Former a16z partner and renowned technology analyst Benedict Evans recently published an in-depth analysis article, pointing out four fundamental strategic dilemmas faced by OpenAI behind its apparent prosperity. He believes that despite OpenAI's large user base and ample capital, it lacks a technological moat, has insufficient user stickiness, faces rapidly catching competitors, and is constrained by the direction of laboratory research in its product strategy, all of which threaten its long-term competitiveness.
Evans pointed out that OpenAI's current business model does not have a clear competitive advantage. The company has neither unique technology nor has it formed network effects, with only 5% of its 900 million weekly active users paying, and 80% of users sending fewer than 1,000 messages in 2025—equivalent to an average of less than three prompts per day. This &#34;a mile wide and an inch deep&#34; usage pattern indicates that ChatGPT has not yet become a daily habit for users.
Meanwhile, tech giants like Google and Meta have technically caught up with OpenAI and are leveraging their distribution advantages to capture market share. Evans believes that the real value in the AI field will come from new experiences and application scenarios that have yet to be invented, and OpenAI cannot create all these innovations alone. This forces the company to fight on multiple fronts, comprehensively laying out from infrastructure to application layers.
Evans's analysis reveals a core contradiction: OpenAI is trying to establish competitive barriers through large-scale capital investment and a full-stack platform strategy, but whether this strategy can be effective in the absence of network effects and user lock-in mechanisms remains questionable. For investors, this means a need to reassess OpenAI's long-term value proposition and its true position in the AI competitive landscape.
<h2>Disappearance of Technological Advantage: Model Homogenization Intensifies</h2>
In his analysis, Evans pointed out that currently about six institutions are capable of launching competitive cutting-edge models, and their performance is basically comparable. Companies surpass each other every few weeks, but none can establish a technological lead that others cannot match. This contrasts sharply with platforms like Windows, Google Search, or Instagram—where network effects have self-reinforced market share, making it difficult for competitors to break the monopoly regardless of how much capital and effort they invest.
This technological equalization may change due to certain breakthroughs, the most obvious being the realization of continuous learning capabilities, but Evans believes OpenAI is currently unable to plan for this. Another potential differentiating factor is the scale effect of proprietary data, including user data or vertical industry data, but existing platform companies also have advantages in this regard.
Against the backdrop of converging model performance, competition is shifting towards branding and distribution channels. The rapid growth of market share for Gemini and Meta AI confirms this trend—these products appear largely similar to ordinary users, while Google and Meta possess strong distribution capabilities. In contrast, although Anthropic's Claude model often ranks high in benchmark tests, its consumer awareness is close to zero due to a lack of consumer strategy and products
Evans compares ChatGPT to Netscape, which once held an early advantage in the browser market but was ultimately defeated by Microsoft's distribution advantages. He believes that chatbots face the same differentiation dilemma as browsers: they are essentially just an input box and an output box, with very limited room for product innovation.
<h2>Fragile User Base: Scale Cannot Mask Lack of Stickiness</h2>
Despite OpenAI's significant lead with 800 to 900 million weekly active users, Evans points out that this figure masks a serious user engagement issue. The vast majority of users who are aware of and know how to use ChatGPT have not cultivated it into a daily habit.
Data shows that only 5% of ChatGPT users pay, and even among American teenagers, the proportion using it a few times a week or less is much higher than those using it multiple times a day. OpenAI disclosed in its &#34;2025 Annual Summary&#34; event that 80% of users sent fewer than 1,000 messages in 2025, which translates to an average of less than three prompts per day, with actual chat frequency being even lower.
This shallow usage means that most users do not see the differences in personality and focus between different models and cannot benefit from features like &#34;memory&#34; that are intended to build stickiness. Evans emphasizes that the memory feature can only bring stickiness, not network effects. While usage data from a larger user base may be an advantage, it is questionable how significant that advantage is when 80% of users use it only a few times a week.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/83e52347-a34c-46db-a1b7-4c9afd893f4c.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" original-src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/83e52347-a34c-46db-a1b7-4c9afd893f4c.png"/>
OpenAI itself acknowledges the problem, stating that there is a &#34;capability gap&#34; between model capabilities and actual user usage. Evans believes this is an avoidance of the fact that product-market fit is unclear. If users cannot think of what to do with it on ordinary days, it indicates that it has not changed their lives.
The company launched an advertising program partly to cover the service costs for over 90% of non-paying users, but more strategically, it allows the company to offer these users the latest, most powerful (and also most expensive) models in hopes of deepening user engagement. However, Evans questions whether providing better models will change the situation if users cannot think of what to do with ChatGPT today or this week.
<h2>Doubts About Platform Strategy: Lack of True Flywheel Effect</h2>
Last year, OpenAI CEO Sam Altman attempted to integrate the company's various initiatives into a coherent strategy, showcasing a chart and quoting Bill Gates: the definition of a platform is that the value created for partners exceeds the value created for itself. Meanwhile, the CFO released another chart showing the &#34;flywheel effect.&#34;
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/99f347e0-2c49-46d0-a2cd-7e4e7681d356.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" original-src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/99f347e0-2c49-46d0-a2cd-7e4e7681d356.png"/>
Evans believes that the flywheel effect is a clever and coherent strategy: capital expenditure itself creates a virtuous cycle and becomes the foundation for building a full-stack platform company. Starting with chips and infrastructure, each layer of the technology stack is built upwards, and the higher you go, the more you can help others use your tools to create their own products. Everyone uses your cloud, chips, and models, and then at higher levels, the layers of the technology stack reinforce each other, forming network effects and ecosystems.
