AMD Earnings: Second-Gen RDNA in 2020, Quarterly Revenue up 50% and 50% debt reduction in 2019

[Brooksie359]

5 hours ago, comander said:

The first bit I have to ask is "how do you know this?" Where is your evidence? How did you develop your expertise? 
Can you provide me with an income statement showing that the graphics division generated a few billion in profit?
So far I see claims but I don't see evidence or even a theoretical model.

The story of "The graphics division kept AMD alive" is at least somewhat true. The reason they were almost dead was that they acquired the graphics division. It's like saying "I volunteered to get cancer for medical research and the free cancer treatment I got with it is keeping me alive". 


For what it's worth, skimming a breakdown on a randomish year - https://wccftech.com/amd-q4-2015-yearly-financial-earnings/

it looks like the graphics & compute division lost $500M in 2015. I'm aware that the consoles would likely be under Semi-Custom. Even if 100% of profits under semi-custom are attributable to graphics (they aren't, AMD made very low margins on each console sale, think $1-2 per console - 300M consoles over 5 years means ~60-100M/year which isn't enough to make up for the fact that the financing cost for the merger were around 500M a year - also half of that APU is a CPU)

https://techreport.com/news/27225/amd-suffers-falling-revenue-announces-7-staff-cut/ in 2014 profits for C&G were 17M. 17 < 500. 

-----

If AMD had an extra $5BN, the extra 500M (which is a much smaller number than 5000M) wouldn't have mattered. 
What you're saying is akin to "Not wasting a $1M inheritance doesn't matter since I have an $8 an hour part time job" - NO. A modest return on that would've dwarved the impact of a low value cash flow.

https://www.macrotrends.net/stocks/charts/AMD/amd/financial-statements

Here's AMD's P&L. They lost 1.5BN total during the years you're claiming ATi helped. ATi didn't help $5BN worth which is what they would've made simply putting those funds in the market. The date that I used for "in the market" would've been basically the worst possible date right before the financial crash by the way. Basically the ABSOLUTE worst that could've been done. If the funds were put in on the deal close date, the lost opportunity would've been more like 15-20BN. 

You can also see that they cut their R&D expenditure A LOT. "hurray we now have 2 companies" doesn't matter THAT much if you have to lay off half of your engineers because you're out of money. 

----

Bulldozer's failure wouldn't have mattered as much if AMD had an extra $5-6BN to spare, even 1-2BN... 
They could've had one reasonably large team working on improving Bulldozer to make it competitive and another team working to get pseudo-Zen (probably based on Phenom) out by 2012-2013 (keep in mind Bulldozer was originally supposed to be out in 2009 - what we got in late 2011 was the second iteration and AMD would've already known that it was a loser by 2009). 

Again with the assumption that they would have simply invested it in the market. You can probably attribute most of the semi-custom under the GPU devision as even the ones that weren't use for consoles would still likely use their GPU design. Anyways you look at any earnings report for AMD from the time after the bulldozer launch and their GPU division was the only one that was doing well at all. If they had invested their money in something else would have panned out better? Maybe but we have no idea. Maybe they would have still made the same design decisions for the fx series and have it panned out horribly and they had invested the money elsewhere that didn't have any sort of liquidity and they go bankrupt. Again their gou division was their only saving grace after they bet on bulldozer and it panned out horribly. I mean look at intel 10nm. they have invested a lot of money into 10nm with little to nothing to show for it but they are lucky that they already had a rock solid cpu architecture so it didn't matter. the point is companies like AMD or Nvidia or Intel do not spend money like that investing in the stock market but rather on new innovations and technology that can easily fail or not turn out well even when a large amount of money is invested in it.

[Crowbar]

7 hours ago, comander said:

For some perspective, I've done lunch with the CEO of a Fortune 100 company (I am not on a first name basis) and straight up had an engineering SVP from that company call me and offer me a raise/promotion when I resigned to join my current company (FAANG/Unicorn). I know enough finance to be dangerous (think making dashboards for ~5BN of expenditures over several years and outlining sales/unit forecasts for the upcoming years - a lot of which were built on shoddy assumptions), have been actively reading about the tech industry presumably about as long as you've been alive and currently work in a space that draws from this knowledge. I'm NOT a somebody (just an above average cog in a massive machine) but I have a feel for how things work behind the scenes. For what it's worth I make about as much as most junior executives though I'm in a technical role. 

