#88 | Amazon Deep Dive | Inside The Everything Store Flywheel

If you think Amazon is just an online retailer, think again. From dismantling physical friction with the internet to building an infrastructure that makes everything from cloud computing to logistics look effortless, Amazon’s true power lies in its hidden, interconnected engine. Jeff Bezos’s long-term vision transformed the world’s shopping habits, but what’s next? How does Amazon defend its dominance, and what vulnerabilities could threaten its empire?

Transcript

SPEAKER_00 0:04

Welcome back to the Dutch Investors to a TDI deep dive into what is arguably the most complex and misunderstood economic engine of today's era. If you're listening to this, you likely interact with this company almost every day, perhaps without even realizing it. Whether you're ordering a replacement part for your kitchen, streaming a movie on a rainy Tuesday, or even just browsing a website that happens to be hosted on their servers. You are feeding a flywheel that has been spinning for decades. But today we are not looking at the retail website or the livery fans. Today we'll be taking a closer look at the Everything Store, better known as Amazon.com. Let's jump in. To understand why Amazon even exists, we have to start with a problem that is as old as civilization itself. Friction. For most of human history, commerce was defined by physical limitations of geography. If you lived in a small town, your choice of books, tools, or clothing was limited by what a local shopkeeper decided to stock on their shelves. You paid a premium for that convenience, and you lived with a restricted selection. Historically, the relationship between a retailer and a customer was one of gatekeeping. The retailer held the inventory and the customer had the need. And this created a massive asymmetry of information and access. Amazon was founded by Jeff Bezos on the radical idea that his friction could be dismantled through, you guessed it, the internet. But it wasn't just about selling things online. That's a common misconception. The fundamental need this company solves is the easy access to the long tail of human production. Why should a reader in a small village have less access to knowledge than a scholar in Manhattan? Jeff Bezos realized that if you could aggregate enough demand in one place, you could justify the massive capital expenditure required to build a logistics network that makes the friction of distance disappear. In the current era, this problem solving has shifted from physical goods to digital infrastructure. Just as small towns lacked bookstores, startups historically lacked the capital to build data centers. And by creating AWS, Amazon Web Services, the company solved the problem of computing friction, allowing anyone with a credit card to rent the same level of scale that previously was only reserved to the world's largest governments and corporations. Amazon exists to be the world's infrastructure, a role it performs by obsessively removing the costs of getting things done for the end customer. And before we dive into the business model, we have to go back to the origin story. The history of Amazon is a study of the Lindy effect, the idea that the longer a non-perishable thing has survived, the longer it is likely to survive in the future. Amazon has survived the dark days of the dot-com crash, the dot com bubble, multiple recessions, and a global supply chain collapse, each time emerging larger and more integrated. Founded in 1994, the company appeared to be just another dime of a dozen online retailer. But the most important moment that defined Amazon's success was not a specific product launch. It was a cultural decision, made during the 1997 and 2001 era. While all the other tech companies around the world were chasing quarterly profits, squeezing as much money out of the customer as possible, Amazon's management made it very clear in their very first shareholder letter that they would focus on the long term. Now, anyone can say that, but they actually did it. They essentially told the market we will be misunderstood for a long, long time, and we will be okay with that. If we had to point to one moment where Amazon's DNA was encoded, it was in 2001. The bubble had burst and capital was scarce. Everyone and their mother suggested that Amazon should raise its prices to show a profit and survive. Instead, Bezos and its team did the opposite. Inspired by the scale economy shared model of companies like Costco and WISE, they lowered prices. They realized that if they could pass their efficiency gains back to the customer, they would create a loyalty loop that no competitor could break. And this was the moment the Amazon flywheel was set in motion. The second important moment occurred in 2006, with the launch of AWS. Historically, many believed AWS was created to sell excess capacity from Amazon's retail service. But this is a myth. AWS was actually a solution to an internal problem, and as the company grew, its engineering teams were spending more time coordinating with each other than actually innovating. They were reinventing the wheel for every new project. So management decided to turn their internal tools into standardized services. By hardening these tools so their own developers could use them efficiently, they created a product that the rest of the world desperately needed. And today, that internal solution provides more than half of Amazon's