Where's the Ryzen Effect? Intel posts Record Financials

[ypsylon]

Most market analyst were in agreement that 2nd quarter will bring something in numbers. Make no mistake Intel is giant and CPU is just one of the branches. They can fund chip making division with other divisions easily.

However it hurts them, in a plethora of numbers you can easily drown the under-performing division. Between March and June it'll be certainly more visible. Especially if AMD floods the market with Ryzen laptops. That was the stalwart, monopoly market for Intel for decades and if AMD can dislodge Intel here (with clearly superior processors) it'll hurt the bottom line. Make no mistake. Moves from companies like Dell or HP which were officially fed-up with Intel lack of supplies for IS and switched to AMD exclusively that's another. It's to early to really notice these shifts. It takes time. By the end of June I'm sure we'll see something. You can't take losses on such scale and hope that creative accountancy will solve all problems.

Intel surely are seething deep down... and it serves them well for stifling the market/innovation. In CPU space they simply have nothing worth looking at compared to competition. I wouldn't be surprised if AMD even pulled some low-power, low core count EPYC for mobile workstations. Why not? Reds just steamrolled Blues in 2019 and 2020 looks about the same in this regard.

IMHO Intel really missed the train of 3647 for HEDT. Instead resurrecting dead X299 platform they should've switched HEDT to 3647. There was plenty of potential. Me personally I don't need 28 core 5GHz Xeon 1kW CPU. But at stock that 28 core is very good chip. Push it to 10nm, work a bit on thermals, triple L3 cache, cut unnecessary server bits, cut the QPI links and convert them into PCIe lanes, lower the PRICE and it would sell easily. So yeah... Intel really were at sixes and sevens during 2019, will see if they have anything to respond in 2020.

 

[qcmadness]

We need AMD and NVIDIA financial data to comapre this quarter.

But lower unit volume in an expanding market does not bold well for Intel.

 

[64K]

We need AMD and NVIDIA financial data to comapre this quarter.

AMD's Financials for Q4 2019 and their 2019 Annual Release should be released in 4 days on the 29th and I'm sure that they will be posted here. Based on what I have seen AMD will probably report around 1/10th the Revenue of Intel and around 1/20th the profit of Intel for the year 2019 but you have to bear in mind that AMD's Financial Report will include their income and profit from their Graphics Business RTG.

 

[Xaled]

It IS definitely the Ryzen effect. Many people bought CPUs just because of the competitive new prices.
Intel won more despite decreasing prices because they sold more. It is that simple.

 

[Makaveli]

The title here is very click bait and sounds to be written by an intel fan boy I expect better.

 

[notb]

AMD's Financials for Q4 2019 and their 2019 Annual Release should be released in 4 days on the 29th and I'm sure that they will be posted here. Based on what I have seen AMD will probably report around 1/10th the Revenue of Intel and around 1/20th the profit of Intel for the year 2019 but you have to bear in mind that AMD's Financial Report will include their income and profit from their Graphics Business RTG.

Expected yearly revenue is $6.6-6.8 bln, but of course there's a decent chance of surprise. Ryzen 3000s sell well.
Anyway, we're moving around that 10% of Intel's you mentioned.

But as for earnings (GAAP)...
Intel made $21 bln.
Market consensus for AMD is around $0.3 bln. That's 1.4%.

And? Disney buys tickets to make their movies look good (multiple levels of fraud) and is rumored to be lying about profits for 6 yrs, now lol.

If you buy your own product - which is quite normal, especially when you're a top manufacturer - you can boost some indicators ("popularity", "market share"), but it doesn't impact earnings.
When Intel buys a $300 CPU, that results in +$300 as revenue and -$300 as operating expenses.
Of course Intel isn't using DIY PCs/servers, so most CPUs (sold to OEMs and "bought back" in computers) are included in revenue. Nothing wrong with that, right?

When Intel makes a chip for internal needs (like testing or an in-house manufacturing device) it isn't sold and is not shown in revenue nor cost of sales. It goes directly to operating expenses (the actual cost of manufacturing, not MSRP etc).

