Microsoft and Sony now pay between $70 and $75 per chip down from a little over $100. At $100 the margins were $15. You do the math. AMD is losing money on console chips.
"It suggests that given Microsoft is paying AMD $100 / Xbox chip (more likely ~$70-$75 currently according to Mercury Research) and that Microsoft has built 12.6M Xbox One consoles to date (i.e.$1.26B in sales to AMD so far); hence a deal would save Microsoft billions (amusingly forgetting that AMD does not produce chipsets out of thin air for free).A “costs saved” thesis, we believe,holds no water. Even if one were to assume that Microsoft would be willing to take on all of AMD’s assorted baggage, what they would actually save in such an event is the gross profit they are paying to AMD on the chip purchases. In fact, AMD’s gross margins on their console chips arelow (we believe in the low 20′s), hence (on perhaps ~$1B or less in purchases per year) Microsoft might save, AT MOST, $150M-$200M through such an action. And, we note that the current purchase agreement has cost-down already built in, suggesting that the price Microsoft is paying today will continue to go down in the futureat the production costs of chips go down."
Microsoft and Sony have both had working Intel-nVidia prototypes for several quarters now.
No lock in there. After all AMD simply delivered Intel invented architecture for which it is a mere second source.
The Intel/GeForce based units run much quieter, cooler and much, much faster. As soon as AMD's demise becomes public knowledge, the console makers will announce their refreshed consoles featuring huge improvements in performance and livability.
They will cost more, though.
Lots of good things.
Oh, like warning that its already feeble Q2 guidance would be more than twice as bad as promised.
Continued cash burn.
Massive margin erosion accelerating rather than reversing as originally promised.
Record low market share in all segments with further losses in Q2.
Hawking 28nm chips when the competition has 14nm in volume.
Being run by a third rate CEO that was runner up to the second rate CEO it took you 6 months to find because no one qualified would tarnish their reputation by taking the job.
So many great things going on at AMD.
Most analysts now project losses for AMD through all of 2016.
So much for all that turnaround smoke Lisa Su was blowing up fanboys' rears.
At this rate, AMD won't last through 2016.
Looks like the waddler knew something you pumpwads didn't.
AMD halts trading and announces the mother of all misses.
This is bad enough to be an event of default under its big loans.
Suck it down fanboys.
Suck. It. Down.
Actually Fury cards are available is you are willing to buy a bundle - AMD APU plus motherboard plus other junk. It seems like AMD is desperate to move unwanted APUs on the back of slightly less unwanted new GPUs and isn't worried about tying one to the other to try and cure inventory issues.
At a minimum reliance damages. The investors who relied on AMD's fraudulent Llano claims would not have bought but for those lies. Their out of pocket loss is the amount they paid for the AMD stock. They have a duty to mitigate, which would mean selling once they realized they had been lied to, so AMD would be liable for the difference between what the investor paid based on AMD's lies, less what the received when they sold their shares after discovering the lies, times the number of shares each injured investor bought in reliance on the lies.
I figure the award if it goes to trail could be upward of a billion dollars. Settlement value is about half that before trial.
Poor-ish yields and buggy pre-final drivers is the word from my sources.
AMD can't afford another fumbled launch. AMD has limited availability to approximately 30,00 early adopters to monitor for QA and QC problems. As they address the inevitable issues that arise with a rushed launch, and correct errata and driver bugs, they'll pour more units into the channel.
But the big problems remain. This product underperforms the equivalent nVidia GPU in 6 out of 10 titles, even at 4k. Those are the headlines today. AMD's own refusal to allow pre-release independent testing and resorting to benchmark configurations one formerly pro-AMD author called "sleazy" and "deceptive" in its own pre-release marketing, shows that AMD was already well aware that the new GPU was little better than the 390X and not particularly competitive with nVidia.
If they pour too many units with shoddy drivers and buggy chips out there too early, the bad press will kill any possibility of selling these successfully. Conversely if they hold back to many while they do the quality work that a reputable company would have done before launch, they increase the revenue shortfall they'll experience this and next quarter due to declining sales and market share in all segments.
There is no happy ending here for AMD. They developed yet another second rate product and launched it before it was ready.
HBM worthless, utterly.
AMD is a year late with its Maxwell answer and fails to answer the question correctly. Typical second rate imitator company!
In the Firestrike test, I see the 980 Ti with a lead at 4K. At 1440p, on the other hand, Nvidia’s flagship beats AMD’s latest beast, turning in 5% stronger performance. That’s a victory.
Of these 10 games, 980 Ti takes the crown on 60% of them. In the 4 titles where Fury wins it only barely edges out the 980 Ti in those titles. At 1440p GeForce 980 Ti gets clear victories in the majority of games, decisively jumping out in front.