If I wanted to, I could probably pass the CFA L1+L2 exams with minimal effort (L3 might be a bit harder), and I'm one class away from being able to get a CPA (I had a minor in accounting in addition a STEM major + minor and a graduate degree in a STEM field)

I've also mentored a few small business owners and seen how pride, ego and hubris coupled with being cash strapped can lead to GROSS inefficiencies and lost potential from the ground level. Being close to bankruptcy creates an environment where clever financial engineering matters more than executing on sound strategy. It's hard to invest for the future when you're trying to keep the lights on and that's what took up most of the focus at AMD for a while (sold off GloFo, spun off business units, sold their corporate headquarters and leased it back). They were doing EVERYTHING they could to not default on upcoming debts (which would've made further financing near impossible). This is NOT efficient/effective.

 

Let's say, hypothetically, AMD had invested the 5BN in the S&P 500 in 2006. From 2006-2016 that would've gone from 5BN to 10BN.
AMD's market cap from 2006-2016 went from 18BN to 1.6BN. The success of Ryzen raised AMD's market cap by around 40BN (in otherwords, Zen/Zen+/Zen2 was worth something like 10-20 all the graphics stuff, conservatively[this is BSed math since it avoids diving into liabilities and other quasi relevant stuff, it's mostly intuition]). 


AMD+ATi combined did not have 5BN of gross profits during that time span. 

Well... it sort of did, the mobile graphics division of ATi had profits around that level. AMD was so cash starved they sold it for $65M in 2009. Let me reemphasize BUYING ATi ALMOST KILLED AMD AND THEY SOLD ASSETS AT DISCOUNT PRICES TO AVOID COMPLETE BANKRUPTCY. Imageon became Qualcomm's Adreno. Every time you read ADRENO(reordered Radeon), realize that it's what a separate and independent ATi could've had. It's arguable that Adreno alone is worth more than AMD's entire graphics division right now. 
https://arstechnica.com/gadgets/2009/01/amd-unloads-mobile-gpu-technology-to-qualcomm/
https://en.wikipedia.org/wiki/Adreno


Their "success in the console market" was NOT worth 10BN - the margins on those sales are low. It was not worth being so cash starved that they couldn't compete in the data center either. 

Again, buying ATi almost killed AMD. If Ryzen were a flop, AMD would likely be in bankruptcy proceedings right now. ATi was a huge risk. It did NOT pay off and the entire industry is worse off because of the hubris of Hector Ruiz, the man who inherited a winning Athlon 64 design and who oversaw the development of Phenom and Bulldozer.

 

 

The only winners from the AMD+ATi deal were Qualcomm, nVidia and Intel.

AMD lost 10 years of competitiveness in the CPU/server/datacenter (which is something like 10-100x as profitable as consoles), ATi lost the entire mobile GPU market which their technology CURRENTLY IS THE LEADING PLAYER IN, etc. It would've been cheaper for AMD to make GPUs in house and/or to keep nVidia as a partner who integrated graphics into a northbridge/PCH (with a dash of onboard cache).


-----

TLDR: AMD REALLY needed have paid 1-3BN less for ATi. That 1-3BN would've gone a VERY VERY VERY long way. Way less financial engineering over the last decade, and a lot more regular engineering. 

For someone that claims to "know enough about finance to be deadly" you sure don't look like it trying to cite articles from 5 years ago as if it's relevant today before AMD launched Ryzen or a brand new GPU architecture that we've yet to see how it scales.... I guess trying to convince us through stories with no merit should swing the opinion though, right? Despite all signs pointing to you just being another dime a dozen forum poster attempting to circle jerk themselves over past "accomplishments".

[Brooksie359]

8 hours ago, comander said:

Alright, what about the assumption that they DIDN'T pay interest at the weighted average cost of capital for the industry (which is around 10%, which works out to about the same amount)

Again, making VERY VERY generous assumptions... where's the other $400M per year across 10 years?
I'm being VERY VERY generous to the position of "ATi kept AMD afloat" and there's still a huge gaping hole. AMD wouldn't have needed to have been kept afloat if it didn't spend $5BN. 

