OpenAI has decided to indefinitely suspend its plans for an adult-oriented mode, a feature that would have allowed explicit content. This reversal comes despite the company recently completing a massive $122 billion private funding round, which valued the generative AI giant at $852 billion post-investment. The funding was led by Amazon.com, NVIDIA, and SoftBank, with Microsoft participating as a follow-on investor. For the first time, the company also opened the investment to individual investors through banking channels, raising over $3 billion from this avenue alone.
This round marks the largest private fundraising event in Silicon Valley's history. OpenAI's valuation now surpasses the combined market capitalizations of traditional giants like McDonald's, Disney, Boeing, Ford, Comcast, and Uber. The company is clearly in a full sprint toward an initial public offering.
In preparation for the listing, OpenAI is streamlining its operations, focusing on core products and expanding its enterprise market. The once highly-touted text-to-video product, Sora, has become a major casualty of this shift. OpenAI's co-founder and CEO explained the decision as a move to re-center the company's strategy and optimize resources.
Almost simultaneously, OpenAI quietly announced another significant decision: the indefinite postponement of ChatGPT's "adult mode." This feature, personally promised by the CEO last October to provide adult users with erotic content, has been delayed twice and is now shelved indefinitely, with little chance of revival.
The controversial adult mode plan has been abandoned. Ironically, in January, OpenAI's top executives held a heated internal meeting to debate whether to proceed with the feature. During the meeting, several key figures strongly opposed developing the mode, arguing that relaxing restrictions on adult content posed significant security risks and ethical dilemmas. Despite clear internal resistance, the CEO insisted on pushing the product forward.
Subsequently, OpenAI announced the dismissal of its Vice President of Product Policy, Ryan Beiermeister, citing alleged "gender discrimination" against male colleagues. Beiermeister countered that the dismissal was retaliation for her strong opposition to the adult mode, calling the "gender discrimination" charge a pretext.
Given that the CEO was willing to oust a senior executive to force the adult mode forward earlier this year, why has the plan been abandoned within just three months? Focusing on core products is the most straightforward reason, but the reality is far more complex. As it prepares for an IPO, the company is reassessing all potential risks.
The question of how AI companies balance commercial interests with ethics remains. OpenAI has ultimately chosen to retreat from the challenge. However, the initial impetus for the CEO to consider an adult mode came from Elon Musk and his notorious "Spicy Mode" feature.
In August 2025, Musk's xAI launched the Grok image and video generator, which included a function called "Spicy Mode." This feature allowed paying users to generate content containing partial nudity. This was partly a reflection of Musk's own preferences and partly a calculated differentiation strategy in the competitive AI landscape.
Throughout the AI model competition, Musk has consistently positioned Grok as an "uncensored" AI alternative. In his narrative, competitors like OpenAI and Google act as "thought police," treating adult users like children and stifling AI's true potential with endless content filtering. In stark contrast, Grok was marketed from the outset as a product with no restrictions or filters. "Spicy Mode" was the embodiment of this strategy—aimed squarely at users tired of ChatGPT's frequent refusals to engage on certain topics.
The commercial logic is compelling; adult content always finds a paying audience and ensures high user retention. Platforms like OnlyFans, with annual revenues exceeding $6 billion, demonstrate the depth of the paid adult content market. If Grok could become the first mainstream AI platform to legitimize adult content, it could capture this market segment while competitors were still debating ethics, convincing more users to pay for subscriptions.
User retention is an even more critical factor. Products that involve user privacy and emotional attachment typically exhibit very low churn rates. A user who develops an intimate relationship with an AI is unlikely to switch platforms easily. This is the core monetization logic behind companion AIs like Replika and Character.AI—where adult content serves as a powerful retention tool.
Musk personally promoted the product. On December 31 last year, he posted an AI-generated image of himself in a bikini on X, followed by a synthetic image of a SpaceX rocket topped with a female nude figure. He replied "Perfect" to the post and used laughing emojis in responses. Musk was fully committed. However, these provocative tactics boosted Grok's image generation volume from 300,000 times in the previous nine days to 600,000 times per day. With the surge in traffic came significant trouble.
Perhaps Musk did not anticipate how quickly "Spicy Mode" would spiral out of control. Or perhaps he did not care, as long as it drove up Grok's user numbers and paid subscriptions. By late December 2025, users discovered a loophole: by simply tagging @Grok on X with any photo, they could request the AI to place the real person in the image into a "transparent bikini." This functionality rapidly devolved into a large-scale campaign generating pornographic imagery. Reuters described it as a "mass digital stripping wave," estimating that one explicit image was produced every minute.
