 2019.

AMD problem is simply that it's operational results are not those of a viable business, so even if AMD got rid of its debt pile tomorrow the net result would be that... they would have to build another pile in order to offset the losses they would incur with their money losing business.

AMD is losing, what, $100-$300 mil a quarter or something like that? If they had no debt load, they could assume a new pile of debt (I think their current debt load is $5 billion or so) to offset the losses and shore up the failing elements of their business. That would be enough for at least a few quarters of better-than-right-now R&D spending. Assuming they did a good job spending that money, that could really help turn around the situation.

But they can't take on any more debt 'cuz who's gonna lend to them with their current debt load?

In a valuation you have basically three main components, future cash flows discounted at the proper WACC rate, added the assets on or off the balance sheet adjusted according to the quality/liquidity of the asset, subtracted from the liabilities on and off the balance sheet, also adjusted. So it doesn't matter much whether AMD debt is bigger than the market cap of the company.

Right, but it does matter when their future cash flows are negative. It gives them far less flexibility wrt future borrowing to correct their core problems. It also means anyone looking to do a buyout will face the same problem once they take the helm, unless they have access to a sufficiently-large pile o' cash or a better line of credit.

Ironically enough, AMD was absolutely right. The future was (is) Fusion.

RIP discrete GPU market/suppliers/AIBs.

AMD has been right about a lot of things, from a technical viewpoint. And let's face it, it doesn't take a genius to figure out that putting your GPU on an external bus is a limitation (NVlink aside).

AMD's problem is that being right about the future hasn't made them enough money.

Yeah... discrete graphics is a shrinking market. That probably explains why AMD is rebranding a lot of their existing cards instead of spending R&D money to develop new ones.

If you look at what they are rebranding, you'll see that most of the rebrands are aimed at the low-to-mid tier market where margins are lower. They're plowing all their latest GPUs into the $500+ segment where margins are high. Neither AMD nor Nvidia have even bothered to release any cards from their latest generation (9xx for Nvidia, 3xx/Fury for AMD) in the sub-$100 category.

 

[frozentundra123456]

Well, it doesnt matter how "right" you are unless you produce and market a product to back it up, neither of which AMD has done.

For instance they hyped Fusion for years before even producing a fully HSA compatible chip, much less software to utilize it.

Both their cpus and dgpus trail in absolute performance and performance per watt, and their APUs are so bandwidth starved that they are perpetually in limbo of better than most people need but really not good enough for demanding uses.

 

[dark zero]

AMD will die... like Nikola Tesla giving the legacy, enough power to continue without them.... but after that leaving Intel without any improvements to deliver.

 

[mrmt]

Assuming they did a good job spending that money, that could really help turn around the situation. Assuming they did a good job spending that money, that could really help turn around the situation.

This is the important part. Debt would be a non-issue if AMD had operating margins around what Nvidia or even Mediatek gets with their business, but as they don't, that light debt load an issue.

Right, but it does matter when their future cash flows are negative. It gives them far less flexibility wrt future borrowing to correct their core problems. It also means anyone looking to do a buyout will face the same problem once they take the helm, unless they have access to a sufficiently-large pile o' cash or a better line of credit.

No one forecasts negative cash flows ad infinitum when doing a valuation, otherwise the net result would be wealth destruction in the long run, which means it would be better to close the shop and sell all assets today before they lose any more value.

I don't think borrowing would be an issue for a new management team. The market is drowning with liquidity and even a junk debtor like AMD is actually improving its debt profile because of good conditions on the debt market, so whoever were to buy AMD wouldn't have issues financing it.

But think about it, this BoD has spent 3 years trying to turn the company around and failing, and this after betting the farm on the worst architecture of all times and missing every single growth trend on the market for the last decade. Those guys are consistently missing every single financial goal they establish for themselves, especially in the last 3 years. Do you think this clueless BoD can come forward and ask investors for even more money?

AMD has been right about a lot of things, from a technical viewpoint. And let's face it, it doesn't take a genius to figure out that putting your GPU on an external bus is a limitation (NVlink aside).

