Walmart, Amazon Take Divergent Approaches to Agentic Commerce

Walmart and Amazon are taking apparently opposite approaches to agentic commerce. Walmart leans to towards an “open” approach while Amazon prefers a more-closed approach. 

Walmart is trying to make its products available wherever consumers choose to shop with AI, partnering with ChatGPT and Gemini, while also building its own agentic agent Sparky,  while Amazon is trying to ensure that AI shopping remains inside an environment Amazon controls by barring most third-party access and relying on its own Rufus agent and other related features.  

It’s an old set of choices. Should an incumbent embrace a new technology platform that may weaken its control over customers, or defend its existing ecosystem and risk losing customers to the new platform?

Neither strategy is universally successful or applicable, though. Much hinges on a firm’s position within an industry. Openness tends to help a company when the new technology shifts value toward a platform or distribution channel it does not control. 

Closed ecosystems tend to work better when the company can make its own platform the preferred destination. 

Open approaches also tend to enable new competition, while closed approaches might tend to favor incumbents with market power. 

But it also is fair to say that both firms attempt to control the interface, the data, and the transaction.

Strategic issue	Walmart's calculation	Amazon's calculation
Customer acquisition	Third-party AI can introduce Walmart to consumers who do not begin their shopping journey on Walmart.com.	Outside agents could divert existing Amazon customers and reduce the importance of Amazon's own app and website.
Competitive position	As a retailer competing against Amazon, Walmart benefits if AI makes retailer choice more fluid.	Amazon benefits from the frictionless integration of its catalog, reviews, payments, Prime benefits, and delivery network.
Customer data	Cooperation can provide incremental demand, although Walmart risks losing some browsing and intent data to AI intermediaries.	Outside agents could capture valuable signals about what customers want, compare, reject, and ultimately buy.
Advertising and sponsored listings	AI recommendations could create another route to product discovery, but may weaken Walmart's control over advertising inventory and ranking.	Amazon's sponsored product placements and retail-media economics could be undermined if an independent agent chooses products using its own ranking criteria.
Price competition	Walmart can compete on price, convenience, grocery, and physical-store reach if agents make comparisons easier.	Agents that systematically compare alternatives could expose price differences and weaken Amazon's ability to retain customers through convenience and ecosystem benefits.
Transaction economics	A new channel is attractive if incremental sales outweigh integration costs and any commissions or loss of control.	Keeping transactions in-house protects more of the economics and preserves control over checkout, customer service, and fulfillment.
Principal risk	Becoming a supplier to someone else's AI platform rather than owning the customer relationship.	Protecting the current platform too aggressively and allowing a new AI intermediary to become the consumer's default shopping destination.

Beyond all that, there can be other issues driving a policy choice. Walmart, for example, encountered issues with a native ChatGPT checkout experience, and now integrates its own Sparky agent. And Amazon might eventually allow some access by external agents, if it can keep control of the customer relationship, or conversely, if it finds it must cede some control to protect sales volume.

Historically, computer operating systems, browsers and mobile phones illustrate cases where distribution, customer relationships or the economic value of the ecosystem shaped choices, generally involving choices between degrees of openness versus closed approaches. 