You absolutely can manipulate earnings.

As I already said: hiding earnings (lowering reported earnings) often makes sense and many companies try.
Boosting earnings (lowering reported costs) is extremely difficult and... pointless. Why would Intel want that? They're paying a dividend. They'll have to give most of that to their shareholders.
You have many theories, so maybe you can give a sensible reason?

The title here is very click bait and sounds to be written by an intel fan boy I expect better.

It sounds quite funny when you call @btarunr an Intel fanboy.
I believe the topic, being clickbait or not, is quite honest. I'm sure he (like many people here) expected Intel to show much worse results.

 

[XiGMAKiD]

The Ryzen effect is Intel keeping their 10nm in development limbo to keep profit high enough by reusing 14nm+++++ to satisfy investor

 

[W1zzard]

The title here is very click bait and sounds to be written by an intel fan boy I expect better.

I woke up, read the post, was like "wtf where's the ryzen effect?" and changed the title

 

[medi01]

This will be the year for us to see what Intel will bring to the GPU market. If they deliver a competitive stack at a reasonable price and proper driver support and there aren't major shortages causing retailers to price gouge then we may see Intel revenue and profit continue to climb.

Uh, nothing even remotely hints at Intel having something serious on GPU market any time soon.
The only "teaser" they had was around power consumption.

That being said, I'd prefer Intel to actually have a competent GPU offering.

Disney buys tickets to make their movies look good

I have read dubious "570 tickets bought online, but nobody has shown up" stories, but financial aspect makes... hard to believe.

It IS definitely the Ryzen effect. Many people bought CPUs just because of the competitive new prices.
Intel won more despite decreasing prices because they sold more. It is that simple.

Intel's margines barely dropped.

 

[yeeeeman]

Do you want a monopoly or something? Comments like this are so confusing.

People are always looking 1 meter in front, no further. The dude doesn't understand that if AMD remains the single player or gets in the same monopoly situation as Intel was, it will do the same things. Small perf bumps with same price.
AMD is what it is today because they didn't have a choice but to risk and have low prices. It is not an act of kindness.

Baity title, Intel's margins are down - they don't break it for each (sub) segment I believe - so you won't see the Ryzen (DIY) effect just from a cursory glance. Secondly, the R&D plus CAPEX will go up, unless Intel wants to stick to 14nm for the foreseeable future. Lastly layoffs, when was the last time a company experiencing record profits did this
Intel’s pending layoffs will hit multiple business groups, Oregon workers

Did I mention notebooks, would be quite the spectacle when AMD goes full throttle with zen2 APU

Ryzen DIY segment is a drop in the ocean. Peanuts compared to the TAM. So stop believing that bragging on reddit with new AMD builds means Intel doesn't sell anything, cause it is just not true.
--------------------
Now, my own opinion on this is split into multiple parts:
1. Intel is a much better know name/brand, whereas AMD is still seen as the cheap, hot and slow variant. Mentalities change very slowly. Some hundred posts on reddit are nothing.
2. Intel still has a lot of succes with OEMs and a LOT better and more support for them.
3. Skylake based products are still pretty good (for gaming still the best), even if they are beaten on price and power.
4. Using a 5 year old, proven process for making your products that are still competitive is something unheard of in the CPU industry. You usually have to continously invest in a new process which means lower profits, but since Intel is still using 14nm to this day, their total expenses are probably very low compared to what AMD needs to pay for a 7nm CPU.
So as an example, if 9900K and 3700x would cost the same, Intel has much higher margins on its product compared to AMD who needs to pay TSMC.
5. Much better marketing.

As a sane person, I wouldn't hate Intel. I would just wish them to come back and kicking with amazing new products, with inovation. AMD current products are indeed amazing compared to Intel ones, but this is just because Intel is still using 5 year old process, 5 year old uArch which is within a few percents of IPC compared to latest Zen 2 core. I would say this is pretty awesome and shows just how good these Skylake based products were in 2015. If it were for Intel to not have fab problems, I doubt Zen 2 would have been the same succes as it is today.

