Jeff Bezos and the Amazon He No Longer Runs

Summary

• Jeff Bezos’ lasting contribution to Amazon is not a single product or founder legend. It is an operating system: long-term capital allocation, customer-oriented measurement, infrastructure reuse, and conversion of internal capabilities into third-party markets.
• The strongest public evidence of post-CEO influence is formal and structural: Bezos remains Executive Chair and Chair of the Board of Amazon, while the company continues to allocate capital across businesses built or approved during his CEO tenure.
• AWS, marketplace services, logistics, Prime, advertising, and technology infrastructure all originated from founder-era choices, but their later results belong to Amazon’s broader teams and the CEO period of Andy Jassy, not to Bezos alone.
• The record also contains unresolved constraints: criticism over labour and safety, antitrust litigation, concerns about vendor dependency, massive capital needs, and the risk that Amazon’s infrastructure model becomes harder to justify if growth slows.

The Founder Question After the Transition

Jeff Bezos is no longer the CEO of Amazon. This fact matters because Amazon is now too large, too regulated, and too operationally distributed to be explained by a founder’s will alone. In July 2021, after more than two decades as CEO, Bezos handed the CEO role to Andy Jassy and became Executive Chair. The public record is clear about the role change. It is less clear about the practical boundary between founder influence and delegated management.

That boundary is the useful subject. Bezos did not merely start a famous online retailer. He built a company that used retail as a training ground for software systems, logistics systems, pricing systems, marketplace governance, cloud infrastructure, advertising inventory, and managerial routines. The resulting Amazon can be described as a stack of dependencies. Customers depend on Prime delivery. Sellers depend on marketplace and logistics access. Developers and enterprises depend on AWS capacity. Brands depend on sponsored placement.

Investors depend on the company’s continuing ability to turn capital expenditure into future operating leverage.

The founder’s strategic signature is visible in that stack, even when he is no longer giving the day-to-day orders. Amazon’s 2024 Form 10-K lists Jeffrey P. Bezos as Executive Chair and traces his full journey: founder in 1994, CEO from May 1996 to July 2021, Chair of the Board since 1994, and Executive Chair since July 2021. It also shows a company whose scale far outruns any simple founder story: $637.959 billion in net sales in 2024, $68.593 billion in operating profit, approximately 1.556 million full-time and part-time employees, and capital expenditure of $77.7 billion.

Those numbers do not prove that Bezos directed Amazon’s current AI infrastructure commitments, advertising priorities, labour practices, or marketplace responses. They prove that the company he handed to Jassy still runs on mechanisms he spent decades building: tolerating short-term spending, expanding capacity ahead of visible demand, turning fixed-cost infrastructure into external revenue, and using measurement to discipline nearly every customer-touching process.

This profile therefore treats Bezos neither as a distant celebrity shareholder nor as an invisible CEO. The evidence supports something more specific. He is the architect of an operating model, the holder of ongoing governance status, and the founder whose early rules still constrain and enable Amazon’s choices. The open question is whether those rules remain a source of cumulative advantage or a source of regulatory, social, and capital-allocation pressure.

The Company He Took Public Was Fragile

Amazon’s later scale can obscure just how uncertain the early business was. In its 1997 Form 10-K, Amazon described itself as the leading online retailer of books after opening for business in July 1995. The same document shows the risk beneath the momentum. By the end of 1997, Amazon had generated more than $164 million in sales from approximately 1.5 million customer accounts in over 150 countries. It also reported a net loss of $27.6 million for 1997 and a cumulative deficit of $33.6 million.

That combination was significant. Bezos was not running a profitable bookshop that gradually moved online. He was running a loss-making public company that was trying to prove that selection, software, logistics, and customer experience could produce scale before established retailers closed the gap. The early risk factors were not cosmetic. Amazon depended on suppliers, technical staff, system development, brand trust, distribution capacity, and the continuing willingness of capital markets to fund growth before profits.