Historical case	Strategic choice	What happened	Lesson for Walmart and Amazon
IBM PC vs. the Microsoft/Intel ecosystem (1980s–1990s)	IBM initially benefited from a PC architecture built around widely available components and outside software. Microsoft and Intel occupied increasingly powerful positions within that ecosystem.	IBM helped establish a market whose economics ultimately favored independent software and chip platforms. IBM lost much of its control over the architecture it had helped popularize. IBM history of e-business; U.S. DOJ Microsoft findings.	A company can create a successful ecosystem without capturing the largest share of its eventual value. Walmart could help normalize AI-mediated shopping while an AI provider captures customer attention and economic rents.
IBM OS/2 vs. Microsoft Windows (late 1980s–1990s)	IBM and Microsoft initially cooperated on OS/2, then diverged. Microsoft concentrated on Windows, while IBM continued to develop OS/2.	Windows gained a decisive advantage from software availability, hardware compatibility, and distribution through PC manufacturers. OS/2 struggled to match that ecosystem. OS/2 history.	Cooperation alone does not win. A company must attract developers, users, and complementary services around its platform. Walmart's partnerships will matter most if they generate durable customer relationships and sales.
Microsoft Internet Explorer vs. Netscape (1990s)	Microsoft initially treated the browser as a threat to Windows and responded by integrating Internet Explorer with Windows and distributing it without a separate charge.	Internet Explorer displaced Netscape's leading position. The U.S. antitrust case documented Microsoft's efforts to protect its operating-system monopoly from the emerging browser platform. U.S. DOJ findings.	Amazon has a rational reason to defend its shopping interface. But incumbents may also use integration and distribution advantages to defend a threatened business model, potentially inviting regulatory scrutiny.
AOL's walled garden vs. the open Internet (1990s–2000s)	AOL initially offered a curated online environment; the wider Internet offered access to independently operated websites and services.	The open web became the broader platform for online activity. AOL later adapted its business rather than retaining its former position as the primary gateway to online life. IBM's history of e-business.	A retailer's controlled environment can be attractive until consumers prefer a more convenient intermediary that searches across competing destinations. Amazon must consider whether its walls add enough value to offset the friction they impose.
Apple iPhone vs. Android (2007 onward)	Apple chose a tightly integrated hardware and software ecosystem. Android enabled a broader network of device manufacturers to participate in a shared platform.	Both strategies succeeded, but in different ways: Apple captured substantial value through integration and premium devices, while Android spread across a much wider range of manufacturers. Apple's developer ecosystem; Android.	This is a crucial counterexample to the idea that openness always wins. Amazon could preserve a highly profitable closed ecosystem if its convenience, trust, and fulfillment advantages remain compelling.
The open web vs. proprietary online services (1990s onward)	Proprietary services controlled the user experience; web standards allowed independent sites, search engines, and new intermediaries to emerge.	Open protocols enabled new categories of businesses and made it possible for companies such as Google to become major gateways to information without owning all the underlying content.	An AI shopping agent could become the equivalent of a search engine for purchases, directing demand across retailers rather than requiring consumers to visit each retailer first.

It might be tempting to conclude that “open” tends to win, but not always. Apple’s  example suggests openness is not a universal rule. Closed ecosystems can thrive when consumers value a coherent experience enough to accept the restrictions. 

And “control” might exist in a few different parts of the value chain:

• The user's AI agent, which understands intent, compares products, recommends and negotiates

• Retailers and their systems who supply product catalogs, prices, inventory, promotions, loyalty and checkout experiences

• Physical commerce infrastructure including warehouses, stores, picking, delivery, returns and customer service. 

It is possible that ultimate success for Walmart or Amazon will hinge on dominance in a couple, but not all, of these parts of the value chain. And either open or closed strategies could work. 

Indicator	Evidence favoring Walmart's strategy	Evidence favoring Amazon's strategy
Where shopping begins	A rising share of purchase journeys originates in general-purpose AI assistants.	Consumers continue to begin most high-intent shopping journeys on Amazon.
Incremental customer economics	AI referrals generate profitable new customers and larger baskets at reasonable acquisition cost.	Amazon's own assistant increases conversion, frequency, and retention more effectively than external agents.
Who controls recommendations	Independent agents can compare retailers and rank products without retailer restrictions.	Retailer-controlled agents provide sufficiently superior recommendations, loyalty benefits, and convenience to keep customers inside their ecosystems.
Advertising and monetization	Retailers can monetize AI-driven demand without surrendering too much margin or data.	Amazon retains valuable sponsored placements, purchase-intent data, and other retail-media revenue within its own environment.
Transaction reliability	Common standards make inventory, payment, substitution, delivery, and returns work across retailers.	Complex transactions work materially better inside each retailer's integrated systems.

