The Saving Advice Forums - A classic personal finance community.

Long on AMD

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Long on AMD

    Bored so I want to make a case for going long on AMD.

    This stock has rallied 20% the past 2 days after a devastating 26% drop 2 months prior. Company market cap is at 11 billion. Their main competitors being Intel and Nvidia which are at market cap of 162 billion/94 billion respectively.

    My case, AMD in the long run at least till 2020 will be a multibagger stock. There are a lot bears out there for this stock due to their past mistakes and not so stellar balance sheet...not to mention the company is currently unprofitable.
    I also feel some of these analysts doesn't understand the technology and AMD's master plan.

    So here's my case for AMD

    1. Infinity Fabric, the AMD pioneered technology that can glue CPU and GPU together with low latency is a game changer to Moore's law.
    • AMD cannot compete with Intel in the R&D department and cannot afford to introduce year after year, huge CPU dies with extra cores to compete which are extremely expensive with every SKU. AMD instead spent the last few years developing the infinity fabric, which allows them to glue multiple CPU cores and GPU cores together with almost perfect scaling. This means they can use their desktop low end processors and just glue them together with no extra R&D to make high core count server grade processors for peanuts. They can afford to be extremely competitive with pricing and performance thanks to this tech.
    • Their GPU sales will always be behind Nvidia even if they introduce something ridiculously awesome. So to pour massive amounts of R&D to compete, they again will use infinity fabric to double their GPU performance as die shrinks year after year.


    2. They have 0% market share in data servers...a 22 billion dollar revenue stream going 100% to Intel.
    • The new EPYC processors has better performance and cheaper than Intel's current offerings.

    • Single socket EPYC server processors can support 128 PCI-E lanes in which Intel cannot match even in their upcoming offering. THIS IS THE KEY! 128 PCI-E = More terabytes of storage per server unit without needing to buy more processors essentially decreasing cost, space, heat, and power consumption!!!! 128 PCI-E lanes also = more GPUs can be connected to a single server for deep learning or high end animation work. To people who think AMD may just take a little market share of the server market is high, this is what companies are salivating over!
    • AMD already partnered up with Drop box, 1&1, Baidu, Dell, Microsoft, and Samsung. These are not your no name customers. They currently love what AMD has to offer.


    3. Intel F-up
    • AMD will release a high end desktop CPU called Threadripper next month, which is just two Ryzen cores glued together. They will offer this part for $850 for a 16 core/32 thread part.

    • Intel's response is universally hated by tech enthusiasts(check out any tech youtube channel and you'll see them bashing intel). The new X299 platform is rushed due to AMD's Ryzen comeback which introduced broken motherboards and chips running extremely hot. Intel's newest offering to combat Threadripper also support less PCI-E lanes and ends up being 2x more expensive. They will also be 5 months late. In the tech world, being 5 months late is an eternity.
    • Intel's 7 years of stagnation and price gouging due to lack of a competitor left many tech enthusiast sour. Many are switching to AMD just on principle alone. Of course the mainstream buyers will still prefer intel..but AMD's low pricing is hard to ignore.

    • Without something like infinity fabric, Intel has to spend more money manufacturing high core count CPUs with lower yields. This prevents them from lowering the price of their high end chip without losing massive amounts of margin. AMD just needs to produce 4 perfect ryzen 7 processors and glue of them together..Intel needs to produce perfect 22 core or 32 core processors.


    4. Dr. Lisa Su as the CEO
    • Dr. Su's road map to profitability is extremely promising. The infinity fabric will allow them to scale well into the future without high manufacturing cost and R&D.

    • She is extremely focused on revenue. You can tell by her abandoning the race for superior GPU for enthusiast this generation because high end GPUs are money losers compared to Nvidia..this is a space they will never win even with a blockbuster hit. Nvidia just has superior brand awareness. Instead she focuses on mainstream GPUs, giving the masses cheap VR and 4k capable GPU without the hefty price tag. The newest Vega GPU will be a flop among gamers(it will not beat Nvidia's 1080Ti). Their determination to develop HBM2 (high bandwidth memory) tells me that their goal is to use infinity fabric on Vega to produce Navi with ease and can double performance with low R&D cost after the next die shrink. This is their master plan with GPUs..so people looking at the short term will definitely be disappointed.

