das hat mal echt Spaß gemacht zu lesen; eine visionäre AMD-Analyse bis 2025
so here is what i see as a "not unrealistically bullish" scenario for AMD stock in the next 5 years:
in 4 years (2024) the TAM is estimated to be $125B split between INTC, NVDA and AMD. this seems conservative to me given that it is already $100B, but lets go with it.
in 2024, AMD will be on Zen5/6. who knows exactly what it will be but it almost certainly will be far superior to any Xeon or Core.
dont mention the threat from ARM or RISC-V or quantum computing. the war will be over by then.
by 2024, Intel will be going through or have gone through tremendous upheaval. their process will likely be at best 1 generation behind (7nm-200mtr/mm2) and possibly up to 3 generations behind (14nm+++++-40mt4/mm2) vs TSMC 3nm-300mtr/mm2. who knows. by that time they might have spun-out their fabs or licensed tech for TSMC. they would not have the profits to sustain all their diversification forays over the years and those would have been divested.
by end of 2020, it is estimated that AMD will have the following rough market shares: 1) server 10-15%; 2) desktop 30-35%; 3) laptop 20-25%; 4) GPU 25-35%.
NVDA is stuck with an inferior SS process. it purportedly has alienated TSMC. it is rumored that RDNA2 will outperform Ampere. with the new consoles games will be optimized for RDNA2. AMD could easily take over gaming especially if they offer bundled pricing for CPU and GPU together, this they should promote aggressively. take no prisoners lisa. get aggressive on the marketing side.
i believe AMD is uniquely and solely positioned to exploit x86 and GPU integration. for compute workloads swapping data in and out of vram takes ages and avoiding this step would be revolutionary. NVDA cannot do this for obvious reasons and can only try to market ARM CPUs which is an uphill battle and will suck up tremendous R&D dollars. INTC has a hot mess GPU effort and a collapsing Fab. they still be stuck in restructuring mode probably for the next several years. their board cannot fire everyone as who will want to replace them? keller came and left. (he basically wanted INTC to publicly acknowledge their fab was a disaster which INTC is not ready to do understandably.) raja is soon to follow as he lost his primary ally. murthy is the last viable option and may have just orchestrated his ascension. it is a hot mess. sad and is a metaphor for the usa vs asia. but i digress.
lets say over the next 24 months Lisa invests heavily and is PERSONALLY FOCUSED on the turnaround of the software side (just like she has done with CPU and Radeon) and makes AMD software first class, including compute and AI. just like she brought in heavy hitters to focus on hardware, she can and should do the same for software, AI and ML.
so lets fantasize a bit and say that by 2025 the world has flipped and AMD is top dog across CPU and GPU.
and lets assume AMD has 70% of the desktop/laptop markets, 60% of the server market and 60% of gaming GPUs and 35% of compute GPUs. yes it might be hard to believe possible. but 5 years is a very long time and 5 years ago it was 2015 and everyone doubted things like Tesla and Netflix and no one even heard of Shopify. things change dramatically at technology turns. we are at such a time in semiconductors. intel was not paying attention to the road ahead and completely missed the turn.
so i plug these numbers into my trusty model. what do i get? $48B in revenues.
the next important assumption is net margin. i assume a 30% net margin (equivalent to both INTC and NVDA currently), this is $14B in net income of $12 EPS.
those of you who do not think 30% net margin is possible, you need to write a model and examine R&D expense and the opex leverage you witness when a company transitions from breakeven (AMD recent history) to extremely profitable (AMD in the future). R&D expense is the magic lever for net margin expansion. it is also the temptation which creates distraction. rather than blow R&D on wild adventures like INTC lisa should remain disciplined and buy back stock with any excess cash flow.
$12 EPS * 25 PE (yes you can disagree with this) = $300 stock price or a 33% CAGR.
think it cant happen? NVDA was $20 5 years ago and $400+ today. and the GPU market is MUCH smaller than CPUs. the market just needs to wake up and understand what is happening. it may not even take 5 years!
bottom line: AMD is sitting in an amazing position to dominate CPU and GPU. the market has been VERY slow to see this and accord it proper value (while it is gives elon $100B market cap value for autonomous driving lol). INTC has a technological problem that cannot be solved any time soon. even if they completely fix their process by 2023 they will be 1 node behind TSMC and have inferior uarch to AMD. TSMC is investing $15-20B PER YEAR in bleeding edge fabs. no way INTC can catch up barring some disaster to TSMC. NVDA lacks CPUs to defend its GPU sales and is stuck with an inferior SS node.
this is how i see it. as lisa keeps repeating "the best is yet to come" i really think she is being sincere about this and is telegraphing you a hint. if you find flaw in my analysis please let me know. the only thing i have not done a lot of work on is whether TSMC can provide the wafers for these types of market shares as that is very hard to project given their aggressive capex. i can only assume TSMC and AMD are in communication about future needs and AMD can spread their SKUs across 3nm/5nm/7nm and still outcompete intel at every price point. and even if i am wrong by half, it still represents a 16% CAGR. so if you are contemplating selling here at $69 you might want to consider how much you could be leaving on the table. the good thing about semicondutors and investing is that development takes so long to materialize that the future was written 5 years ago. and here we are.
www.reddit.com/r/AMD_Stock/comments/...potential_tldr_300_5y/
Zurück
