A16z: AI cannot avoid advertising, there is huge monetization pressure behind it Author: Bryan Kim Compiled: Deep Tide TechFlow Guide: The Internet is a universal access miracle leading to opportunity, exploration and connection. And advertising pays for this miracle. Bryan Kim, a partner at a16z, pointed out that OpenAI announced plans last month to launch advertisements for free users, which could be the "biggest news not considered news" so far in 2026. Because if you keep paying attention, signs of this happening are everywhere. Advertising is the best way to bring Internet services to as many consumers as possible. The data shows that the conversion rate of consumer AI subscription companies is very low (5-10%). Most people use AI for personal productivity tasks (writing emails, searching for information) rather than high-value pursuits (programming). 5-10% of paid users in 800M WAU are already 40-80m people, but to reach one billion users, advertising is needed. The full text is as follows: The Internet is a universal access miracle leading to opportunity, exploration and connection. And advertising pays for this miracle. As Marc has long advocated, 'If you take a principled stance on advertising, you are also taking a stance on broad access.' Advertising is the reason we have beautiful things. Therefore, OpenAI announced last month its plan to launch advertisements for free users, which may be the "biggest news without news" so far in 2026. Because of course, if you keep paying attention, signs of this happening are everywhere. Fidji Simo joined OpenAI as the CEO of Applications in 2025, which many interpret as' implementing advertising, just like what she did on Facebook and Instacart '. Sam Altman has been previewing the launch of advertisements on commercial podcasts. Technology analysts like Ben Thompson have been predicting ads almost since the launch of ChatGPT. But the main reason why advertising is not surprising is that they are the best way to bring Internet services to as many consumers as possible. The long tail of LLM users' 'luxury beliefs', a term that became popular a few years ago, refers to taking a certain stance not for fundamental reasons, but for optical reasons. There are many examples of this in the technology industry, especially in advertising. Despite all the moral twists on Bingo words such as "sell data!" or "track!" or "attention harvesting," the Internet has always relied on advertising, which most people like. Internet advertising has created one of the greatest "public goods" in history, and the cost is negligible - occasionally I have to see advertisements for cat sleeping bags or hydroponic living room gardens. People who pretend this is a bad thing usually want to prove something to you. Any Internet history enthusiast knows that advertising is the core part of the ultimate monetization of the platform: Google, Facebook, Instagram and TikTok all started for free, and then found monetization through targeted advertising. Advertising can also be a way to supplement the ARPU of low value subscribers, just like Netflix's newer $8 per month option, which introduces advertising into the platform. Advertising is very good at training people to expect most things on the Internet to be free or very cheap. This pattern can now be seen appearing in cutting-edge laboratories, professional modeling companies, and smaller consumer AI companies. From our survey of consumer AI subscription companies, we can see that converting subscription users is a real challenge for all of these companies: So what is the solution? As we know from past consumer success stories, advertising is often the best way to expand services to billions of users. Understanding why most people don't pay for AI subscriptions can help us understand what people do with AI. Last year, OpenAI released data on this. In short, most people use AI for personal productivity: tasks such as writing emails, searching for information, tutoring, or providing advice. At the same time, high-value pursuits such as programming only account for a small portion of total queries. According to rumors, we know that programmers are one of LLM's most loyal users, and some even adjust their sleep schedules to optimize daily usage restrictions. For these users, a monthly subscription fee of $20 or $200 does not seem too high, as the value they receive (equivalent to a group of efficient SWE interns) may be several orders of magnitude higher than the subscription fee. But for users who use LLM for general queries, suggestions, and even writing assistance, the actual payment burden is too great. Why do they pay for answers to questions like 'Why is the sky blue?' or 'What is the reason for the Peloponnesian War?' when Google search used to provide you with a good enough answer for free. Even with writing assistance (some people do use it to complete email and routine work), it often cannot complete enough of a person's work to justify paying for subscriptions. Moreover, most people typically do not require advanced models and features: you do not need the best inference model to write emails or suggest recipes. Let's take a step back and admit something. The absolute number of people paying for products like ChatGPT is still huge: 5-10% of 800M WAU. 5-10% of 800M is 40-80m people! The most important thing is that the Pro's price point of $200 is ten times what we believe is the upper limit for consumer software subscriptions. However, if you want ChatGPT to reach one billion people (or more) for free, you need to introduce products other than subscriptions. The good news is that people actually love advertising! Ask ordinary Instagram users, they may tell you that the ads they receive are very useful: they get the products they truly want and need, and make purchases that truly make their lives better. Classifying advertising as exploitative or invasive is regressive: while we may feel this way about television advertising, most of the time targeted advertising is