Walmart has an incentive to open distribution because it can gain from a change in the shopping gateway. 

Amazon, on the other hand, has an incentive to control distribution because it is the existing gateway.

But much hinges on where users experience value. If starting a shopping journey inside Amazon continues to make sense, Amazon wins. If users start to trust their agents, then perhaps Walmart wins. And it seems inevitable that neither Amazon nor Walmart approaches win all the time. 

It might matter whether the user values selection across retailers (perhaps choices are narrower but price differences can be significant), compared to the value of curation when the product universe is large. 

The key variables might include:

• how much comparison matters

•  how costly a mistake is

• how much context the agent needs

• whether the retailer has a unique advantage in price, assortment, or fulfillment.

Not every shopping experience favors each strategy. 

Shopping experience	Likely advantage	Why
Routine grocery replenishment	Walmart openness	An agent can compare prices across grocers, assemble a basket, and optimize substitutions, store proximity, and delivery. Walmart's stores and grocery scale are valuable if the agent can access them.
Repeat purchases of known products	Amazon curation, with a caveat	One-click reordering, purchase history, reviews, and reliable delivery make Amazon convenient. But if an independent agent can automatically find the cheapest equivalent across retailers, Walmart benefits from openness.
Staples and commodity products	Walmart openness	For paper towels, detergent, batteries, and pantry staples, product differences may be small relative to price, availability, and delivery cost. Cross-retailer comparison is valuable.
Novelty and product discovery	Amazon curation	A broad catalog, customer reviews, recommendations, and product comparisons can help consumers discover unfamiliar products. A strong independent AI agent could, however, reproduce much of this capability across retailers.
Rare, complex purchases	Amazon curation, potentially	For cameras, specialized tools, electronics, or hobby equipment, product specifications, compatibility, and reviews matter. Amazon has extensive product information and shopping history. An independent agent may win if it searches specialist retailers and manufacturer sources more comprehensively.
Urgent purchases	Depends on local fulfillment	The winning retailer is the one that can actually deliver the right item fastest. Walmart's physical-store network may be decisive for local essentials; Amazon may win on products already stocked in its fulfillment network.
Large, mixed baskets	Walmart openness	An agent can optimize the entire basket across retailers, rather than selecting the best individual product at one retailer. Walmart's grocery, household, pharmacy-related merchandise, and general merchandise breadth can help.
Premium or highly differentiated goods	Depends on brand access	Product authenticity, warranties, authorized distribution, expert advice, and service may matter more than the lowest price. Neither Walmart's openness nor Amazon's curation automatically wins.
Gift shopping	Open agents, potentially	An agent can search multiple stores using recipient interests, budget, delivery deadlines, and gift availability. Amazon's reviews and recommendations still help, but cross-retailer discovery may offer more distinctive choices.
Fashion, beauty, and home décor	Mixed	Personal taste, visual discovery, sizing, returns, and brand identity complicate pure price optimization. Retailer-specific expertise can matter, but a cross-retailer agent could broaden discovery and compare return policies.

And different types of purchases might create value in ways that either approach can tend to capture:

• Frequency favors whichever platform has the best history, automation, and fulfillment.

• Commodity products favor open, multi-retailer agents.

• Decision complexity might favor the agent with the best data and reasoning

• Fulfilment (delivery speed, returns, warranties, local availability matter favors the provider with  strongest operational capabilities.

And there also are the issues around user preferences. Some users will want to delegate more; some less. Some will trust agents more; some less. 

The point: it is not certain which approach is “better” or “more successful” long term. Both can work, sometimes, for some purchases.

Is Sustainable Frontier Model Market Share Leadership Already Emerging?

Has market leadership in the frontier artificial intelligence model market already begun to take shape? And are calls to slow the development of the latest frontier models going to help solidify that market pattern? It seems possible. 