 

If you invested in the stock market (S&P 500) at the WORST possible time, right before the stock market crash, and held on, it would've done better. Literally doing the WORST POSSIBLE JOB IMAGINABLE (with the constraint that the people in corporate treasury were avoiding calls/puts). 
 

 

I'm assuming that FX would be just as screwed up BUT that there'd be money to make a new uarch in a reasonable timespan starting the moment they knew it was FUBAR (instead of a few years later). 

AMD's R&D budget dropped by half during the FX years because they didn't have the money to develop stuff. Another team of engineers working on a parallel uarch would've done WONDERS. Literally just copy the design philosophy of nehalem/sandybridge using existing parts (which is what they did with Zen to some degree). Ultra conservative, non-ambitious. 

 

Dangerous. As in, I get the general idea and am generally correct but don't have enough experience to avoid common pitfalls. I admitted that outright. If someone said "you will die if you don't become a finance/strategy analyst at a top company (IB, VC, PE, Google Capital, etc.) within a year" I could do it. It's not a good fit for me though. 

And yes, citing financial statements and industry articles is a lot more valid than "this is what I think" with 0 theoretical motivation and 0 evidence. 

ATi caused a $500M a year gap for AMD (5.2BN at a 10% opportunity cost) and the free cashflow from ATi was not enough to cover it. This absolutely KILLED their credit rating. By 2016 AMD's debt was considered a junk bond and banks were legally NOT allowed to invest in it. The only way AMD was able to raise capital to keep the lights on was to accept it at a VERY HIGH interest rate. I haven't run the numbers but I wouldn't be surprised if the increased interest expenses actually outweighed the FCF from the graphics division. 
https://www.reuters.com/article/idUSFit951590?type=companyNews

I'm a huge AMD fan, I'm excited to see them back in the game again but... yeah, they made a stupid decision in 2006. I'll admit that teenage me thought it was a cool thing at the time. Cool doesn't mean "not stupid". 
 

 

Again they started the design of zen right after the launch of bulldozer panned out horribly so the idea that if they didn't have ATI they would have been able to start developing zen earlier is simply wrong. Again with the stock market comparison which is absolutely unrealistic. You dont see nvidia or intel investing in the stock market but rather they invest in new technologies and sometimes those don't work out well I mean look at 10nm or physx or any other investment in new tech that didn't work out. 

[Nicnac]

Redwaveredwaveredwaveredwaveredwave

[Crowbar]

14 hours ago, comander said:

Alright, what about the assumption that they DIDN'T pay interest at the weighted average cost of capital for the industry (which is around 10%, which works out to about the same amount)

Again, making VERY VERY generous assumptions... where's the other $400M per year across 10 years?
I'm being VERY VERY generous to the position of "ATi kept AMD afloat" and there's still a huge gaping hole. AMD wouldn't have needed to have been kept afloat if it didn't spend $5BN. 

 

If you invested in the stock market (S&P 500) at the WORST possible time, right before the stock market crash, and held on, it would've done better. Literally doing the WORST POSSIBLE JOB IMAGINABLE (with the constraint that the people in corporate treasury were avoiding calls/puts). 
 

 

I'm assuming that FX would be just as screwed up BUT that there'd be money to make a new uarch in a reasonable timespan starting the moment they knew it was FUBAR (instead of a few years later). 

AMD's R&D budget dropped by half during the FX years because they didn't have the money to develop stuff. Another team of engineers working on a parallel uarch would've done WONDERS. Literally just copy the design philosophy of nehalem/sandybridge using existing parts (which is what they did with Zen to some degree). Ultra conservative, non-ambitious. 

 

Dangerous. As in, I get the general idea and am generally correct but don't have enough experience to avoid common pitfalls. I admitted that outright. If someone said "you will die if you don't become a finance/strategy analyst at a top company (IB, VC, PE, Google Capital, etc.) within a year" I could do it. It's not a good fit for me though. 

And yes, citing financial statements and industry articles is a lot more valid than "this is what I think" with 0 theoretical motivation and 0 evidence. 