Bezos’ early decision that still explains Amazon was to make the company legible as a long-horizon infrastructure play, not as a simple retailer with a website. The 1997 shareholder letter, later republished by Amazon and regularly appended to annual shareholder communications, defined the company’s priority as long-term market leadership. Bezos emphasised customer focus, aggressive investment, the value of cash flows, and a willingness to make bold choices even when they depressed near-term financial appearance.

This was not mere rhetoric. Amazon had to spend before it could fully prove demand. It expanded distribution capacity, hired employees, deepened its catalogue, built recommendation and review systems, and treated the website as a measurable selling machine. The goal was not to win by selling books online. The goal was to use books as the first category of a repeatable system: acquire customers, broaden selection, reduce friction, collect behavioural data, improve the interface, expand logistics, and keep reinvesting.

This distinction matters because it separates founder myth from an observable managerial pattern. The myth says Bezos foresaw the entire Amazon empire. The record shows a narrower but more durable habit: when Amazon found a capability that could be measured, scaled, automated, and offered to another party, it tended to make it central. Retail selection became marketplace selection. Internal logistics became third-party fulfilment. Internal computing capacity became AWS. Placement in search results became advertising. Internal automation became a management norm.

The young company also baked in a dependency on Bezos himself. The 1997 document warned investors that Amazon’s performance depended substantially on senior management, and specifically on Bezos as President, CEO, and Chair of the Board. That warning was rational at the time. In a small, high-growth, loss-making company, the founder’s credibility helped hold together investors, employees, suppliers, and customers. The later problem was different. Once Amazon became enormous, founder dependency could no longer be the central explanation. The company needed mechanisms that outlived him.

The Operational Doctrine: Customer Measurement, Scale, and Optionality

Bezos’ operational doctrine is often summarised in slogans about customer obsession and long-term thinking. The public record supports those phrases, but they are too soft if not translated into operational behaviour. At Amazon, customer focus meant more than service language. It meant using customer behaviour to justify investment, discipline internal priorities, and make the company’s internal processes measurable.

This doctrine shaped capital allocation. Bezos was willing to accept thin margins and losses when management believed the spending would increase selection, convenience, speed, or infrastructure capacity. The company did not view retail as a finished format. It viewed it as a set of constraints to be reduced: limited selection, slow delivery, uncertain search, poor product information, distrust of online payment, inventory risk, and customer-service friction.

The same logic explains Amazon’s later willingness to operate across adjacent layers of commerce. Third-party sellers could expand selection without Amazon owning all inventory. Logistics services could make sellers’ inventory behave more like Amazon’s own. Prime could turn shipping from a single transaction into a subscription relationship. Advertising could monetise product discovery inside the store. AWS could convert internal technical scale into a service sold to developers and enterprises.

The operational doctrine was not risk-free. A system designed to remove friction for customers can create new dependency for counterparties. Sellers gain access to demand but become exposed to Amazon’s rules, fees, placement systems, logistics terms, and competition from private brands. Brands gain advertising reach but may have to pay for visibility that once came from ordinary search-result placement. Enterprises gain cloud speed but inherit migration costs, outage exposure, and long-term dependency on a hyperscale provider.

This dual character is central to Bezos’ record. He built mechanisms that offered convenience and lower friction at massive scale. Those same mechanisms concentrated bargaining power inside Amazon’s platform. The company became valuable because others could plug into it. It became controversial because plugging in often meant accepting Amazon’s terms.

The doctrine also made Amazon unusually comfortable with low near‑term visibility on future profit pools. In the early years, investors could see growth but not stable profit. Later, they could see infrastructure spending but not always the future split of returns among retail, AWS, advertising, logistics, and subscriptions. Bezos’ answer was not to make every unit profitable at all times. It was to make the system larger, more useful, and harder to replace.

This approach worked in important ways. It also left Amazon exposed to the charge that its advantages are not only operational but structural: control of the marketplace, control of logistics options, control of advertising inventory, control of cloud capacity, and control of the rules that determine who is visible to customers. Bezos’ significance lies in building a company whose strategic choices became questions of market structure.