Consider the advantages of a slowing of frontier model releases for existing leaders. If new release pace slows to “a new model every 18 months to 24 months,” then the capital intensity of the business might well slow from the breakneck pace of “a new model every six months to 12 months.”

The caveat, of course, is that all the leading contenders need to agree to respect the new pattern. But such agreements can temper the competitive-investment dilemma where each leading firm has to respond if a major competitor ups investment, even if all the leading models are relatively comparable at each stage of development, as suggested by Stanford's Human-Centered Artificial Intelligence institute. 

Savings would develop as model developers spent less on training costs, for example, which might also drive less spending on graphics processing units and data center facilities. That, in turn, would help developers reach revenue targets more easily. 

On the other hand, such slowing could also enhance the value of existing high-performance computing assets.  If capability scaling slows, the economic life of existing infrastructure increases. In other words, older GPUs should have extended economic life and value, for example. 

Longer frontier model cycles also should mean the useful economic life of each existing model lengthens, allowing the existing models to remain near the performance frontier for a longer period, enhancing the value of each model. 

And all of that might actually help leaders remain leaders, while making it harder for upstarts to break into leadership, as costs related to governance and testing will increase. 

As seen in other industries, large compliance costs raise the costs of doing business, and tend to favor larger firms better able to bear the costs of such compliance. 

Also, if the frontier model releases slow their pace, while more money and effort has to be spent on safety and testing, there is less opportunity for an upstart to move faster and take share, assuming the upstarts have less capital to commit, compared to the model leaders. 

At the same time, the market leaders retain a lead in monetizing their products, for a longer time, at greater scale, while upstarts are disadvantaged, in that regard. 

In fact, most capital-intensive industries tend to have a small number of leaders. Perhaps the only surprise for the capital-intensive frontier model business is that we might be seeing market structure stabilization already. 

 

In addition, in part because of network effects, software industries also tend to have a small number of leaders. The rule is so common as to have a name: “The rule of three.”

“A stable competitive market never has more than three significant competitors, the largest of which has no more than four times the market share of the smallest,” BCG founder Bruce Henderson said in 1976. 

And such structures have profitability implications as well. “Costs characteristically decline by 20 percent to 30 percent in real terms each time accumulated experience doubles," Henderson posited in 1968. 

In most markets, argue Bain consultants, two firms have 80 percent of the profit. In other words, market share also often is a proxy for profitability. “On average, 80 percent of the economic profit pool was concentrated in the hands of just one or two players in each market,” say Bain and Company consultants. In other words, it really matters if a firm is number three in any market.

In mature, competitive markets where there are barriers to entry; economies of scale or are capital intensive, three companies tend to dominate, with the largest player holding 40-50 percent of the market, the second-largest holding 20-30 percent, and the third-largest holding 10-20 percent market share, the rule suggests. 

That virtually perfectly corresponds to a market share pattern of 4:2:1, as the number-three provider tends to have less than 10 percent share, in that pattern. Of course, real life does not always confirm to our predictions, but the theory is often quite useful. 