ATi caused a $500M a year gap for AMD (5.2BN at a 10% opportunity cost) and the free cashflow from ATi was not enough to cover it. This absolutely KILLED their credit rating. By 2016 AMD's debt was considered a junk bond and banks were legally NOT allowed to invest in it. The only way AMD was able to raise capital to keep the lights on was to accept it at a VERY HIGH interest rate. I haven't run the numbers but I wouldn't be surprised if the increased interest expenses actually outweighed the FCF from the graphics division. 
https://www.reuters.com/article/idUSFit951590?type=companyNews

I'm a huge AMD fan, I'm excited to see them back in the game again but... yeah, they made a stupid decision in 2006. I'll admit that teenage me thought it was a cool thing at the time. Cool doesn't mean "not stupid". 
 

 

What do you not understand that AMD's position prior to 2017 is meaningless? Stating the obvious that the company was in the hole prior to this date brings nothing of value to your piss poor argument.

 

And again no one cares about what you claim to have or could do for employment. You have zero evidence that suggests any part of what you are stating is or possibly could be reality. At the end of the day you're no better than any other mouth piece on this forum attempting to convince others that your opinion is the correct one with nothing to support it.

 

As it sits right now in 2020, AMD is in a good place and only picking up momentum. The "yeah but XXXX arbitrary date" reaches that you have been grasping at to hinge your argument on are irrelevant. I'm not going to bother addressing the hypotheticals you've brought up because unless you sit as a share holder chances are that you know anything more are slim to none.

[Brooksie359]

26 minutes ago, comander said:

Do you have any evidence for this? So far you have provided 0 links vs around a half dozen from me. I think I remember this time a bit better than you. 

Also Bulldozer was originally meant for a 2009 launch(on 45nm) NOT a 2011 launch(on 32nm). They should have known by 2009.

https://www.anandtech.com/show/3863/amd-discloses-bobcat-bulldozer-architectures-at-hot-chips-2010

"Due out in the first half of 2009, AMD's Bulldozer core is the true revolutionary successor to the K8 architecture. While Barcelona and Shanghai are both evolutionary improvements to the current core, Bulldozer is the first ground-up redesign since the K7."

 https://news.softpedia.com/news/AMD-to-Release-Bulldozer-in-2009-83957.shtml

"AMD will start sampling the Bulldozer chip in late 2009," - "However, the recent wave of layoffs that hit the chip manufacturer's headquarters are likely to make the company miss its deadlines. According to previous reports, the Austin and California offices are the most affected locations."

Also, from what I understand AMD historically had a bigger R&D budget (think multiple teams working on multiple things) and that took a huge hit (got cut by half - see financial statements I linked to a while back PLUS the articles talking about massive layoffs) right around the time the ATi merger happened. The ATi merger assumed FCF(free cashflow) which never materialized. 

Are you saying that AMD cutting their R&D in half didn't affect their ability to pivot/adapt?

I am saying having more R&D wouldn't have made them start development of zen any faster. Bulldozer was a bet that didn't pay off and regardless of how much R&D they had it wouldn't have stopped that from being the case. Also they did pivot and adapt through the use of their GPU division to pick up the slack when their cpus were not competitive so to say that the gpu division didn't help them adapt is also plainly false. 

[Crowbar]

24 minutes ago, comander said:

Is it fair to characterize that you don't think having their R&D budget cut in half while doing more stuff (GPU) had no impact on their ability to do research and development on new products?
 

I never said that it didn't help them adapt. I'm saying that it was not enough. 
Smacking a person with a cane until they're crippled and then giving them that cane to help them walk is not a good strategy. 

AMD would not have been crippled if they didn't pay $5BN for a company that was only worth $3BN and which effectively required $500M worth of financing payments while only bringing in around $100M worth of profit (on a good year, on the worst years it was worse). 

 

How is it meaningless? Their position in 2017 is a function of R&D started years and even decades earlier, a copyright/patent portfolio accrued over decades and the finances they inherited based on past operations and financing activities. 

 

I provided financial statements and industry articles from reputed sources. 
 

This is true. It's also true that they're NOT in as good of a place as they could have been in a counterfactual where they didn't shoot themselves in the foot. 

https://ir.amd.com/financial-information/sec-filings

https://ir.amd.com/financial-information/sec-filings?field_nir_sec_form_group_target_id[471]=471&field_nir_sec_date_filed_value=&items_per_page=10&page=1

I'm going to assume you've never read (or even skimmed) any of AMD's 10Ks.

 

Are you really still employing the "yeah but" strategy that always fails here?