Marketplace and Logistics Became a Control Surface

Amazon’s marketplace strategy is one of the clearest examples of founder-era optionality turning into a long-term control surface. The company began by selling goods directly to customers. It then allowed third parties to sell through Amazon’s store, expanding selection while shifting some inventory risk off the company. Over time, third-party seller services became one of Amazon’s main revenue sources. In 2024, Amazon reported $156.146 billion in revenue from third-party seller services.

This was not an ancillary side business. It changed Amazon’s shape. The store no longer needed to be simply a retailer’s shelf. It became a marketplace in which Amazon controlled search, payments, trust systems, vendor access, fee structures, and customer relationships. This produced real value for small and medium-sized businesses that could reach national and global demand without building their own traffic engine. It also made those businesses dependent on Amazon’s operational choices.

Logistics intensified the dependency. In 2006, Amazon announced services that let small and medium-sized businesses access fulfilment, customer service, and Amazon’s website features. The business idea was simple: external sellers could use Amazon’s logistics infrastructure to make their products easier for customers to buy and receive. The strategic consequence was broader. Sellers using Amazon logistics could compete more directly within Amazon’s customer promise, while Amazon gained more volume, more data, and more control over the delivery experience.

From the perspective of Bezos’ decisions, the important choice was not simply to launch a service for sellers. It was the choice to treat Amazon’s internal capabilities as external products. This pattern later became central to AWS and advertising as well. If Amazon had built something hard for itself, it asked whether that capability could be sold, rented, or attached to a broader ecosystem.

The marketplace and logistics also show why Bezos’ record should not be reduced to admiration. The same design that creates opportunities for sellers also creates vulnerability. A merchant can use Amazon to reach customers it could never economically reach alone. But the merchant also operates in an environment of ranking, fees, logistics, reviews, and enforcement measures that Amazon controls. The stronger Amazon’s customer relationship becomes, the weaker the merchant’s independent bargaining position may become.

This tension later became part of the antitrust and regulatory record. In its 2024 Form 10-K, Amazon disclosed that the Federal Trade Commission and state attorneys general had alleged a monopoly position over online superstores and marketplace services, and anticompetitive practices related to pricing policies, advertising, Prime structure, and promotion of Amazon products. Amazon states that it contests the allegations of wrongdoing and intends to defend itself vigorously. These allegations do not constitute a final judicial finding. They are nevertheless important because they target the very operating surface that Bezos helped create.

AWS Was a Delegated Invention, Not a Single Person’s Product

AWS is the most important test of attributing Amazon’s post-founder value. It is tempting to tell the story as another example of Bezos seeing the future before anyone else. That version is too simple. AWS was built by teams, led for years by Andy Jassy, and developed through engineering, sales, security, pricing, data centre, and customer‑support work that no single founder could have accomplished. The better attribution is organisational: Bezos approved and protected the conditions under which such a business could emerge.

Amazon launched major web services products in the mid‑2000s, turning internal infrastructure capabilities into commercial services for developers and enterprises. The strategic pattern matched the logic of marketplace and logistics. A capability built to support Amazon’s own operations could become a paid platform for third parties. Instead of treating servers, storage, databases, and developer tools only as internal cost centres, Amazon made them external products.

The result was not immediately obvious to outside observers. Cloud infrastructure was capital‑intensive, technical, and remote from Amazon’s original identity as a retail company. It required data centres, networks, reliability, security, developer trust, and a pricing model that would invite customers to move workloads off their own infrastructure. It also required patience. The profit pool would not be proven at launch.

By the end of 2020, AWS had become one of Amazon’s central economic engines. Amazon’s 2020 Form 10‑K reported AWS net sales of $45.370 billion and AWS operating profit of $13.531 billion. The same year, Amazon as a whole reported $386.064 billion in net sales and $22.899 billion in operating profit. AWS accounted for a much smaller share of sales than retail operations but a very large share of operating profit.

