Sam Altman has finally acted on a tension that has defined the AI boom: people want powerful chatbots, but they do not want to pay for them. OpenAI is rolling out advertising inside ChatGPT, a move its chief executive once described as a “last resort,” and he is now openly acknowledging that “a lot of people want to use a lot of AI and don’t want to pay.” The shift marks a turning point for how cutting edge AI will be funded, and for how much commercial pressure users will have to tolerate in exchange for a $0 bill.
From “Last Resort” to live product
For years, Sam Altman resisted the idea of turning ChatGPT into an ad-supported product, warning that it would only be considered if every other option failed. At an event at Jan in 2024, he framed ads as something he would do only if it was the “only way to get everyone in the world to be able to use this technology,” a clear signal that he saw it as a compromise rather than a goal. That language, echoed in later coverage of his comments, set expectations that OpenAI would exhaust subscriptions and enterprise deals before letting marketers into the chat window.
Yet the company is now moving ahead with what earlier reports described as a Last Resort business model, confirming that the line has been crossed. During that Harvard University appearance, Altman was explicit that advertising made him uneasy because of the incentives it creates, but he left the door open if it was the only way to keep access broad. The fact that OpenAI is now embracing that option suggests the financial and infrastructure pressures around ChatGPT have become too large to ignore.
Altman’s blunt admission: users want AI, not invoices
Altman is not pretending this is about anything other than user behavior and money. In recent comments, he put it plainly: “A lot of people want to use a lot of AI and don’t want to pay,” a line that has quickly become shorthand for the company’s new realism about its audience. That quote, attributed directly to Sam Altman, captures the core tension: the appetite for generative AI is enormous, but the willingness to cover its costs is limited. I read that as an admission that the original bet on subscriptions alone was not enough to sustain the scale of usage OpenAI has encouraged.
OpenAI has been equally direct in its own framing, saying “It is clear to us that a lot of people want to use a lot of AI and don’t want to pay, so we are hopeful a business model like this will work.” That line, shared in coverage of the company’s ad plans, underscores how the free tier has become both a growth engine and a financial liability. The company is now testing ads in ChatGPT’s free experience and in a new $8 per month ChatGPT Go plan in the United States, a detail highlighted in reporting on the Jan announcement. The message is simple: if users insist on $0 access, they will now help pay with their attention instead.
Historic losses, trillion‑dollar commitments, and a $200 warning sign
Behind the rhetoric is a balance sheet that has been straining under the weight of AI’s infrastructure demands. One detailed account notes that OpenAI now has roughly $1.4 trillion in spending commitments on data centers and related infrastructure, a staggering figure that would be hard to support even with aggressive subscription growth. When I look at that number, it is clear why the company can no longer treat monetization as an afterthought. The cost of serving billions of queries with large models is not theoretical; it is locked into long term contracts and hardware.
There were earlier warning signs that the existing pricing structure was unsustainable. An analysis of ChatGPT’s finances pointed out that the fate of the ChatGPT Pro plan was “particularly symptomatic,” noting that Sam Altman publicly admitted in September 2025 that the $200 paying premium customers were unprofitable. If even high end subscribers could not cover their own compute costs, then a free tier with no offsetting revenue was always going to be structurally fragile. The new ad model is, in that light, less a surprise than a delayed acknowledgment of the economics Altman had already described.
How the ads will actually look inside ChatGPT
OpenAI is trying to reassure users that the new commercial layer will not quietly bleed into the model’s core answers. The company has said it is starting to test ads in ChatGPT free and in the Go plan, and that it is “sharing our principles early on how we’ll approach ads, guided by putting user trust and transparency first.” That language comes directly from a post by Altman on X, where he outlined that the tests would begin in Jan. Reporting on the rollout adds that ads are coming to the free tier and the new $8 per month ChatGPT Go plan in the US, and that these Ads will be visually separated from the answer, a detail highlighted in coverage by Benj Edwards, who noted the figure 177 in that context.
Early descriptions of the interface suggest that free users in the United States will see promotions at the bottom of responses, clearly labeled as advertising and visually distinct from the chatbot’s text. One analysis framed this under the heading “What This Means for Users,” explaining that free ChatGPT users in the United States should expect ads at the bottom of responses, with OpenAI trying to keep the core interaction intact while accepting the “baggage of advertising incentives.” That perspective, shared in a Jan commentary, captures the tradeoff: the chat window stays mostly clean, but the business logic around it changes. Separate reporting notes that OpenAI has told regulators and the public that ads will be “visually separated from ChatGPT’s answers,” a point repeated in coverage by Khaosod English, which also referenced an AP Photo by Michael Dwyer showing the ChatGPT homescreen.
Trust, backlash, and the politics of “last resort”
The move has already triggered a debate about whether users should trust OpenAI’s assurances that ads will not distort answers. One widely shared piece opened with the line “Here we go,” reminding readers that OpenAI CEO Sam Altman once called advertising a “last resort” and warning that ChatGPT is about to be “stuffed with ads.” That commentary, attributed to Rich Stanton, reflects a broader skepticism that any ad funded system can fully insulate its core product from commercial influence. I see that concern as especially acute in AI, where even subtle shifts in training data or ranking could tilt answers toward paying partners.
Altman has tried to get ahead of that criticism by addressing the change directly. In one account of his response, a section titled “Altman Addresses the Change” notes that following the announcement, he posted on X to stress that “Most importantly, we will keep the product experience great,” while acknowledging that the company had resisted ads less than two years ago. That reassurance, described in coverage that quoted Altman Addresses the, is an attempt to square his earlier reluctance with the current reality. Another report revisited his Harvard University remarks, where Altman called ads a Last Resort for a business model, underscoring how far the company has moved in a short time.
More from Morning Overview