operating profits. So how does Amazon actually make money? Now that we understand the history, we can take a look at the engine. If you look at the top line, it looks like a retail business. But if you look at the bottom line, it's more of a technology and services company. Amazon operates a multi-sided platform. Think of it as a digital landlord that also happens to run a giant store on the ground floor. Historically, the company made money by buying products at wholesale and selling them at a markup, the 1P or online stores model. Today, that is just bait for the trap. The real money is made in the services that sit on top of the retail traffic. You could say there are three revenue pillars. Currently, about 62% of items sold on the platform come from third-party sellers. So that are sellers, which is different from first-party sellers. A first-party product is something Amazon sells, and a third-party product is something someone else sells. Amazon provides them all, and the sellers pay rent. But it's a little more than just rent. Sellers pay a referral fee, usually around 15%, a fulfillment fee to use Amazon's warehouses, and increasingly they pay for advertising, just to be seen by customers. And today Amazon is capturing roughly 50% of every dollar a third-party seller makes. The next pillar is the cloud infrastructure, AWS. And this is the high-margin heart of the business. AWS provides the servers, databases, and machine learning tools that run everything from Netflix to the CIA. It's a pay as you go model, which sounds simple, but it creates massive switching costs once a company's entire workflow is built on Amazon-specific APIs. In 2025, AWS generated$128.7 billion in revenue, still growing 20% year over year. Now the third pillar, and the fastest growing part, is the advertising segment. When you search for, I don't know, coffee maker, and you see a sponsored result, that is pure profit for Amazon. Because the customer is already there to buy, the advertising is incredibly effective, and this revenue has exploded, generating over$68 billion in revenue annually and growing fast. Currently, around 37 to 38% of revenue comes from online stores first party sales, around 24% comes from third-party sellers, which is high margin, while first party sellers have more volume but are lower margin. AWS accounts for about 18% of revenue, which is a very high margin and massive switching costs. Advertising, around 10% currently, but it's the fastest growing one. And obviously, you have the subscriptions like Prime, which is recurring revenue, about 7% currently, with over 240 million members globally. You can imagine the logistics side as a giant heavy concrete slab. It's very expensive to build and to maintain. But the advertising and AWS segments are the advertising and AWS segments are the high-pressure jets that lift that slab off the ground. Without the retail volume, the advertising probably wouldn't exist. And without the engineering challenges of retail, AWS wouldn't have borne. They are interconnected and interdependent parts of a single organism, in this case Amazon. Now that we understand how they built the castle, let's look at the modes protecting it. This is where we need to be skeptical. Because is the mode really as wide as management says it is? The most profound competitive advantage Amazon has is what we've already mentioned skill economy shared. In a normal business, which Amazon obviously isn't, as you get bigger, you might keep the extra profit for yourself to give to shareholders. But Amazon does the opposite. They take the savings from their massive scale and use them to lower prices even further. This creates a barrier to entry that is almost impossible to hurl. If a new competitor wants to challenge Amazon, they don't just have to be better, they have to be cheaper while having smaller scale, which is a mathematical nightmare for any startup. Another massive mode is the logistics mode. Historically, Amazon was a customer of UPS and the postal service. In the current era, Amazon has become the logistics industry itself. In 2025, Amazon Logistics officially became the largest domestic parcel carrier by volume in the United States, delivering a staggering 6.7 billion packages. So why is this a mode? Well, logistics is a density game. The more packages you have on a single street, the cheaper it is to deliver each one. Because Amazon controls the order and the delivery, they have the ultimate density. Legacy players like UPS are actually retreating from this last mile commodity, delivery to focus on more lucrative B2B in healthcare logistics because they cannot compete with Amazon's unit economics. However, if we look closely, we can see some cracks. The retail mode is being attacked by a new breed of you could call it dopamine commerce. Chinese companies like Timu and TikTok Shop are bypassing the traditional retail model entirely. TikTok Shop moved$90 billion in merchandise globally in a single quarter in 2025. So while Amazon is a safety reflex for consumers, Timu has an addiction reflex. Timu's ability to use tax loopholes has allowed them to ship unbranded goods at prices Amazon often cannot match. Which is crazy to think about. Amazon's search-based dominance could be in trouble in the future. Furthermore, in the cloud battlefield, the multi-cloud trend is a real thing. Historically, companies were happy to be all in on AWS. But today, companies and CTOs are nervous about