The pattern persisted after Bezos stepped down as CEO. In 2024, Amazon reported AWS net sales of $107.556 billion and AWS operating profit of $39.834 billion. That means AWS sales more than doubled between 2020 and 2024 and AWS remained the company’s most important profit centre. The same 2024 document shows how the cloud business determines Amazon’s capital posture. Amazon reported $77.7 billion in capital expenditure in 2024, primarily related to technology infrastructure to support AWS growth and additional logistics capabilities, and said it expected capital expenditure to increase in 2025, mainly because of technology infrastructure.

AWS is therefore both the strongest argument for Bezos’ founding architecture and the strongest warning against over‑attribution. Bezos’ company built the conditions: a tolerance for infrastructure spending, a habit of turning internal systems into external services, and a willingness to sell tools to developers at scale. But Andy Jassy’s leadership of AWS was central before he became Amazon CEO. AWS’s later results also belong to leaders such as Adam Selipsky, Matthew Garman, Werner Vogels, and thousands of engineers, operators, sales teams, data centre specialists, and support staff.

This distinction matters because the present‑day economics of Amazon’s AI infrastructure are often interpreted as an extension of the AWS bet. That is reasonable at the pattern level. Amazon is again spending heavily on technology infrastructure ahead of fully visible returns. It is again asking investors to believe that capacity, scale, and customer demand will justify the near‑term capital intensity. But the public evidence does not show Bezos personally directing specific AI spending after 2021. It shows a company using a founder‑era pattern under a successor CEO.

The relevant question is whether that pattern still works when inputs become more expensive. AI infrastructure may require specialised chips, large energy commitments, accelerated data centre construction, and competition with other hyperscale providers. The Bezos pattern rewards capacity before demand is fully mature. It can also punish overbuilding if demand, pricing, power availability, or customer adoption disappoint. AWS made that risk look brilliant in hindsight. AI infrastructure has not yet earned the same historical confidence.

The CEO Transition Was an Organisational Test

The transition announcement in February 2021 came after an exceptional year for Amazon. Pandemic demand had pulled more consumers and businesses toward online retail, cloud services, and digital operations. Amazon announced 2020 annual net sales of $386.1 billion and operating profit of $22.9 billion, and then stated that Bezos would become Executive Chair and Andy Jassy would become CEO in the third quarter.

The timing matters. Bezos did not leave the CEO role after a visible operational collapse. He left after Amazon had proven its scale in retail, Prime, marketplace, logistics, and AWS, and after Jassy had already steered AWS into a major enterprise‑infrastructure business. The succession was therefore not a search for an outsider to fix a founder’s company. It was a handover to an internal operator whose credibility came from one of the founder‑era’s most successful businesses.

Bezos’ 2020 shareholder letter, published in 2021, also signals how he wanted the transition to be understood. He highlighted the number of Prime subscribers worldwide, the number of small and medium-sized businesses selling in Amazon’s store, and the annualised run rate of AWS. He also cited employee success and workplace safety as areas where Amazon needed to do better. The combination is telling. Bezos wanted the record to show invention and scale, but he also acknowledged that labour and the employee experience were not solved just because Amazon had won a union vote in Bessemer.

The transition created a new attribution problem. When Amazon expanded after 2021, how much of it was still down to Bezos? When Amazon cut positions, adjusted capacity, slowed some physical‑store ambitions, or redirected infrastructure spending, how much belonged to Jassy? The public evidence cannot answer every operational detail. The clearest line is formal: Bezos remained Executive Chair; Jassy became CEO. Strategic continuity can be inferred from the company’s architecture, but day‑to‑day accountability shifted.

This line matters for judging both credit and blame. Bezos deserves credit for building an organisation that could survive succession without abandoning its core mechanisms. He should not be credited with every AWS sale, advertising dollar, warehouse‑automation improvement, or AI capacity decision after 2021 unless specific evidence ties him to them. Equally, he should not be treated as the sole decision‑maker for every layoff, product cancellation, or regulatory response after 2021.

The governance signal remains significant, however. The Executive Chair title is not a default honorary label in a founder‑led company. It gives Bezos proximity to board‑level strategy, leadership selection, long‑horizon initiatives, and investor discourse. It also preserves a symbolic centre. Amazon’s employees, executives, regulators, investors, and competitors all know that the founder remains formally inside the corporate structure.

