AMD reports FY 2019 results

MangoSeed

[H]ard|Gawd
Joined
Oct 15, 2014
Messages
1,447
Very solid results on the back of Ryzen and Navi sales. AMD paid back a ton of debt which means they are really confident in cash flow going forward. Free cash flow for AMD means more R&D, more competition and better products for us. Whoop!

https://www.anandtech.com/show/15445/amds-fy2019-financial-report

Today AMD announced its 4Q 2019 revenue of $2.13 billion, up 18% from the previous quarter and up 50% from the same quarter last year. This is accompanied by a 45% gross margin for Q4, AMD's highest on record, up from 38% from Q4 last year and up from 43% in Q3. Operating income for the quarter was up a staggering 1143% (not a typo) from $28 million a year ago to $348 million this quarter, and net income was up 347% to $170 million. The resulted in earnings per share of $0.15, up 275% from a year ago.

Within Q4, AMD paid down a lot of debt - down from $1087m in 3Q19 to $563m in 4Q19, which would lead to suggestions that AMD could pay off its debt by the start of 3Q20. This also corresponds with a positive free cash flow, increasing from $1209m in 3Q to $1503m in 4Q.
 
Good, the sooner they get positive cash flow the better. They need to continue to capitalize on Ryzens success and keep that momentum for at least 5 years before they start chipping away at Intels market dominance.
 
Good, the sooner they get positive cash flow the better. They need to continue to capitalize on Ryzens success and keep that momentum for at least 5 years before they start chipping away at Intels market dominance.
It isn't even going to be 5 years. AMD is on fire with their CPU lineup and they're not slowing down with the Ryzen 4000 series on the horizon. Intel is on their umpteenth security flaw and their manufacturing seems to be hopeless. Krzanich's retirement wasn't a coincidence.
 
It isn't even going to be 5 years. AMD is on fire with their CPU lineup and they're not slowing down with the Ryzen 4000 series on the horizon. Intel is on their umpteenth security flaw and their manufacturing seems to be hopeless. Krzanich's retirement wasn't a coincidence.

I'm fairly certain that Joker is referring to the Server space. That's where the real money is and Enterprise customers are very slow to move and prefer a steady hand over "best performing". Overall platform and reliability matter far more to a lot of clients rather than raw speed. AMD has to keep going up and Intel needs to keep faltering for there to be a real shift in marketshare.
If AMD captures even 5% of Server that would be an insane win. As that would be an absurd amount of cashflow.
 
It isn't even going to be 5 years. AMD is on fire with their CPU lineup and they're not slowing down with the Ryzen 4000 series on the horizon. Intel is on their umpteenth security flaw and their manufacturing seems to be hopeless. Krzanich's retirement wasn't a coincidence.

Yet Intel just posted record profits.

I'm glad AMD is doing well. Just don't count Intel out yet.
 
Yet Intel just posted record profits.

I'm glad AMD is doing well. Just don't count Intel out yet.
Intel has their hands in a lot of pies, including enterprise and oem, embedded, networking, ai, and others, so that's not surprising. You aren't going to get rid of Intel, even just competing with them is impressive, especially in the enterprise cpu arena. Hopefully zen and it's grandchildren aren't just a fluke. lol
 
I'm fairly certain that Joker is referring to the Server space. That's where the real money is and Enterprise customers are very slow to move and prefer a steady hand over "best performing". Overall platform and reliability matter far more to a lot of clients rather than raw speed. AMD has to keep going up and Intel needs to keep faltering for there to be a real shift in marketshare.
If AMD captures even 5% of Server that would be an insane win. As that would be an absurd amount of cashflow.

Spot on. The one thing that Intel has over AMD right now is a wide variety of Enterprise products tailored for the data center covering all price ranges.

AMD needs to diversify this segment of their product portfolio in order to merely have a chance at making a measurable dent in Intel's established stranglehold.
 
Intel has their hands in a lot of pies, including enterprise and oem, embedded, networking, ai, and others, so that's not surprising. You aren't going to get rid of Intel, even just competing with them is impressive, especially in the enterprise cpu arena. Hopefully zen and it's grandchildren aren't just a fluke. lol
WAIT are you telling me the global markets of these multi-hundred-billion dollar megacorps extend beyond the Microcenter display case?
 