    • She knows that the money is in the developers and data centers. She prioritize them first by producing high grade server CPUs with features companies crave for, and graphics chip tailored more to animators and deep learning vs enthusiast gamers.


    For all the reasons above, I believe AMD will one day become a 50 billion dollar company...eating away Intel's market share. They will also dominate the console market (like current gen) because Nvidia will just be too expensive for console makers to use. I don't see them taking away Nvidia's market share too much since that's a company more prepared for AMD's comeback. Time will tell with that one.

    Disclaimer: I have 20k invested with AMD.
    Last edited by Singuy; 06-22-2017, 11:11 AM.

  • #2
    I guess we'll have to see how the gaming community responds to the Intel I9.
    Brian

    Comment


    • #3
      Originally posted by bjl584 View Post
      I guess we'll have to see how the gaming community responds to the Intel I9.
      I can tell you how. At first everyone will try to prove that i7 is better for gaming than i9 (as it happened in the case of i5 and i7) and later they will move on and embrace i9.

      Comment


      • #4
        The enthusiast community right now dislike the I9 due to bios instability and insane temps(almost 100c) when overclocked. Also it's drawing 50 watts more power than advertised(perhaps can be fixed with bios). The only processor that provides gamers with 44 PCI express lanes are the newly released 7900K. So for people who wants to do SLI/crossfire for 4k gaming and future proofing is SOL with any of Intel's cheaper processors for the x299 platform. Each high end GPU would like to have full 16 PCI lanes for max performance. All of intel's sub thousand dollar processors are gimped with only 28 PCI lanes. The 7900k is retailing at $999. Yes the 7900k is faster than AMD's 300 dollar processor at 1080p gaming by about 20%..but who buys a 1k cpu and not do 4k gaming? When it comes to 4k gaming, PCI lanes is MORE impt since you are GPU bound. Even AMD's 250 dollar chip is just a few % points slower than intel's 7900k here.

        AMD is offering 64 PCI-lanes for ALL of their Threadripper processors in which the most expensive one will most likely retail at $849. So there may be a $550 part, a $650 part, a $750 part and etc and they will all have access to the full amount of lanes for SLI/Crossfire.

        Also the gaming community is not really important (PC game sales are 1/3 of consoles). AMD won the console bid due to them being cheap.

        The money is really at the enterprise end..pc sales have been declining. Data centers is growing at an exponential rate however since everything is going mobile(youtube, netflix, could based storage, GPU deep learning, e-
        commerce). At the end of the day, whoever holds the efficiency crown wins the game. AMD's 128 PCI-lanes and low power draw as a single socket processor looks to be the efficiency champion. If a company were to go Intel, they will need 2x more server racks just to hold the same storage and GPU cards as AMD's. Going AMD is almost a no-brainer until Intel gives them the I/O they need. Either Intel likes to price gouge or they just have no way of providing the business end with the PCI-E lanes they need with their lower end server chips. Unlike AMD who will provide you with full 128 lanes at even their lowest end 400 dollar server chips, Intel's offering is 1/4 of that for some reason.

        Talking about mobile, Ryzen 1700 cpu mobile edition is a full fledge desktop 8 core CPU in a notebook. ASUS's just went on pre-order. They also published some benchmarks which is 100% faster, completely crushing the current high end intel mobile quadcore CPU in multi-tasking workloads (cinebench 15).

        So here's AMD's master plan.

        This year, take up market share in the data server space and mobile laptop space. Release a GPU that is faster than Nvidia's GTX 1080...which means AMD loses here but price it aggressively.

        In 2 years when the die shrinks, their ryzen 2 will have better IPC and higher clocks. They will most likely fuse 2x more cores together vs this year. Game companies will start utilizing more than 4 cores..most likely 8 cores(since the AMD jaguar CPU in the xbox/ps4 are already 8 cores..game companies will be force to utilize all of them for the most performance). All the console ports are already optimized for AMD graphics(thanks to their monopoly with Xbox/PS4).

        Their data chips will mostly have 64 cores at the high end, fusing 8 ryzen 2s together, giving companies 256 PCI-LANES!!

        They will fuse 2x GPU together for their high end, 2x performance which will likely beat or match Nvidia's offering but at low R&D and manufacturing cost.

        They will also develop a very high end APU with Ryzen cores and GPU cores fused together..perhaps the first era to tap into GPUs for CPU computes seamlessly with certain applications.