The post‑CEO Amazon therefore sits between two interpretations. One says Bezos left and Jassy now runs the company operationally. That is true in the CEO sense. The other says Amazon remains Bezos’ company because its main profit pools and decision habits were built under his leadership. That is true in the architectural sense. The most accurate assessment is that Bezos delegated the controls without dismantling the machine.

What Amazon Inherited: Retail, Prime, Advertising, and Labour Scale

In 2024, the breakdown of Amazon’s sales showed how many founder‑era systems had become distinct economic engines. Online stores generated $247.029 billion. Third‑party seller services generated $156.146 billion. AWS generated $107.556 billion. Advertising services generated $56.214 billion. Subscription services generated $44.374 billion. These line items are not independent accidents. They reinforce one another.

Online stores create customer traffic. Marketplace sellers expand selection. Logistics and Prime increase delivery reliability and frequency. Advertising monetises visibility in a high‑intent purchasing environment. AWS finances and normalises large‑scale technology infrastructure. Subscription services deepen recurring customer relationships. The operating model is circular: each activity creates demand, data, or infrastructure that can reinforce another activity.

That circularity is a Bezos legacy. It also makes Amazon hard to regulate or evaluate using ordinary category boundaries. Is Amazon a retailer, a marketplace, a logistics provider, a cloud company, a media distributor, an advertising platform, or an AI device and services company? The answer is yes, but the most useful description is that Amazon is an infrastructure broker spanning consumer demand, seller access, enterprise computing, and delivery capacity.

Advertising is a particularly clear example of late monetisation of founder‑era infrastructure. Bezos did not build Amazon as an advertising business in the 1990s. But the store he built eventually controlled enough purchase intent to sell paid placement and measurement to brands and sellers. Advertising revenue did not require Amazon to create a separate consumer destination from scratch. It flowed from controlling the customer interface.

This creates strong economic and policy questions. For investors, advertising can improve margins because it monetises traffic Amazon already owns. For sellers and brands, advertising can look like another toll layer inside a marketplace they depend on. The more Amazon’s search and recommendation systems matter, the more valuable paid visibility becomes. The more valuable paid visibility becomes, the more sellers can feel compelled to buy it.

Labour scale is the other face of the same model. Amazon’s roughly 1.556 million employees at the end of 2024 show the physical reality behind the digital interface. The convenience customers experience rests on warehouses, delivery systems, data centres, customer service, and technical operations. Bezos’ 2020 letter acknowledged that Amazon needed a better vision for employee success and safety. That acknowledgement should not be treated as a finished solution. It should be read as evidence that the labour demands of the operating model became too visible for founder rhetoric alone.

The company’s operational choices after the CEO transition show both continuity and correction. Amazon continued to expand its capital‑intensive technology infrastructure, but it also recorded impairment and severance charges related to physical stores and position eliminations. That is what large‑scale strategy looks like after the founder era: not a triumphant march of invention, but a portfolio of commitments, write‑downs, adjustments, and capacity decisions.

Bezos’ role in that inheritance is best summarised this way: he built the rules by which Amazon seeks cumulative advantage, but the organisation must now manage the consequences at a scale where every rule has external effects. A pricing change can affect sellers. A logistics policy can affect workers and local markets. A cloud outage can affect businesses. A data centre plan can affect electricity demand and public permits. A Prime design choice can become an antitrust issue.

Strategic Leverage After CEO Control

A founder’s leverage after a CEO transition can take several forms. It can be legal control through voting rights. Economic influence through ownership. Board authority. Cultural authority. Or the continuing power of systems designed during the founder’s tenure. For Bezos, the public evidence is strongest on board role, economic association, cultural authority, and system legacy. It is weaker on direct command of current operations.