Last edited:
What Intel has is a bunch of people in the Enterprise space that are afraid and unwilling to purchase a superior product from AMD.
 

it's "whoopee!" not "whoop!" :barefoot:

anyway, all the comments that talk about the demise of Intel is beyond foolish, makes no sense whatsoever other than a hate for Intel. I greatly disliked when they deceived folks with their i7 "k" series CPU's that overheated due to cheap TIM but I don't hate them.

Maybe put things in perspective ... might save some folks an ulcer down the road a piece ;)
 
Spot on. The one thing that Intel has over AMD right now is a wide variety of Enterprise products tailored for the data center covering all price ranges.

AMD needs to diversify this segment of their product portfolio in order to merely have a chance at making a measurable dent in Intel's established stranglehold.


All amd needs todoks get model parity with Intel when it comes to ordering a CPU.

If I could spec a same model of server with an amd epyc CPU instead of an Intel CPU when building out a new set of servers for a hardware refresh I would and give that as an. Option to my director.

Amd needs to work with its vendors to get these high density cpus more visibility.

For instance if I want to retire a hundred or so small sql servers and put it on to a massive box. A pair of 64 core epyc cpus would give me more than enough density and ram and i/o capacity to do it.

Especially if it can run mssql better or as good as Intel per tick.

That's just one example.

Esxi hosts are per socket licensed. If I can get more cpus in a socket that's a win win.
 
I don't believe AMD has ever been behind Intel on a technical level. Yes even the dozer years. CMT was ahead of its time. (yes those chips lol) Intel had the right solution for a single threaded time. AMD bet on software developers going multi threading a lot faster then what ended up happening. So sure Sandy Bridge was the better chip... but the AMD was technically superior, there is a reason why 2019/20 benchmarks of 8 year old Intel vs AMD chips put AMD in the lead. (obviously taking the lead 8 years after launch is not ideal lol). From the K5 right through Zen 2 you could argue AMD has made all the steps forward... when they bet right such as dropping some FPU performance in favor of improved caching and BP way back in the K6 they made a ton of money. With the first athlons their alpha inspired design scored AMD mile stones like 1ghz and introduced design ideas that everyone picked up. Athlon64... X86_64. Then they bet on software threading 5 or 6 years to early and it hurt.

They seem to be back on track now though with Lisa Su... I find it interesting that AMDs hayday as enthusiasts see it back in the Athlon-Athlon64 days the company was helmed by one of the industries then best CPU minds in Dirk Meyer, and now that they are rolling again the same can be said of Su. AMD works best when the people at the top really know the tech. Being the baby fish in the market they have to make bold moves on the technical side of things, and to do that you have to really understand what your betting on. Meyer left just as the company choose to throw in on the Bulldozer arch 100%... rumors where at the time Meyer was forced out cause he wanted to push one more Athlon core revision and push harder into the server market. I really believe if he had stayed on and we had gotten one more revision of the Alphalon... AMD could have had a lot different last 7 or 8 years. A phenom 3 could have brought just enough IPC gains to compete much move favorably against the Intel bridge chips of the day as basically ALL software was still single threaded. A Bulldozer a couple years later with a step up fab process and a world where at least the big major benchmark friendly software would have been more threaded would have looked a lot better.

Anyway good to see both AMD and Intel making good money... with some luck hopefully the next few years will see some real innovation. AMD can't afford another bulldozer age of 5+ years of mediocrity. And I agree for them to really crack into the big money server markets they are going to need 5 years of Zen 2 type server chip smack downs. Intels lead on the software side is deep. I know AMD can't afford to throw money at the software side of things like Intel can, but they need to step up there for sure. For us PC knuckle heads its easy to drop a few grand and get whatever is fastest that day... for a company dropping 6 7 and sometimes 8 figures on hardware, software stability and support means lower ongoing costs. AMD needs to wise up to that. Having server chips that are 10 20 or 50% faster isn't enough.
 
Last edited:
I think AMD is well aware if its challenges ahead. They are looking at a BIG change in the landscape for the enterprise environment. BIG changes like that require a calculated approach and most importantly, time.
 
  • Like
Reactions: ChadD
like this
I like AMD and I love what they are doing right now, but I am worried for them. Even if tomorrow they launched a new revolutionary CPU that had 2x the performance watt per watt and thread per thread as any Intel CPU Xeon or otherwise and had a line of GPU's that trounced nVidia the absolute best they could manage is to get up to like 5 maybe 6% of the global CPU sales and possibly 30% of the global GPU market because they just don't have access to Intel's vast manufacturing capabilities, they just can't produce the volume that would be needed to meet that kind of demand. While selling off their fabrication is probably what let them stay afloat to begin with not having access to a dedicated Fab will forever keep them to a smaller fraction of the market share, and I am not sure how they are going to overcome this hurdle going forward.
 