At 4k GeForce 980 Ti retains its 6:4 title advantage, with even larger victory margins over Fury, despite the latter having HBM. More critically, the 4 titles where Fury wins, the victory margin at 4k is even slimmer than the slim margin at 1440p.
The conclusion is that Fury X is over-priced for its modest improvement over 390X (15% performance boost for 50% price increase).
Suck it down fantards. Suck. It. Down.
You seem to have missed QCOM's comments - they say they already acquired what they wanted from AMD long ago - the mobile graphics patents - and have zero interest in liquid-cooled 28nm GPUs or three-generation-behind x86 CPUs. Instead they are eyeing FPGA and PLD companies.
Remember when the fantards were pumping a SeaMicro Verizon server deal like wuttz pumps ruckpral?
Did AMD's abandonment of SeaMicro cause pain for Verizon? If so, does Verizon have grounds to sue?
The truly amazing thing is that this is written by Demrjian who after Chris Tom of AMDZone was one of AMD's top journalist fan boys.
For Charlie to use terminology like Fiji will get "blown out of the water" and that AMD's own benchmarks are "embarrassing", "sleazy" and "deceptive", it has to be really, really bad. Kind of like the Pope publicly proclaiming that Jesus wasn't the son of God and that Mary was a a floozy like ruckpral's mother.
This is the key takeaway: "The problem is that AMD took a few games and cranked up the memory requirements so high that they overran the 4GB buffer on the Nvidia card artificially hobbling performance. The only reason the 390X comes anywhere close to the Nvidia 980 is because it has 8GB and isn’t hamstrung by these artificial tests. In a fair fight the AMD card would be obliterated by the 980."
AMD stinks of desperation. Lisa Su is a #$%$ and is running AMD into the ground.
And absolutely no one cares. As Semi-Accurate wrote, in real world benchmarks for current titles, Maxwell blows Fiji out of the water and AMD knows it.
Comically, Fiji is a combination of dusty old TSMC and GF 28nm process and even dustier and older UMC 65nm process. What a watt-sucking hunk of junk! No wonder the rats are leaping over the rails into the sea. Not only is AMD sinking, but it's also burning.
Devinder is a moron. Multiple analysts who have reviewed the cross license agreement provisions related to change of control have confirmed that AMD cannot undergo *any* change in control related to the x86 intellectual property without Intel's permission.
Barron's: "Would the x86 license survive an acquisition? While change in control would invalidate the existing cross license agreement with Intel, we think in most circumstances Intel would be very willing to renew the agreement as Intel is interested in expanding the use of x86."
Forbes: "With that in mind, splitting up AMD into separate PC and server processor companies would be extremely hard as both markets are reliant on the same technologies. The first order problem would be that both companies would need a cross license from Intel to cover x86 patents (there is no “x86 license” per se). Intel would have to agree to license its patents to the two entities, which would be tricky to negotiate, but is essential for both to produce x86-based products."
Fortune: "The company has struggled since failing to see the development of the mobile computing market, which leaves it with a graphics chip business and the traditional x86 server business where it competes with Intel. It has licensed the technology to make x86 chips from Intel, and any deal might jeopardize that agreement."
AMD cannot be bought or spin out its x86 business without Intel's consent! There is no credible dissent.
The top end Carrizo APU is intended for $700 laptops. It'll have sub-Core i3 CPU performance and low end GeForce/Radeon discrete GPU levels of graphics performance.
But for $700 you can get a Core i7 equipped laptop that will out CPU Carrizo by 6-7x, that includes discrete AMD or nVidia graphics that will out game Carrizo's graphics by 3-4x, for that same $700, and the Intel-based system will have 150% of the batter life to boot.
As with every APU before it, Carrizo is a solution to a problem that never existed. They will stack up in the warehouse until AMD is forced to dump them at cost or for a loss. Just like Kaveri. Just like Richland. Just like Trinity. Just like Llano.
Despite months of fan boy pumping to the contrary, AMD's new Fiji GPUs do not outperform comparable nVidia GPUs. In fact, they barely match them despite being 6 months later to market. nVidia's next chip, Pascal, is just 6 months away and will surpass nVidia's current and AMD's upcoming Fury GPUs by 10x (not a typo!) in performance.
AMD claimed to be delivering revolutionary performance with Fiji and instead has only managed to barely catch-up in performance; AMD still trails badly in efficiency and heat dissipation. Look for 1s next quarter, fan boys. This peez of sheet is going down faster than your wife on the mail man!
I think you're confusing a middle aged immigrant Fry's sales clerk with an AMD product manager. And you've overstated the AMDers wages by about 12%.