        While everyone is trying to build massive dies to compete..AMD just needs a die shrink and glue everything together. Infinity Fabric is the 800lb gorilla.
        Last edited by Singuy; 06-23-2017, 09:37 AM.

        Comment


        • #5
          Almost seems like the (most recent) time to get in was between October 2014 to right around February 2016.

          Are you just chasing performance at this point? How much higher can it go?

          Looking at the maximum life time stock price chart (~1980 to present) and eyeballing it, it seems like $20 has been a threshold (not taking into any stock splits), with a few spikes to $40 over the years. I'm ignoring the massive spikes in early 80's ($65) and 2000 ($92).

          Comment


          • #6
            I look at market caps and what value the company provides. Intel is at 180 billion, Nvidia is at 90 billion..AMD is at 13 billion.

            AMD has a negative balance sheet and is bleeding money. Analysis are still bearish on AMD(Goldman downgraded them just 2 weeks ago).

            They have a revenue of 4.6 billion..and if they take 20% of the data center shares, we can expect an additional 6-10 billion in revenue.

            Data centers are not glowing at the same rate they were 20-40 years ago. Youtube was not even a website 25 years ago.

            There have been reports of almost 100% yield on the manufacturing side.
            So we have a company with high end server parts that are better, can have almost 100% yield at the manufacturing side, and can scale in the future for cheaper...and also has no where to go but up in the highly lucrative data center side...

            Their 13 billion dollar market cap is just the beginning.

            Comment


            • #7
              Originally posted by Jluke View Post

              Looking at the maximum life time stock price chart (~1980 to present) and eyeballing it, it seems like $20 has been a threshold (not taking into any stock splits), with a few spikes to $40 over the years. I'm ignoring the massive spikes in early 80's ($65) and 2000 ($92).
              It's probably wise to ignore some of the massive spikes in 2000 due to the tech bubble, but 2000 was also the year AMD beat out Intel on the CPU front. Their Thunderbird CPU was faster than Intel's Pentium 4s and they never caught up until their core2duos. AMD was smoking hot for a good 5 years with no competition, releasing the first 64bit cpu at this time and also a high performance dual-core X2 CPU which CRUSHED Intel's Pentium Ds.

              Comment


              • #8
                Btw, I would like to encourage others to make deep analysis on stocks you are passionate about and then having an open discussion about it. I believe this can bring some value to this section of the forum.

                Comment


                • #9
                  AMD is better.

                  Comment


                  • #10
                    The problem for AMD are that:
                    • there are fewer Gamers than there are Regular People, and
                    • most programs need higher MHz on single-threads, and
                    • most programs need good integer performance.


                    Sadly, those are all bad trends for AMD. (I'm writing this on an AMD system, but a 5 year old AMD system. It's the RAM and disk storage that I keep upgrading, because the CPU and video card are Fast Enough.)

                    Comment


                    • #11
                      I'm not seeing anything here that would make me want to go long on AMD:

                      Comment


                      • #12
                        Bears have no logical reasons except they failed to buy cheap till now.. very bullish in coming weeks..

                        Comment


                        • #13
                          I am bullish on AMD, but only tenatively so....

                          Threadripper is a home run, but without a doubt, Intel will counter-punch hard on the next gen. Even now, current gen Intel CPUs are still quite good, and energy-efficient to boot.

                          Vega is on par with nVidia's GTX 1080, but not the Ti's. And only just.

                          And of course, you can hardly find any of their current gen Radeons because cryptocurrency miners have literally bought them all... to the point that nVidia realized their folly and created their own line of GPUs dedicated only to mining.

                          Now, the question is, how much of all this fairly obvious and public news have already been baked into their stock price, and how much growth potential is there from this point out? I don't know, but I know I will be watching.

                          Regardless of where you stand on this stock, I am glad AMD is still fighting the good fight, as competition will only benefit us customers.

                          Comment


                          • #14
                            I support the whole bullish case here. I joined their ranks at $4.00 back at 2016 and still on board. Definitely long AMD here.

                            Comment


                            • #15
                              A couple of big server deals just announced today. MI25s going to Baidu and Eypc being used at Tencent/JD.com

                              AMD is ready for another lift off again after the brutal beating post earnings thanks to the tech sell off and Trump's mouth going off on NK.

                              Comment

                              Working...
                              X