I like AMD and I love what they are doing right now, but I am worried for them. Even if tomorrow they launched a new revolutionary CPU that had 2x the performance watt per watt and thread per thread as any Intel CPU Xeon or otherwise and had a line of GPU's that trounced nVidia the absolute best they could manage is to get up to like 5 maybe 6% of the global CPU sales and possibly 30% of the global GPU market because they just don't have access to Intel's vast manufacturing capabilities, they just can't produce the volume that would be needed to meet that kind of demand. While selling off their fabrication is probably what let them stay afloat to begin with not having access to a dedicated Fab will forever keep them to a smaller fraction of the market share, and I am not sure how they are going to overcome this hurdle going forward.

Actually that's kind of wrong... and let me explain a bit.

AMD not having manufacturing facilities allows them to focus on development even when they arn't selling through to the market very fast. So when there is a lull or not a lot of demand they just decrease the amount they produce. No need to fire people or worry about layoffs and the associated legalities with unions and all of that. instead they can just not buy as much product.

Then not having their own fabs also means they don't have to absorb the cost (instead its spread our among a fabs customer base) they can just say they want to go with x technology for production and run with it. then they can spread their orders over vendors that can meet the demand for product that they have. They don't have to build more fabs to have them sitting empty at the end of the product cycle. Instead they can focus on development and leave the production woes to their business partners to fulfill.

Yes that means that their per CPU cost is a litle higher at the time of order but it also removes the overhead of running massive multi billion dollar fab's on their own that they absorb all of the cost of.

Look at TI. It is a HUGE company with many of it's own fabs. They do a LOT of CPU and chip production for other companies to help offset the overhead of having as many fabs as they do. Yet they still have fabs with empty parking lots due to not enough demand.
 
Actually that's kind of wrong... and let me explain a bit.

AMD not having manufacturing facilities allows them to focus on development even when they arn't selling through to the market very fast. So when there is a lull or not a lot of demand they just decrease the amount they produce. No need to fire people or worry about layoffs and the associated legalities with unions and all of that. instead they can just not buy as much product.

Then not having their own fabs also means they don't have to absorb the cost (instead its spread our among a fabs customer base) they can just say they want to go with x technology for production and run with it. then they can spread their orders over vendors that can meet the demand for product that they have. They don't have to build more fabs to have them sitting empty at the end of the product cycle. Instead they can focus on development and leave the production woes to their business partners to fulfill.

Yes that means that their per CPU cost is a litle higher at the time of order but it also removes the overhead of running massive multi billion dollar fab's on their own that they absorb all of the cost of.

Look at TI. It is a HUGE company with many of it's own fabs. They do a LOT of CPU and chip production for other companies to help offset the overhead of having as many fabs as they do. Yet they still have fabs with empty parking lots due to not enough demand.
I get that when they are in a lul it lets the focus on development and keep costs low but it also means when they are booming they can't possibly keep up with demand at this stage how many fabs can do the 7nm process and how many different companies are bidding for time at those plants, going forward how many are going to be able to do 5nm so on and so on, as demand on the limited plants grow either costs go up or yields go down which plays a large role in their ability to get parts to market at specific price points. Pure manufacturing capabilities is what keeps Intel where they are even with an older process and less competative product stack they can still flood the market and even they are having a hard time keeping up with demand.
 
All amd needs todoks get model parity with Intel when it comes to ordering a CPU.

If I could spec a same model of server with an amd epyc CPU instead of an Intel CPU when building out a new set of servers for a hardware refresh I would and give that as an. Option to my director.

Amd needs to work with its vendors to get these high density cpus more visibility.

For instance if I want to retire a hundred or so small sql servers and put it on to a massive box. A pair of 64 core epyc cpus would give me more than enough density and ram and i/o capacity to do it.

Especially if it can run mssql better or as good as Intel per tick.

That's just one example.

Esxi hosts are per socket licensed. If I can get more cpus in a socket that's a win win.

I agree...AMD really needs to pump up their involvement with their ODM/OEM partners to push the availability of Epyc equipped products.


They also need to expand the lineup:
Epyc 2nd Gen starts at $500 for 1P capable and $650 per processor for 2P capable.

In comparison, Intel's Xeon starts for less than half of that. Granted, I'm NOT comparing core/thread counts here, just starting price tags...
 