 

I'm also not impressed with your hindsight warrior tactics...

 

Your argument is meaningless because it doesn't matter. What AMD did 10+ years ago is not current and quite frankly many large tech giants often purchase other IP's and initially take on a loss doing so. It's how you get your foot in the door and acquire intellectual property such as patents along with talent to set the ball rolling.

 

The fact that I have to explain this to you when you apparently "know enough about finance to be dangerous" indicates the exact opposite and  continuing to grasp at straws for the sake of argument quoting outdated articles rather then recognize that it's 2020 doesn't bode well for you either.

 

[Brooksie359]

1 hour ago, comander said:

I'm NOT saying I would have made the right choice regarding the acquisition back in the day. (part of why "I know just enough to be dangerous" was said). I am saying that anyone with a good education (e.g. bachelor's + masters in relevant subjects from top universities), knowledge of the industry and experience in the industry or an adjacent one SHOULD be able to reflect upon the past and draw reasonably good conclusions. 
 

 

I understand the idea of investing for the future. 
My statement was that THIS particular investment was a bad idea. Not every investment works out. 

As was stated in AMD's financial filings - clashflow impairment is a real concern and overpaying for ATi caused an impairment. If AMD had managed to get ATi for $2BN instead of $5.3 I'd actually be calling it a good buy. That extra 3BN would have allowed them to get lower interest rates. It would have allowed them to hold onto properties like Imageon. It would have allowed them to put something like 200M more a year into operations. 

The big issue I have with the ATi investment is that they paid 2x what it was worth. AMD itself agreed with this and felt they could justify it when they took MULTIPLE GOODWILL IMPAIRMENT CHARGES. 

Again you even state yourself that anyone can see a bad decision in hindsight but that is to ignore the fact that nobody knows if not investing in ATI all that time ago would have made them better off now unless you are some sorta fortune teller.

[Crowbar]

1 hour ago, comander said:

I'm NOT saying I would have made the right choice regarding the acquisition back in the day. (part of why "I know just enough to be dangerous" was said). I am saying that anyone with a good education (e.g. bachelor's + masters in relevant subjects from top universities), knowledge of the industry and experience in the industry or an adjacent one SHOULD be able to reflect upon the past and draw reasonably good conclusions. 
 

 

I understand the idea of investing for the future. 
My statement was that THIS particular investment was a bad idea. Not every investment works out. 

As was stated in AMD's financial filings - clashflow impairment is a real concern and overpaying for ATi caused an impairment. If AMD had managed to get ATi for $2BN instead of $5.3 I'd actually be calling it a good buy. That extra 3BN would have allowed them to get lower interest rates. It would have allowed them to hold onto properties like Imageon. It would have allowed them to put something like 200M more a year into operations. 

The big issue I have with the ATi investment is that they paid 2x what it was worth. AMD itself agreed with this and felt they could justify it when they took MULTIPLE GOODWILL IMPAIRMENT CHARGES. 

Having a post secondary degree has nothing to do with the level of ones intelligence or ability to make informed decisions. Many entrepreneurs especially in the age of the internet or even those that are just financially savvy have never stepped foot in a college and done very well for themselves. Regardless this is beyond the point. I question how you draw your conclusion outside of being a hindsight warrior that ATI wasn't valued at 5 billion when it was acquired. It takes zero logic (or education as you have stated) to claim that purchasing anything at half the price would have been better. This is common sense. In most cases the market determines the price of any given sale. Who are you to say there weren’t other bidders or offers on the table for ATI?

 

But again, none of this is relevant since it's not the early 2000's any more.

[Brooksie359]

17 hours ago, comander said:

My argument was that the ATi acquisition, at the price it was done at, was a bad decision, not that I make perfect decisions 100% of the time. 

For that to be a valid conclusion, either the counter factual with a lower acquisition price would have to be materially better OR no acquisition at all would've been better. 

I think it should be largely self-evident that a lower price would be better and AMD itself (as a company along with the CEO who went through with the deal as an individual) acknowledge that it overpaid. I don't think this is worth debating. 