 

[notb]

Uh, nothing even remotely hints at Intel having something serious on GPU market any time soon.
The only "teaser" they had was around power consumption.

I think the "teaser" and all the other leaks suggest that a ~GTX1650-1660 card is around the corner. I don't see why June 2020 wouldn't be realistic.

Intel is focusing on power consumption, because their key market are laptops. In other words: if they're ever going to make gaming/workstation GPUs, they have to work in notebooks first, desktops second.

Intel's margines barely dropped.

They did, but looking at history, it's within statistical fluctuations.
If we see a gross margin of 50% at some point, that would be clearly coming from competition. But it would be better for this market, if AMD's margins went up, not Intel went down. Of course that would be better for AMD as well. Their net profit is just sad.

 

[bug]

People are always looking 1 meter in front, no further. The dude doesn't understand that if AMD remains the single player or gets in the same monopoly situation as Intel was, it will do the same things. Small perf bumps with same price.
AMD is what it is today because they didn't have a choice but to risk and have low prices. It is not an act of kindness.

This.
Two players doesn't mean much competition either, ideally I'd like to see about a dozen of them. But we have to be realistic here, the entry barrier is just too high for that to ever happen. Too bad the likes of Cyrix and VIA couldn't keep up.

 

[ssdpro]

You have to remember the marketing image portrayed online is not the reality. You might think AMD is gaining major ground and theoretically they are. But sales are up across the board - enough for Intel to grow and AMD to grow. And AMD growing from almost nothing to something is mega but still small compared to Intel's (growing) share. AMD still needs to keep on growing. This doesn't happen in 2 years it will take a decade or more to be truly competitive to scale.

 

[notb]

You have to remember the marketing image portrayed online is not the reality. You might think AMD is gaining major ground and theoretically they are. But sales are up across the board - enough for Intel to grow and AMD to grow. And AMD growing from almost nothing to something is mega but still small compared to Intel's (growing) share. AMD still needs to keep on growing. This doesn't happen in 2 years it will take a decade or more to be truly competitive to scale.

How exactly was AMD "almost nothing" before?
They had almost 10% market share in PC CPUs (both desktops and laptops), which basically doubled in the Zen era. But having 10% of something as important and (theoretically) lucrative as x86 processors is hardly a thing to ignore.
On the GPU side they have even higher share, albeit here growth potential is smaller even with 7nm Navi (because Polaris was relatively less ancient than pre-Zen CPUs were).
And they dominate in consoles.

The only segment where they're going "from nothing to something" are servers. But here they currently have roughly 5% market, which is lower than what silly looking FXes managed in 2016.

 

[Makaveli]

Expected yearly revenue is $6.6-6.8 bln, but of course there's a decent chance of surprise. Ryzen 3000s sell well.
Anyway, we're moving around that 10% of Intel's you mentioned.

But as for earnings (GAAP)...
Intel made $21 bln.
Market consensus for AMD is around $0.3 bln. That's 1.4%.

If you buy your own product - which is quite normal, especially when you're a top manufacturer - you can boost some indicators ("popularity", "market share"), but it doesn't impact earnings.
When Intel buys a $300 CPU, that results in +$300 as revenue and -$300 as operating expenses.
Of course Intel isn't using DIY PCs/servers, so most CPUs (sold to OEMs and "bought back" in computers) are included in revenue. Nothing wrong with that, right?

When Intel makes a chip for internal needs (like testing or an in-house manufacturing device) it isn't sold and is not shown in revenue nor cost of sales. It goes directly to operating expenses (the actual cost of manufacturing, not MSRP etc).

As I already said: hiding earnings (lowering reported earnings) often makes sense and many companies try.
Boosting earnings (lowering reported costs) is extremely difficult and... pointless. Why would Intel want that? They're paying a dividend. They'll have to give most of that to their shareholders.
You have many theories, so maybe you can give a sensible reason?

It sounds quite funny when you call @btarunr an Intel fanboy.
I believe the topic, being clickbait or not, is quite honest. I'm sure he (like many people here) expected Intel to show much worse results.