Spot on. The one thing that Intel has over AMD right now is a wide variety of Enterprise products tailored for the data center covering all price ranges.

AMD needs to diversify this segment of their product portfolio in order to merely have a chance at making a measurable dent in Intel's established stranglehold.

And they will but, it is a slow and steady process, year over year and considering where they were just 4 years ago...... :D
 
it's "whoopee!" not "whoop!" :barefoot:

anyway, all the comments that talk about the demise of Intel is beyond foolish, makes no sense whatsoever other than a hate for Intel. I greatly disliked when they deceived folks with their i7 "k" series CPU's that overheated due to cheap TIM but I don't hate them.

Maybe put things in perspective ... might save some folks an ulcer down the road a piece ;)

Nope, it is Whoop! despite the obvious fact that you do not want it to be. :)
 
Lisa Su on Cramer today.. Lisa repeatedly said "In for the long haul" and as an investor - you have to be too. She knows exactly what AMD is up against and I believe they will play the "Console sales slow because nextgen" card for the rest of the year. Reality shows Intel -> AMD adaption is slower than what investors want or realize. All for the reasons discussed in this thread. Great comments and insight.

On the supply / demand issue - One of the hardest things in the world is predicting demand for your product. AMD clearly underestimated 3k Zen demand, Papermaster has admitted. However, with Apple jumping to 5nm, AMD has purchased this capacity from TSMC and will have much greater flexibility in this regard in the immediate future, Sure- it's better to have your own fabs to spin up for demand, but we've seen the disadvantage as well. Huge R&D overhead and technological hurdles make owning your own fabs costly and distracting from your core business. Therein lies the inefficiency. Going fabless was a gutsy move for AMD, but the right one for them. One of many great calls made to get them where they are now.

AMD clearly has more work to do, especially on the Enterprise space, it's not just enough to have a great product but they need to compete at all levels. Bringing on a former ex-Intel for VP of their Server unit is a great step in that direction.

Let's hope AMD keeps grinding away and making all the right moves as we've all seen the benefits over the past couple years.
 
Lisa Su on Cramer today.. Lisa repeatedly said "In for the long haul" and as an investor - you have to be too. She knows exactly what AMD is up against and I believe they will play the "Console sales slow because nextgen" card for the rest of the year. Reality shows Intel -> AMD adaption is slower than what investors want or realize. All for the reasons discussed in this thread. Great comments and insight.

On the supply / demand issue - One of the hardest things in the world is predicting demand for your product. AMD clearly underestimated 3k Zen demand, Papermaster has admitted. However, with Apple jumping to 5nm, AMD has purchased this capacity from TSMC and will have much greater flexibility in this regard in the immediate future, Sure- it's better to have your own fabs to spin up for demand, but we've seen the disadvantage as well. Huge R&D overhead and technological hurdles make owning your own fabs costly and distracting from your core business. Therein lies the inefficiency. Going fabless was a gutsy move for AMD, but the right one for them. One of many great calls made to get them where they are now.

AMD clearly has more work to do, especially on the Enterprise space, it's not just enough to have a great product but they need to compete at all levels. Bringing on a former ex-Intel for VP of their Server unit is a great step in that direction.

Let's hope AMD keeps grinding away and making all the right moves as we've all seen the benefits over the past couple years.

I find this interesting, considering that AMD use to FAB for Intel, back in they day. :)
 
Yeah, I've kniown they were in the right hands since Lisa Su started paying down enough debt during the AMD stock buy frenzy back in 2016 (cut their most critical debt in-half by introducing new stock), but it's good to see it dissappear FIVE YEARS before their renegotiated 2025 payoff deadline :D

If they keep doubling revenue every four year, Intel is going to have BIG compney that will be passing it by around 2035!
 
Yeah, I've kniown they were in the right hands since Lisa Su started paying down enough debt during the AMD stock buy frenzy back in 2016 (cut their most critical debt in-half by introducing new stock), but it's good to see it dissappear FIVE YEARS before their renegotiated 2025 payoff deadline :D

If they keep doubling revenue every four year, Intel is going to have BIG compney that will be passing it by around 2035!

It was ironic that cramer and the other "financial wiz" questioned Su on lowering the debt so much. I'm sure it is / was high risk debt paid at a high rate, since it was aquired when AMD was on the verge of failure. Why wouldn't you pay it off?

Also entertaining is the impatient investors today selling off, they do this every quarter earnings announcement - expecting Enterprise Server to suddenly jump 200% on a superior product. Guess they missed the part on 25% yearly revenue increase as you mentioned too. So, shares are on sale today.
 