As Executive Chair, Bezos remains formally tied to Amazon’s governance. As founder and long‑tenured former CEO, he remains associated with the company’s preferred investor narrative: long‑term invention, customer focus, and willingness to invest heavily. As a significant shareholder, even after reported stock sales, his economic relationship with Amazon remains material. As the author of the original shareholder doctrine, he continues to supply the language through which Amazon explains its capital intensity.

But strategic leverage should not be confused with direct control. The company has an appointed CEO, appointed segment leaders, and a board structure. Jassy is responsible for current management decisions. The current AWS leadership is responsible for cloud execution. The leaders of advertising, logistics, retail, devices, and operations own their domains. The public record does not support describing Bezos as secretly running Amazon after 2021.

What they do support is describing him as a continuing constraint on imagination inside and outside the company. Amazon can change, but it changes inside a grammar designed by the founder. Long‑term investment remains easier to defend because Bezos made it part of the company’s identity. Massive infrastructure spending remains more plausible because AWS proved that internal scale can become external revenue. Vendor dependency remains embedded in the commercial architecture because marketplace and logistics became central to selection and delivery. Advertising growth remains tied to control of the purchasing interface.

This is founder control in its most durable form: not the founder making every decision, but the founder having built the decision environment in which successors operate.

This form of leverage has limits. If Amazon’s AI and cloud investments disappoint, investors will hold Jassy and current management accountable. If regulators impose marketplace changes, legal outcomes will reshape the model regardless of the founder’s preferences. If social pressure raises logistics costs or changes warehouse practices, Amazon will have to adapt operationally, not cite old shareholder letters. If customers or enterprises diversify away from Amazon dependencies, the company’s historical flywheels could weaken.

Bezos’ post‑CEO influence is therefore asymmetric. He continues to lend legitimacy to long‑cycle bets and to the company’s self‑image as an invention engine. He cannot single‑handedly guarantee that those bets will overcome higher capital costs, regulatory hostility, or infrastructure constraints. The machine he built still works, but it now operates in a market more sceptical of concentrated platform power.

The Facts Against the Founder Myth

The founder myth makes Bezos too perfect. It implies that Amazon’s rise was a straight line from vision to dominance. The facts are messier. Amazon lost money for years, faced credible competition, depended on capital markets, benefited from broader internet adoption, and built its advantages through teams and institutions as much as through the founder’s choices.

It also faced criticism that cannot be dismissed as envy of success. The FTC and state attorneys general challenged Amazon’s marketplace practices. Workers and advocates criticised warehouse conditions and bargaining power. Sellers raised concerns about fees, visibility, enforcement measures, and competition with Amazon’s own products. Policy observers asked whether Prime, marketplace search, logistics, and advertising created interlocking dependencies that restrict competition.

None of these problems alone proves that Bezos’ model was illegal or overall harmful. They show that Amazon’s success created governance problems. A company can improve customer convenience while concentrating power over suppliers and sellers. It can lower some prices while increasing dependency on its infrastructure. It can create cloud services that help startups grow while becoming a critical dependency for enterprises and public services. It can create jobs while facing serious questions about safety, pace, and worker voice.

Bezos’ 2020 shareholder letter is useful because it does not entirely hide that tension. He defended Amazon’s value creation but acknowledged that the company needed a better vision for employees. That acknowledgement matters because it came at the point of transition. Bezos left the CEO role having built a company whose customer and investor story was powerful, but whose social and platform‑power story remained contested.

The facts also challenge another myth: that Amazon’s strategic breakthroughs were obvious from the start. AWS was not a natural extension of an online bookshop for most observers. Logistics services for third‑party sellers were not just retail operations; they were a platform choice. Advertising was not an original identity; it was a later monetisation of accumulated customer attention. The reason Bezos matters is not that every outcome was premeditated. It is that Amazon repeatedly had permission to discover large businesses inside its own operational problems.

That permission is perhaps Bezos’ most important managerial contribution. Many companies build internal capabilities and leave them trapped inside cost centres. Amazon looked for ways to expose those capabilities to external demand. That required technical skill, but also a financial and governance culture willing to tolerate ambiguity. Investors had to accept periods when the company spent heavily without clear near‑term evidence. Employees had to operate in a measurement‑heavy environment. Customers and sellers had to accept Amazon as an intermediary. Regulators eventually had to ask whether the intermediary had become too powerful.