Actually I said sounds like it was written by a fan boy I didn't come out and say he was.

And to be honest this results are only a surprise to those that just gamers and people that don't actually follow these companies.

When you see all the products and services intel offers in area's outside of desktop cpu's these numbers are not a surprise.

 

[notb]

Actually I said sounds like it was written by a fan boy I didn't come out and say he was.

Even so, he isn't and certainly doesn't sound as one. Not here and not in any text he authored on this site lately.

When you see all the products and services intel offers in area's outside of desktop cpu's these numbers are not a surprise.

Compared to 2018 they're flat in CCG which consists of pure consumer chip sales. No services. So it's far from bad. And yes, this is a surprise. Not currently, because it was known for a while that this is a very good year.
But if you asked anyone a year ago, it was natural to expect AMD will take more of the market.
Instead, AMD just halted in 2019. Maybe 2020 will be better (it is expected to be).

Other markets were in line with expectations. Stable datacenter and big growth in new businesses.

 

[Vayra86]

A big thing is consistency. AMD is going to have to keep the momentum for a long time, and while they have done that for 3 generational releases now, it will have to last a lot longer, basically an entire lifecycle for whatever market they want to own. We are not even close to half that right now since Ryzen released, and let's face it, 1st gen wasn't exactly faster or better in any conceivable way versus what Intel had at the time. It was close. But not better.

The whole Intel is big and other divisions is true, but it has little weight in this discussion, the numbers simply don't lie. Things don't happen overnight, just think of your own way of purchasing systems. We research, analyze the market too and consider what the best purchase might be, and many people, if the time isn't right and if they are informed, can and will wait if they need to. Look at how long we've sat on our quad cores... There were options faster than that. But out of our reach / not realistic and also not exactly necessary.

The real moment of truth is when Intel actually comes up with something radically new that they are supposed to iterate on again; if AMD can waltz over thát, then they will get the market share in a big way.

 

[medi01]

 

[notb]

The real moment of truth is when Intel actually comes up with something radically new that they are supposed to iterate on again; if AMD can waltz over thát, then they will get the market share in a big way.

I think the main challenge for AMD is to finally target other groups than people who write on forums like this one.
Revenue fluctuates, with spikes after new Zen launches. And if you look at Zen users on this forum, many of them had CPUs from all three generation. Some openly said a 3700X is temporary and they're going to go for 3950X when it is available / price drops.
So it would be interesting to see how AMD copes with a different kind of client - one that replaces a PC every 3-5 years.

And of course AMD can't keep this momentum in performance gains. They'll slow down and start iterating just like Intel.
This should already become visible with Zen 3 - the first Zen generation for which they won't have a strong external factor boosting potential.

In case of Intel this is a lot more obvious. Their only real problem is 10nm capacity. If they were able to make all their CPUs on denser node (like AMD does), we wouldn't have this discussion.
And they'll have to sacrifice the HEDT segment, which is an important PR loss, but nothing directly significant for sales.

/atrocious screenshot/

You should definitely look into mobile image editors. Give Snapseed a chance.

 

[bug]

View attachment 143313

Besides the obvious typo, is there another point to your post?

 

[svan71]

Yes Intel despite your lack of innovation and constant security flaws but AMD posted records as well, it's called a booming economy, thank the President who promised it.

 

[mtcn77]

This should already become visible with Zen 3 - the first Zen generation for which they won't have a strong external factor boosting potential.

Is this a joke I'm failing to comply? Zen 2 is the first seriously good chip they got that even the bottom of the pile is sizing up the ladder. Zen 3 on the other hand will be market shifting. The needed highlights are all there, they could even launch the same chip with a discrete pbo profile, one that fits pb2 and not limited by fit temperature control(62°C>95°C)

Ryzen 7 3700X high temperatures

Hello everybody, Yesterday, I finished my new build featuring the Ryzen 7 3700x, it contains: Ryzen 7 3700X MSI X470 Gaming Pro Carbon Be Quiet! Dark Rock 4 16 GB DDR4 Tridenz Z RGB 3200 Mhz CL16 Fresh Copy of Windows 10 Fractal Design Meshify C with 2 intakes 1 exhaust Now, after some...

community.amd.com

After applying metal liquid between the cooler and IHS things improved by almost 20°C in the range of 65-80°, that was when the CPU draws from 80 to 120watts, (80° became 60 at 120watts!) but improved only by 3°C when the wattage jumps over 150watts, (90° instead of 93-94°), also frequencies improved accordingly.