I'm fairly certain that Joker is referring to the Server space. That's where the real money is and Enterprise customers are very slow to move and prefer a steady hand over "best performing". Overall platform and reliability matter far more to a lot of clients rather than raw speed. AMD has to keep going up and Intel needs to keep faltering for there to be a real shift in marketshare.
If AMD captures even 5% of Server that would be an insane win. As that would be an absurd amount of cashflow.
AMD has already done a good chunk of sales to Google, Twitter, Amazon AWS, and I believe Dell as well. They definitely need to keep going and get order quantities up, but they have been making inroads for over a year now.
 
What Intel has is a bunch of people in the Enterprise space that are afraid and unwilling to purchase a superior product from AMD.

Well duh. You don't order hundreds to thousands of something, costing hundreds of thousands to millions, until its proven to be just as reliable as the last stuff you bought.

The fact my 3900x home built PC has been rocked solid means jack sh!t in my work world. So does the comments on these forums when it comes to buying corporate products. They are completely proprietary.

I'll watch the complaint forums for HP Ryzen ProDesks and Dell Ryzen OptiPlex for another year before I risk my career.
 
.
I'll watch the complaint forums for HP Ryzen ProDesks and Dell Ryzen OptiPlex for another year before I risk my career.

Keep Watching, yet another year. And keep patching and losing more performance, paying higher cost of ownership. How many years do you need?
 
The problem I see 5-10 years down the line for AMD is that Intel will move passed these current fabrication hurdles and will have 10/7/5nm lithography ready to go in their own foundries. The last part is super important because if you're going to cater to Enterprise in a meaningful way, it helps not to be dependent on an outside supplier for your production needs. This is where Intel has a huge advantage over AMD and it is why AMD will never take them down. Plus Intel is a top down solutions provider and continues to expand. AMD may be on the right track but they will always be a minor player in the CPU and GPU markets IMO.
 
The problem I see 5-10 years down the line for AMD is that Intel will move passed these current fabrication hurdles and will have 10/7/5nm lithography ready to go in their own foundries. The last part is super important because if you're going to cater to Enterprise in a meaningful way, it helps not to be dependent on an outside supplier for your production needs. This is where Intel has a huge advantage over AMD and it is why AMD will never take them down. Plus Intel is a top down solutions provider and continues to expand. AMD may be on the right track but they will always be a minor player in the CPU and GPU markets IMO.

What is good for the goose is good for the gander. If Nvidia can do it and they are fabless too - your argument crumbles on the GPU end to say the least.
 
What is good for the goose is good for the gander. If Nvidia can do it and they are fabless too - your argument crumbles on the GPU end to say the least.

NVidia isn't anywhere near Intels level.
 
NVidia isn't anywhere near Intels level.
Ok, so what's the argument for Nvidia/GPU? Not sure why you included them in a fab means #1 player argument. Saying AMD will never be a major player in GPU implies Nvidia, or maybe Intel? LOL
 
Ok, so what's the argument for Nvidia/GPU? Not sure why you included them in a fab means #1 player argument. Saying AMD will never be a major player in GPU implies Nvidia, or maybe Intel? LOL

I was referring to marketshare rather than capability with respect to NVIDIA. Looking at raw numbers, NVIDIA invests far more into R&D than AMD does and they have a stranglehold on the GPU market. Navi hasn't really made a huge dent. But as far as Intel goes, the foundry ability will always give them an edge in enterprise and they are a top down solutions provider and their corporate customers are addicted. They may have shortages now but that won't last and they are also now building discrete GPUs which will eat into both AMD and NVIDIAs lunch. I just find it funny when people try to count Intel out, they aren't even close to being out of the game.
 
And as far as AMD not ever going to be a major player - they have been for years. Enterprise is not the only CPU market.
 
And as far as AMD not ever going to be a major player - they have been for years. Enterprise is not the only CPU market.

Lets look at it another way, market cap:

NVIDIA: $151.758B
AMD: $56.272B
Intel: $287.885B

AMD has CPUs, GPUs and APUs yet they trail NVIDIA in market cap who is solely a GPU manufacturer. Then compared to Intel, they aren't even a quarter of their market cap. That makes them a distant third.

Look at AMD's projected server marketshare for end of 2020:

AMD-Market-Share-Q2-2019-640x252.png


It's a nice gain but still a pittance.
 
Back
Top