The founder myth celebrates the first half and neglects the second. A decision‑focused assessment must account for both.

What Could Change the Assessment

The assessment of Bezos could shift in either direction with better evidence. If Amazon disclosed board minutes, executive testimony, or primary source statements showing that Bezos personally approved or redirected major AI infrastructure, marketplace, advertising, or logistics decisions after 2021, the case for continued direct influence would strengthen. The current public record does not go that far.

If final court rulings or settlements forced Amazon to change its marketplace, Prime, advertising, or seller‑service practices, the assessment of the founder‑era model would also change. The current record contains serious allegations and Amazon’s denial of wrongdoing. It does not yet contain a definitive structural remedy that would rewrite the company’s operating surface.

If later audited filings show that AWS and AI infrastructure spending produces sustained operating profit growth, Bezos’ original infrastructure logic will look more valuable as a legacy. If the spending does not produce adequate returns, that same logic will look more vulnerable: a culture trained to build ahead of demand can overbuild when the marginal cost of capacity rises.

The social record could also change the assessment. Amazon’s scale rests on the labour of warehouses, delivery, technology, and customer service. If the company materially improves safety, retention, pay, and worker voice while preserving customer speed, the operating model becomes more defensible. If labour conflict and injury concerns persist, the customer‑convenience narrative remains incomplete.

Finally, the assessment would change if Bezos’ governance role became clearly symbolic rather than substantive. For now, the official title, board role, ownership relationship, and founder authority make him more than a historical figure. But the article should not claim to know what Amazon does not disclose. The evidence supports ongoing strategic proximity, not invisible command.

Assessment

Jeff Bezos built Amazon by repeatedly asking a specific question: what internal capability, if scaled, could become a market? In retail, the capability was online selection and customer trust. In marketplace, it was seller access to demand. In logistics, it was fulfilment capacity. In AWS, it was IT infrastructure. In advertising, it was control of purchase intent. In AI infrastructure, Amazon is again testing whether capacity built at enormous cost can become the next layer of external dependency.

That pattern is Bezos’ strongest verified legacy. It matters more than personal anecdotes, management folklore, or founder charisma. It explains why Amazon became more than a retailer and why the company remains strategically coherent after the CEO transition. It also explains why Amazon is controversial. A company that sells access to its own infrastructure ends up governing the terms on which others reach customers, run workloads, and compete for visibility.

Bezos’ current role must be assessed with discipline. He remains Executive Chair and Chair of the Board. The company still uses the capital‑allocation logic he made credible. Its biggest profit engine, AWS, was born from the founder‑era decision to commercialise internal infrastructure. Its seller, logistics, Prime, advertising, and subscription systems still reflect a model built under his leadership. That is genuine strategic leverage.

But the operational record also calls for restraint. Bezos did not personally build every system, close every sale, write every line of code, run every warehouse, or make every decision after 2021. Andy Jassy and Amazon’s current leaders are responsible for the CEO era after July 2021. AWS results belong to a large organisation. Marketplace and logistics results belong to thousands of operators and millions of sellers as well as to Amazon. Regulation, labour, energy, and capital markets now constrain Amazon in ways that founder doctrine cannot simply override.

The fairest view is that Bezos no longer runs Amazon, but Amazon still runs on choices he normalised. He turned patience into a funding strategy, infrastructure into a product strategy, and customer dependency into a platform strategy. Those choices produced one of the most important companies in the modern economy. They also produced a company whose power now demands explanation, limits, and evidence.

For readers tracking market structure, Bezos matters less as a personality than as an ongoing operational thesis. The post‑founder‑control Amazon is still testing its bet: if a company makes itself useful enough, invests long enough, and converts enough internal systems into external markets, scale will compound faster than criticism, regulation, and capital intensity can slow it down. The answer is not settled. That is why Bezos remains a live subject even after he left the CEO chair.