actually quite good content. I am using OpenAI as an example here (because they have always been one of the most candid laboratories in terms of comprehensive disclosure of usage trends). But this logic applies to all cutting-edge laboratories: if they want to expand to billions of users, they ultimately need to introduce some form of advertising. The consumer monetization model is still unresolved in AI. In the next section, I will introduce some methods. My general rule of thumb in consumer application development for possible AI monetization models is that before introducing advertising, you need at least 10 million WAUs. Many AI laboratories have already reached this threshold. We already know that the advertising unit is about to enter ChatGPT. What could they look like, and what other advertising and monetization models are feasible for LLM? 1. Higher value search and intent based advertising: OpenAI has confirmed that this type of advertising (recipe ingredients, travel hotel recommendations, etc.) will soon be launched for free and low-cost users. These advertisements will be distinguished from the answers in ChatGPT and will be clearly marked as sponsored. Over time, advertisements may feel more like prompts: you will be prompted with the intention to purchase something, and the agent will complete your request end-to-end, selecting from a list of sponsored and non sponsored content. In many ways, these ads remind people of the earliest advertising units in the 1990s and 2000s, as well as the content perfected by Google through its sponsored SEO advertising units (it is worth mentioning that Google still earns the vast majority of its revenue from its advertising business and only entered subscription after more than 15 years of its history). 2. Instagram style context based advertising: Ben Thompson pointed out that OpenAI should have introduced advertising to ChatGPT earlier. Firstly, it will enable non paying users to adapt to advertising earlier (when they have a real lead in Gemini's capabilities). Secondly, it will enable them to take the lead in building truly outstanding advertising products that predict what you want instead of opportunistically providing ads based on intent based queries. Instagram and TikTok can provide amazing advertising experiences, showing you products you never knew you wanted but absolutely needed to buy immediately, and many people find advertising useful rather than intrusive. Given OpenAI's possession of personal information and memory, there is ample opportunity to build similar advertising products for ChatGPT. Of course, there are differences in the experience of using these apps: can you convert the more "backward" advertising experience on Instagram or TikTok to a more engagement focused model using ChatGPT? This is a much more difficult problem, and also a more profitable one. 3. Alliance Business: Last year, OpenAI announced a partnership with marketplace platforms and individual retailers to launch an instant checkout feature that allows users to make purchases directly through chat. You can imagine this being built into your own dedicated shopping vertical, where agents actively search for clothing, household items, or rare items you are tracking because their availability is limited, and model providers receive revenue sharing from the markets showcased in this service. 4. Games: Games are often forgotten or obscured as their own advertising units, and we are unsure how they fit into ChatGPT's advertising strategy, but they are worth mentioning here. Application installation advertising (many of which are mobile games) has been a significant part of Facebook's advertising growth for many years, and games are inherently so profitable that it is not difficult to imagine a large advertising budget here. 5. Goal based bidding: This is interesting for fans of auction algorithms (or former blockchain gas fee optimizers who want to switch to LLM). What if you could set a bounty for a specific query (for example, $10 for Noe Valley real estate alerts) and have the model invest a huge amount of computation on specific results? You will receive perfect price discrimination based on the determined value of the problem, and can also ensure better chain of thought reasoning for searches that are particularly important to you. Poke is one of the best examples of this: people must explicitly negotiate a subscription service with chatbots (although this does not map to computational costs, it is still an interesting illustration of what it may look like). In some ways, this is already the way certain models work: both cursor and ChatGPT have routers that select models for you based on the complexity of the interpreted queries. But even if you select a model from the drop-down menu, you cannot choose the underlying computational cost that the model will invest in the problem. For highly active users, examples of specifying the value of a question to them in US dollars may be attractive. 6. AI Entertainment and Companion Subscription: AI users demonstrate two main use cases where they are willing to pay: coding and companionship. CharacterAI has one of the highest WAU counts among any non laboratory AI company. They can also charge a subscription fee of $9.99 for their service, as they offer a mix of companionship and entertainment. But even if people do pay for partner apps, we haven't seen partner products cross the threshold where they can reliably monetize through advertising. 7. Pricing based on token usage: In the field of AI creative tools and coding, pricing based on token usage is also a common monetization model. This has become an attractive pricing mechanism for companies with premium users, allowing them to differentiate and charge more fees based on usage. Monetization remains an unresolved issue in AI, with the majority of users still enjoying the free tier of their preferred LLM. But this is only temporary: the history of the Internet tells us that advertising will find a way out.