being locked into a single provider. They are increasingly using Microsoft Azure or Google Cloud for specific tasks to avoid dependency. While AWS growth has re-accelerated, the market is moving towards a mature oligopolis. Now this can also be a good thing in the long run, since if there's just a single player, they could get fined for being a monopoly. So who knows, right? We'll see. A company, ultimately a collection of people, and the incentives that drive them. And Amazon's culture is legendary, but is it sustainable? Especially considering its current size. Jeff Bezos famously said that it is always day one at Amazon. Day two is stasis, followed by irrelevance and death. To prevent day two, the company uses a set of 16 leadership principles that are surprisingly influential in daily operations. For example, frugality is why their office desks were historically made of cheap doors. And bias for action is why they are willing to fail at things like the Firephone to find a success like Alexa or Prime. A couple other examples are be obsessed about the customer, learn and be curious, deliver great results, think big, and hire and develop the best. Just to name a few. So let's briefly touch on management and skin in the game. The transition to CEO Andy Jesse was a logical move. He was actually architect of AWS. Historically, Amazon has been one of the most shareholder-aligned companies in terms of pay. Jesse's base salary is just$175,000. His real wealth comes from over 2.2 million shares that fast over a 10-year period. Jeff Bezos still owns roughly 8-9% of the company. And 99.5% of compensation for Andy Jesse is linked to performance. There are no cash bonuses and they are entirely stock incentivized. And bonuses are based on the stock performance, which usually is not a good thing, but in Amazon's case, it forces them to keep innovating and keep day one active. So we don't mind it, and looking back, they've done great so far. But we have to ask, can this culture remain as the company is so extremely large now? Recent periods have seen large-scale layoffs and employee complaints about a secretive low performance list. Some observers argue that under Jesse, the company has focused more on trimming the fat and operational excellence than on the messy invention that defined the Bezos era. If Amazon becomes just another efficient company, it may lose the magic touch that allows it to enter and disrupt new industries. Definitely something to keep track of. Now, let's talk about risks and opportunities. What could kill Amazon in say 10 years from now? And can it still become a 10-bagger from here? Now, for the bear case, would say there are three main killers. First, the antitrust breakup. The US government is increasingly skeptical of Amazon's dual role. You cannot be the referee, the platform, and the player selling your own Amazon basics products at the same time. In 2025, Amazon already faced special charges related to FTC legal settlements. 2. The death of search. If generative AI changes how we find products, the search bar on Amazon.com becomes less valuable. More than 300 million customers already used the AI assistant Rufus in 2025, showing there is a real shift. 3. The logistics bloat. Capital expenditures in 2025 have reached a staggering$131 billion. If consumer spending slows, they could be left with a massive physical network that is underutilized, turning their greatest strength into a fixed-cost anchor. On the flip side, the bool case or the tailwinds are massive. Amazon Leo, also known as Project Kuiper. This satellite internet project was rebranded to Amazon Leo in late 2025. With over 200 satellites already in orbit and surface beginning in five countries in early 2026, it aims to plug broadband into the high-margin AWS engine. Another massive tailwind is healthcare. Amazon's acquisitions of one medical and the pharmacy push could turn a$4 trillion industry into a high-efficiency digital product. Lastly, advertising. So how does the market value a company that refuses to show gap profits? Because historically valuing Amazon on PE ratios has been a fool's errand. If they wanted to show a PE of 10, they could simply stop investing in new warehouses and satellites tomorrow, but that would kill the flywheel. The market has been willing to pay a premium for Amazon's operating cash flow. So here's our verdict of Amazon. We still think Amazon is a high-quality business, in the truest sense. It still has a culture almost impossible to replicate, and a logistical mode that has now surpassed the century-old incumbents. Net income from fiscal 2025 reached almost$80 billion, a massive leap from the 2.7 billion loss just three years earlier. Now any investor shouldn't just look at net income. We would recommend looking at operating profits, which is more tied to the quality of the business and performance. Amazon's transition to advertising AWS is largely complete, but there is still more room in the tank. The risk is that management becomes too focused on protecting the castle and not enough on building the next one. But given the incentives, it is hard to bet against the Everything Store, and as long as they continue to share their economies of skill with the customer, we think the flywheel will keep spinning. This was the Amazon TDI Deep Dive. And if you enjoyed it, please give us a 5-star review on Spotify or Apple Podcasts, or wherever you get your podcasts, and we'll see you in the next one.