95°C is almost the point where 200w tower coolers are not held back by the interface thermal paste. Being able to boost thereabouts is where coolers make up for the most of overclocked cooling capacity.

Under 55°C it sits at 4350MHz all cores and occasionally some cores jump to 4.4

@65°C it draws 80Watts and sits at 4150MHz all cores

@75°C it draws 90Watts and sits at 4100MHz all cores

@85°C it draws 100-110Watts and sits at 4050MHz all cores
@90°C it draws 150-160Watts and sits at 3980MHz all cores
@94°C+ almost 170watts, throttling rules and it sits at 3900MHz all cores

 

[notb]

Is this a joke I'm failing to comply?

What I meant was: earlier AMD had a fairly easy job delivering performance jumps.
Zen came out after years of negligence. It seemed a bit half-baked, but delivered a massive boost over FX lineup.
Zen+ was a proper launch, making Zen more robust and faster (if only this could be done with VBIOS!).
Zen2 benefits from 7nm.

We'll see what Zen3 turns out to be. If it's just a "Zen+" kind of polishing, this won't be as impressive as some think - simply because Zen2 isn't as bad as Zen was.
It's the same architecture, similar node, the same socket, with the same number of cores. So we'll see how much AMD team can really give us in a simple architecture iteration.

 

[Vayra86]

What I meant was: earlier AMD had a fairly easy job delivering performance jumps.
Zen came out after years of negligence. It seemed a bit half-baked, but delivered a massive boost over FX lineup.
Zen+ was a proper launch, making Zen more robust and faster (if only this could be done with VBIOS!).
Zen2 benefits from 7nm.

We'll see what Zen3 turns out to be. If it's just a "Zen+" kind of polishing, this won't be as impressive as some think - simply because Zen2 isn't as bad as Zen was.
It's the same architecture, similar node, the same socket, with the same number of cores. So we'll see how much AMD team can really give us in a simple architecture iteration.

Great points. But let's look at Core for a minute, too. That was iterated on as well and included some sizeable gains. I think Zen has a few more rabbits to pull out of the hat, whereas Intel is in fact still iterating on Core. Zen 3 was also pretty surprising in how they approached the latency problem. Its dangerous to give AMD too much credit here, but the fact their design is so radically new is definitely giving them more potential right now. They've basically created new 'low hanging fruit'. Even if they might not know it today.

Exciting times in any case

 

[londiste]

Great points. But let's look at Core for a minute, too. That was iterated on as well and included some sizeable gains. I think Zen has a few more rabbits to pull out of the hat, whereas Intel is in fact still iterating on Core. Zen 3 was also pretty surprising in how they approached the latency problem. Its dangerous to give AMD too much credit here, but the fact their design is so radically new is definitely giving them more potential right now. They've basically created new 'low hanging fruit'. Even if they might not know it today.

Zen is not radically new. Zen2 brings chiplets to mainstream but this is a packaging not architecture question. Performance seems to come in large part from how the latency problem was addressed - much larger L3 cache. How Zen3 will improve things is a bit of an open question right now but rumors suggest that distancing from 4-core CCX will be part of it, with bolstered frontend and larger L1/L2 caches.

In an interesting way, Intel is following AMD's playbook. Sunny Cove (Ice Lake) uses larger caches among other things that play a role in increased performance. Willow Cove (Tiger Lake) supposedly has additional cache size increases and improved frontend. Since Intel's execution stage is somewhat narrower and should be more prone to compute execution bottlenecks, both also widen the execution stage, largely with store ports but their powerful execution port allocation was also reworked a bit for Ice Lake and probably will get some more work done for Willow Cove.