In busines the only angle that matters to be right is from the commercial viewpoint.

 

[tential]

Do you think AMD is in a sustainable trajectory?

With proper investment into the company it's possible AMD can be competitive again.

I mean, I'm working in a business that literally just crashed into the ground right now and they've gotten millions of dollars now in investments to build it back up after private equity stepped in and changed some things.

So yes, 100% possible to turn the AMD ship around.

 

[Rakehellion]

Apple should buy AMD.

 

[mrmt]

With proper investment into the company it's possible AMD can be competitive again.

I mean, I'm working in a business that literally just crashed into the ground right now and they've gotten millions of dollars now in investments to build it back up after private equity stepped in and changed some things.

So yes, 100% possible to turn the AMD ship around.

So it's not a matter of money, the new management is changing a lot of things compared to the old business, probably because the old business wasn't in a sustainable trajectory, just like AMD.

 

[mrmt]

I'm not sure that the data really supports this, though.

Attach rates aside Nvidia business strategy is supporting this view. We are not seeing Nvidia putting even more money on the dGPU market, but we are seeing them trying to desperately open new markets for their technology.

 

[MiddleOfTheRoad]

Consoles most likely only went x86 because 64bit ARM wasnt ready at the time. Going x86 is against what both MS and Sony wants. A licensed design they control completely and can fab whenever and whereever they want.

That has to be the dumbest thing I've ever read.

Software developers were begging Microsoft and Sony (and now probably Nintendo) to go to x86. The software development tools for ARM are insanely primitive.... And the majority of console buyers wouldn't be satisfied with tarted up mobile games.

The cost of the x86 license is peanuts... MS and Sony don't honestly care.

 

[shady28]

I'm not sure that the data really supports this, though.

The low-end of the dGPU market is obviously going away/pretty much gone, but I think the dGPU for gamers and professionals will continue to thrive for a good long while.

What this will probably translate into, though, is higher prices for mid-range and high-end discrete chips to offset the loss of the low-end volumes. Fortunately, this is a market where people have shown that they are willing to pay for performance, and hardware requirements for games are continually rising.

That's a dead-end game. The money needed to design and construct those high-end GPUs is a product of economies of scale. At its heart a 980 Ti GPU is essentially still a Maxwell design, which was intro'd with the 970/980/750/750Ti.

What you're saying is like saying Cadillac would be just fine if GMC / Chevy / Buick went away, not realizing that a Cadillac engine is nothing more than a tweaked Chevy engine.

For perspective, if you total up the Steam HW survey quantity of 970s, 960s, 750, and 750Ti - it comes up to 9.07%.

The 980 by comparison has 0.99%. The 980 Ti and Titan don't even rank, meaning below 0.3%.

If we start looking at mobile dGPUs as well - the 840M is the highest with 1.3%. That means there are at least 4x more 840Ms out there than 980 Ti's. *at least*

So basically you are talking about 98% of Nvidia's unit sales in the 970 and below category + mobile. Without those sales, Nvidia would have to stop R&D.

But the iGPU is a long ways off from matching up against ~$100 dGPUs. A GTX 740 or 750 will blow them out of the water. That'll change but we're not there yet. The only time I see favorable comparisons for iGPUs is with odd graphics settings vs some crap dGPU like an r7 240 DDR3.

 

[Phynaz]

Ironically enough, AMD was absolutely right. The future was (is) Fusion.

RIP discrete GPU market/suppliers/AIBs.

But, did AMD create this future?

 

[ultimatebob]

But, did AMD create this future?

AMD might have started it, but Intel and Apple caught up quick.

 

[Ken g6]

Ironically enough, AMD was absolutely right. The future was (is) Fusion.

RIP discrete GPU market/suppliers/AIBs.

I'd say AMD was half-right. They thought the future was hot fusion (125W TDP). The future is actually cold fusion (4.5W TDP).

 

[Tsavo]

GE is going to buy AMD.

They want to expand into the Space Heater / Toaster Oven / Hot Water Heater Internet of Things, and AMD's products are perfect for this.

I cannot wait to toast my Pop Tarts at 4GHz.