However, Evans candidly states that he does not think this is the correct analogy; OpenAI does not possess the kind of platform and ecosystem dynamics that Microsoft or Apple once had, and that flywheel diagram does not actually demonstrate a true flywheel effect.
In terms of capital expenditure, the four major cloud computing companies invested about $400 billion in infrastructure last year and announced plans to invest at least $650 billion this year. OpenAI claimed a few months ago that it has a commitment of $1.4 trillion and 30 gigawatts of computing power for the future (with no specific timeline), while the actual usage by the end of 2025 is projected to be 1.9 gigawatts. Due to the lack of large-scale cash flow from existing businesses, the company aims to achieve these goals through financing and utilizing others' balance sheets (partly involving &#34;recurring revenue&#34;).
Evans believes that large-scale capital investment may only secure a seat at the table rather than a competitive advantage. He compares the costs of AI infrastructure to those in the aircraft manufacturing or semiconductor industries: there are no network effects, but the craftsmanship of each generation of products becomes more difficult and expensive, ultimately leaving only a few companies able to sustain the investment required to stay at the forefront. However, while TSMC has a de facto monopoly in the cutting-edge chip sector, this has not provided it with leverage or value capture capabilities in the upstream technology stack.
Evans points out that developers must build applications for Windows because it has almost all the users, and users must buy Windows PCs because it has almost all the developers—this is the network effect. But if you invent an outstanding new application or product using generative AI, you only need to call the foundational model running in the cloud via an API; users do not know or care what model you used.
<h2>Lack of Product Leadership: Strategy Constrained by the Laboratory</h2>
At the beginning of the article, Evans quotes a statement from OpenAI's product head, Fidji Simo, in 2026: &#34;Jakub and Mark set the long-term research direction. After months of work, amazing results emerge, and then researchers contact me saying, 'I have something cool. How do you plan to use it in chat? How will it be used in our enterprise products?'&#34;
This statement sharply contrasts with Steve Jobs' famous quote from 1997: &#34;You have to start with the customer experience and work backward to the technology. You can't start with the technology and then figure out where you're going to try to sell it.&#34;
Evans believes that when you are the product head of an AI lab, you cannot control your roadmap, and your ability to set product strategy is very limited. You open your email in the morning to find out what the lab has produced, and your job is to turn it into a button. Strategy happens elsewhere, but where?
This issue highlights the fundamental challenge faced by OpenAI: unlike Google in the 2000s or Apple in the 2010s, OpenAI's smart and ambitious employees do not have a truly effective product that others cannot replicate. Evans believes that one interpretation of OpenAI's activities over the past 12 months is that Sam Altman has a profound awareness of this and is trying to convert the company's valuation into a more enduring strategic position before the music stops.
For most of last year, OpenAI's answer seemed to be &#34;everything, all at once, executed immediately.&#34; Application platforms, browsers, social video apps, collaborations with Jony Ive, medical research, advertising, and so on. Evans believes that some of these appear to be a &#34;full-scale assault,&#34; or merely the result of quickly hiring a large number of aggressive individuals. Sometimes it also gives the impression that people are replicating the forms of previously successful platforms without fully understanding their purposes or dynamic mechanisms.
Evans repeatedly uses terms like platform, ecosystem, leverage, and network effects, but he acknowledges that these terms are widely used in the tech industry and their meanings are quite vague. He quotes his medieval history professor from college, Roger Lovatt: &#34;Power is the ability to make people do things they do not want to do.&#34; This is the real issue: does OpenAI have the ability to get consumers, developers, and businesses to use its systems more, regardless of what the systems themselves actually do? Microsoft, Apple, and Facebook once had this ability, and so did Amazon.
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/53c550d8-a058-4a00-9b9e-203477221182.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" original-src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/53c550d8-a058-4a00-9b9e-203477221182.png"/>
Evans believes that a good way to interpret Bill Gates' statement is that what platforms truly achieve is leveraging the creativity of the entire tech industry, so you don't have to invent everything yourself, allowing for the large-scale construction of more things, all done on your system and under your control. The foundational models are indeed multipliers, and a lot of new things will be built with them. But do you have a reason to make everyone have to use your product, even if competitors have built the same thing? Is there a reason to ensure that your product is always superior to competitors, regardless of how much money and effort they invest?
Evans concludes that without these advantages, the only thing you have is daily execution. Executing better than others is certainly a desire; some companies have achieved this over longer periods and even convinced themselves that they have institutionalized it, but this is not a strategy

MSFO

AAPB

AAPU

Benedict Evans stated that the lack of technological moats, insufficient user stickiness, the absence of flywheel effects in platform strategy, and product strategy being constrained by laboratory research directions are all threatening OpenAI's long-term competitiveness. Evans pointed out that the real issue is whether OpenAI has the ability to get consumers, developers, and enterprises to use its systems more, regardless of what the systems actually do. Microsoft, Apple, and Facebook once had this capability, and so did Amazon

- Benedict Evans highlights four fundamental strategic dilemmas facing OpenAI despite its large user base and capital.  
- The company lacks a technological moat, user engagement, and effective competitive advantages, as evidenced by low user retention—only 5% of its 900 million weekly active users pay for the service.  
- Evans argues that OpenAI's attempts to establish barriers through capital investment may not yield lasting competitive benefits, necessitating a reevaluation of its long-term value proposition and position in the AI landscape.

Beneath the surface glamour, OpenAI's "Four Dilemmas"