Industry / market	Approximate structure	Classification	Why the structure develops
Rental cars	48 / 22 / 14% historically	4:2:1	Fleet scale, purchasing, utilization, locations
Agricultural machinery	John Deere / AGCO / CNH	Rule of Three	Manufacturing scale, dealer networks, installed base
Household appliances	Whirlpool / Electrolux / GE	Rule of Three	Manufacturing scale + distribution
Credit ratings	S&P / Moody's / Fitch	Rule of Three, highly concentrated	Reputation, regulatory recognition, scale
Automobiles	Several firms with relatively comparable shares	Oligopoly	Huge fixed costs, but brands, geography and product segmentation prevent 4:2:1
Commercial aircraft	Boeing / Airbus / Embraer	Rule of Three-ish	Enormous R&D, certification and manufacturing barriers
Large commercial engines	GE Aerospace / Rolls-Royce / Pratt & Whitney	Rule of Three-ish	Massive R&D, installed base, certification and aftermarket
U.S. wireless	Verizon / AT&T / T-Mobile	Rule of Three	Spectrum scarcity + network scale + capital intensity
Desktop operating systems	Windows overwhelmingly dominant	Winner takes most	Developer ecosystem + applications + compatibility + switching costs
Search engines	Google overwhelmingly dominant, Bing distant #2	Winner takes most	Query data, relevance, advertiser ecosystem, defaults
Social networking	A few enormous platforms	Winner takes most within individual networks	Network effects are extremely powerful
Mobile operating systems	Android / iOS dominate	Two-sided winner-take-most	Developers + users + apps + hardware ecosystems
Enterprise CRM	Salesforce + Microsoft + Oracle/SAP/etc.	Oligopoly / emerging winner-take-most	Switching costs and installed applications matter, but multiple ecosystems survive
Cloud infrastructure	AWS / Azure / Google	Rule of Three-ish	Enormous capital investment + scale + enterprise relationships
GPUs for AI	Nvidia overwhelmingly ahead, AMD/others behind	Winner takes most—so far	Software ecosystem + CUDA + performance + supply scale
AI frontier models	OpenAI / Anthropic / Google currently prominent, with others competing	Hybrid	Huge compute costs favor concentration, but rapid model improvement and low switching costs favor competition
AI model APIs	Several major suppliers + open-weight models	Not yet settled	Models can be substituted relatively easily; price/performance changes rapidly
AI applications	Many thousands of potential providers	Potentially winner-takes-most at the platform level, fragmented underneath	Distribution, workflow integration and proprietary data may matter more than model ownership
AI agents	Not yet established	Potentially winner-takes-most	Persistent user context, permissions, transactions and ecosystem integration could create powerful switching costs

The “rule of three” seems to explain outcomes in capital intensive industries. But network effects might explain outcomes in software markets that are not so capital intensive, but are highly reliant on scale effects. 

Market structure	Typical share pattern	Main source of advantage	What happens to #3 and smaller firms?	Examples
Rule of Three / 4:2:1	~57% / 29% / 14%	Scale economies, experience curve, fixed costs	#3 remains viable and profitable	Rental cars, machinery, appliances, credit ratings
Oligopoly, but not 4:2:1	~35% / 30% / 20% / 15%	Scale + differentiation + regulation	Several firms remain viable	Autos, parcel delivery, airlines
Winner takes most	60–90% / 10–30% / tiny	Network effects, ecosystem, standards, switching costs, data	#2 can survive, but #3+ often become marginal	Search, desktop OS, social networks, dominant software platforms
Winner takes all	80–100% / near zero	Strong network effects + high switching costs + standardization	Alternative platforms largely disappear	Some historical OS/platform markets
Fragmented / turbulent	Many firms	Innovation, differentiation, low switching costs	Entrants continually replace incumbents	Much of software, consumer technology, early-stage AI

AI is interesting because it contains both forces simultaneously. Data center, cloud computing as a service and frontier AI models are the most capital intensive, and most likely to take a “Rule of Three” shape. 

On the other hand, graphics processing units and agentic AI might tend to take a “winner takes most” structure. Vertical applications might not fit either pattern. 

Layer	Likely structure	Economic reason
Power/data centers	Rule of Three / oligopoly	Capital intensity and physical constraints
GPU/accelerators	Winner-takes-most	Software ecosystem + performance + scale
Cloud	Rule of Three	Capital intensity + enterprise relationships
Frontier models	Three-to-five major competitors	Massive fixed costs but rapid technological change
Model APIs	Competitive oligopoly	Low switching costs and rapid price/performance improvement
AI assistants	Potential winner-takes-most	User relationships, memory, distribution
AI agents	Potentially winner-takes-most	Workflow integration, permissions, transactions
Vertical AI applications	Many winners	Industry specialization and differentiated workflows