The next bit would be assessing whether a counterfactual where the two companies never merged would've been better for AMD. 
The evidence that I'm putting forth here is that AMD was not able to execute well for YEARS due to a lack of positive cashflow. This situation was acknowledged as a risk in AMD's financial statements preceding the merger, and after the merger it was noted as a reality. 
Some side effects include NOT having a viable Bulldozer replacement (in theory it should've been evident by 2008 - 1 year before BD's initially planned launch that something was wrong; because of a lack of cash AMD was forced to push BD ahead and couldn't put resources towards its replacement, with more funds and more focus[Rory Reed, AMD's former CEO went on record noting that the years following the merger had a lot of disorganization and people sitting idle along with graphics and CPU divisions doing things that hurt the other], a BD alternative could've been put forth 3-5 years AFTER it was recognized that BD was not a viable path forward[so 2011-2014 or so]). The evidence I have for the BD launch being planned for 2019 came from Anandtech in 2007. 3-5 year CPU development cycles are the norm and I'm using that as a given, go scowl around Quora for CPU architects or go dig into press releases if you want evidence of that. 

The only way that my conclusions fall apart are if having substantially more cash, fewer distractions (organizational issues related to the merger, mass layoffs, etc.) and a shifted plan in 2008/2009 for the CPU roadmap wouldn't have generated an effective, market shifting product. I think a "good enough" product would've been probable enough. 

It is possible that the alternative to Zen (plus its descendants) wouldn't be as good as Zen (over say 2010-2025 in terms of FCF or EBITDA). It is possible that the future of the GPU section of the existing business will be HUGE. 

I am acting on the assumption that the fusion product lineup was not that great (llano fell flat and AMD had a huge backlof of inventory they couldn't sell, the BD-based APUs never sold that well, etc.)


I would argue that my assumptions are not aggressive and are all reasonably likely. The non-assumptions are backed up by sources. 

---

meanwhile your claims end up relying on MUCH stronger implicit assumptions (ultimately that RTG and/or merged CPU-CPU products will have such a huge future impact so as to offset the bloodbath which was 2007-2017). Note quite this strong but...  https://aeroleads.com/blog/bootstrapped-startup-profitability/dilbert-by-scott-adams-9/

 

 

---

 

  

If you want to go off of the IQ contest... I had a near-perfect GRE score (top vigintile in each section and presumably top 1% overall among people going for MS/PhDs) and am at one of these places: https://www.technologyreview.com/lists/companies/2017/intro/#kite-pharma
Multiple litmus tests essentially came to the same result, largely independently. I'm either very good at finessing my way into systems that generally keep people with my background out (non-affluent upbringing, blue collar family, etc.) and/or there's an underlying common factor.

Degree doesn't imply good judgement. It does correlate with it well though and it is valid to cite it as supporting evidence for it.
 

 

So... yes and no. Investment bankers are REGULARLY pitching mergers and acquisitions and management consultants are regularly (or at least used to be circa 2007) suggesting M&A as a way to create efficiencies. A lot of this is basically BS to accumulate fees. For some context I know a few people who've worked on wallstreet (IBD, ER, etc.) and considered that path for myself. Same for management consultants at places like McKinsey.

And.. most mergers fail because execution is REALLY hard (and many of the assumptions going into a deal are overly optimistic - bankers are paid a percent of valuation on deal-close, they're incentivized to get a deal to close at a high price). 
https://www.forbes.com/sites/forbescoachescouncil/2019/06/24/most-mergers-fail-because-people-arent-boxes/#13f31ce55277

For what it's worth I worked at an F100 when it went through a $70BN deal and... I rolled my eyes and decried it as stupid. Many of the reasons I thought it was stupid are cited in the article above and I lived through the realities of combining two organizations. 

Why isn't the near-past relevant to the future? That strikes me as a very strong assumption. 

How does that argument hold much water? Yeah anyone would know that if they had the ability to buy the company for cheaper it would be better than buying it for more. That's common sense but the problem is would the buy be possible for less? And would they be better off not buying if they couldn't? 

[Brooksie359]

1 hour ago, comander said:

They'd have been better off doing no deal than doing the deal at the price they did it at. 

AMD probably would've had to have have paid a premium no matter what but... let's just say instead of a 70% premium if they'd done only a 20% premium (so 1.5-2BN cheaper) that would've helped and it might've been viable.

Again that is not certain that if they hadn't acquired ATI that they would be any better off than they are now. We can speculate all we want but at the end of the day it is all speculation.