Valuing Amazon is an extremely challenging exercise, mostly because of their high reinvestment. Who are we to make a DCF model in which we assume Amazon will stop reinvesting the majority of cash flows over the next 10 years?
Moreover, a 10-year model is extremely conservative in my opinion: because the company has never optimized for returns, a 10-year period of excess returns seems too short, which means you're missing a lot of value in your model.
Totally agree that valuing Amazon is an extremely complex exercise — their high reinvestment rate, lack of transparency, and the fact that they operate across such different business models (retail, AWS, ads, Prime, etc.) make it nearly impossible to model precisely. But that’s also what makes it such a fascinating challenge.
Regarding your point on the 10-year excess returns period: I think you're absolutely right that it's likely too conservative for a company like Amazon. That’s why we chose to model a return on capital (ROC) above our WACC for the terminal value, based on our belief that Amazon will continue to generate excess returns over that timeframe, even if they don’t explicitly optimize for profitability.
At the same time, we’re convinced that trying to predict Amazon’s exact growth rate or margins in year 15 would be overconfident — we simply don’t feel qualified to make that kind of precise forecast. Our approach shifts the uncertainty, but doesn’t eliminate it. And we’re conscious of the fact that it has its own drawbacks.
Really appreciate your thoughtful comments — always great to hear your perspective!
From my perspective, you need to capitalize and amortize R&D, just like you need to do the same for at least a portion of SaaS companies' marketing expenses--because the benefits of such investments will last far beyond a year.
Secondly, if you build up an operating margin forecast, making assumptions on what mature operating margins can be for AWS, advertising and retail, along with reasonable growth rates for each of those over the next 5-10 years--I think you'll come up with an operating margin forecast well north of 16%.
Totally agree with your first point — capitalizing and amortizing R&D is definitely a valid and logical approach, especially when the benefits of those investments extend well beyond a single year. That’s why we also included a valuation comparison with Damodaran’s model, where R&D is capitalized. We strongly considered doing the same, but ultimately decided against it in this case.
Our reasoning was shaped by what we’ve seen with other big tech names like Alphabet, where side projects (e.g., Waymo, Verily, etc.) often have huge long-term potential but don’t generate near-term ROI. Because Amazon similarly invests in moonshots and long-horizon innovations, we choosed the conservative approach.
As for your point on margins — again, we completely see where you’re coming from. If you break down Amazon’s segments and assume mature operating margins for AWS, advertising, and even retail (at scale), you can justify an operating margin north of 16%.
That said, we chose to stay conservative here too. Amazon has historically reinvested heavily into new initiatives, many of which are funded by profits from their high-margin segments. These expansions often pressure margins in the short term and can even lead to temporary losses — but have proven to be extremely valuable over the long run.
We believe this cycle of reinvesting in new business areas is far from over. Over the next decade, we fully expect Amazon to launch entirely new verticals, and that dynamic is already priced into our forecast: we do model a margin increase, but not an aggressive one, precisely because we think continued reinvestment will remain part of the Amazon DNA.
Appreciate your thoughtful input — really sharp points!
Yes.
Valuing Amazon is an extremely challenging exercise, mostly because of their high reinvestment. Who are we to make a DCF model in which we assume Amazon will stop reinvesting the majority of cash flows over the next 10 years?
Moreover, a 10-year model is extremely conservative in my opinion: because the company has never optimized for returns, a 10-year period of excess returns seems too short, which means you're missing a lot of value in your model.
Just my thoughts
Totally agree that valuing Amazon is an extremely complex exercise — their high reinvestment rate, lack of transparency, and the fact that they operate across such different business models (retail, AWS, ads, Prime, etc.) make it nearly impossible to model precisely. But that’s also what makes it such a fascinating challenge.
Regarding your point on the 10-year excess returns period: I think you're absolutely right that it's likely too conservative for a company like Amazon. That’s why we chose to model a return on capital (ROC) above our WACC for the terminal value, based on our belief that Amazon will continue to generate excess returns over that timeframe, even if they don’t explicitly optimize for profitability.
At the same time, we’re convinced that trying to predict Amazon’s exact growth rate or margins in year 15 would be overconfident — we simply don’t feel qualified to make that kind of precise forecast. Our approach shifts the uncertainty, but doesn’t eliminate it. And we’re conscious of the fact that it has its own drawbacks.
Really appreciate your thoughtful comments — always great to hear your perspective!
From my perspective, you need to capitalize and amortize R&D, just like you need to do the same for at least a portion of SaaS companies' marketing expenses--because the benefits of such investments will last far beyond a year.
Secondly, if you build up an operating margin forecast, making assumptions on what mature operating margins can be for AWS, advertising and retail, along with reasonable growth rates for each of those over the next 5-10 years--I think you'll come up with an operating margin forecast well north of 16%.
Totally agree with your first point — capitalizing and amortizing R&D is definitely a valid and logical approach, especially when the benefits of those investments extend well beyond a single year. That’s why we also included a valuation comparison with Damodaran’s model, where R&D is capitalized. We strongly considered doing the same, but ultimately decided against it in this case.
Our reasoning was shaped by what we’ve seen with other big tech names like Alphabet, where side projects (e.g., Waymo, Verily, etc.) often have huge long-term potential but don’t generate near-term ROI. Because Amazon similarly invests in moonshots and long-horizon innovations, we choosed the conservative approach.
As for your point on margins — again, we completely see where you’re coming from. If you break down Amazon’s segments and assume mature operating margins for AWS, advertising, and even retail (at scale), you can justify an operating margin north of 16%.
That said, we chose to stay conservative here too. Amazon has historically reinvested heavily into new initiatives, many of which are funded by profits from their high-margin segments. These expansions often pressure margins in the short term and can even lead to temporary losses — but have proven to be extremely valuable over the long run.
We believe this cycle of reinvesting in new business areas is far from over. Over the next decade, we fully expect Amazon to launch entirely new verticals, and that dynamic is already priced into our forecast: we do model a margin increase, but not an aggressive one, precisely because we think continued reinvestment will remain part of the Amazon DNA.
Appreciate your thoughtful input — really sharp points!