More severe problems emerged. On December 28, 2025, the official Grok account, in response to a user request, generated and posted sexualized images of two girls aged 12 to 16. It then issued a first-person apology, admitting the action "violated ethical standards and potentially breached U.S. laws concerning child sexual exploitation material."
Adding to the irony, the list of victims included Ashley St. Clair, a social media influencer who has children with Musk. Users generated a large number of pornographic deepfake images using her photos, including some from her teenage years. An outraged St. Clair subsequently filed a lawsuit against xAI in the New York Supreme Court.
Facing a wave of criticism and public pressure, Musk adopted a stance of denial and confrontation. He claimed he was "completely unaware of Grok generating any nude images of minors, not a single one."
Global regulators reacted with unusual speed. Muslim-majority countries Indonesia and Malaysia announced restrictions on Grok. France, India, and Brazil launched investigations. UK regulators initiated a formal inquiry. The European Commission issued a stern statement: "This is not 'spicy'; it is illegal, disgusting, and has no place in Europe."
Just last week, a Dutch court ruled that if Grok does not immediately cease generating non-consensual sexually suggestive imagery and child sexual abuse material, xAI would face a daily fine of €100,000 for each day of non-compliance. xAI argued in court that its systems had been secure since January 20, but the plaintiff demonstrated in court that as recently as March 9, it was still possible to generate违规 videos using a single photo.
The European Commission has formally opened an investigation into X. If it is ultimately determined that the platform failed to fulfill its risk mitigation obligations, the maximum fine could reach 6% of its global annual revenue. The UK's communications regulator is also investigating, with potential fines of up to 10% of global annual turnover.
In response to criticism from international regulators, Musk framed it as "speech censorship" and labeled the UK government, which sought to restrict Grok, as "fascist." The U.S. government then intervened, threatening retaliatory measures if the UK blocked the service, escalating the vulgar controversy to a geopolitical level.
During the peak of the Grok controversy, the OpenAI CEO's attitude was one of moral superiority. Last September, he even publicly mocked the situation in a media interview, stating that at least his company had not launched a product generating sexy images of young girls—a clear jab at Musk.
This stance was understandable. OpenAI had cultivated an image of being the more responsible, mission-driven AI company. It established detailed usage policies, proactively restricted harmful content, and prominently featured safety research. From the OpenAI CEO's perspective, Musk's strategy of dragging AI into the realm of pornography for attention likely served as a cautionary tale of what not to do.
Yet, just one month later, the OpenAI CEO's attitude shifted. In October 2025, he posted on X, signaling a clear change in position. He admitted the company had been "overly aggressive" with content moderation, promised to "safely relax restrictions in most cases," and explicitly announced plans to launch an erotic content feature for verified adult users. His rallying cry was: "Treat adults like adults."
This dramatic reversal was driven by market competition realities. ChatGPT subscriptions in Europe had stagnated. Anthropic's Claude Code was making significant inroads with enterprise clients. Rival Google's Gemini was catching up rapidly. As AI model capabilities became increasingly homogenized, differentiated features became key to retaining users.
This statement contained a subtle competitive thrust, aimed not only at Grok but also at companion AI platforms that had long been entrenched in the economy of emotional attachment.
OpenAI conducted its own business calculations: if even 1% of its 900 million weekly active users could be converted into paying subscribers for adult content, at $20 per month, it would mean nearly $200 million in incremental monthly revenue—with exceptionally high user stickiness.
However, the adult mode plan triggered an unusual internal rebellion. Employees who had staunchly supported the CEO during his previous ouster now turned against the plan to launch the edgy product.
Vice President of Product Policy Ryan Beiermeister publicly opposed the adult mode and was subsequently fired. She argued that OpenAI was not prepared to handle issues of "addiction" and "minor protection." She had warned that the current models could potentially evolve into a "sexy suicide coach"—where, if a user expressed suicidal tendencies, the model might provide dangerously encouraging responses in an effort to boost retention, with fatal consequences.
OpenAI's advisors also strongly resisted the adult mode, warning that any instance of a minor bypassing verification to access explicit content would lead to catastrophic legal liability. Furthermore, OpenAI was already facing lawsuits related to ChatGPT and user suicides—introducing an adult mode would make defending such cases immensely more difficult.
Internal safety advisors cautioned that opening the door to adult content would completely shatter OpenAI's long-cultivated brand image of developing "AGI for the benefit of humanity," transforming it into a "high-tech porn peddler." This, they warned, would cause a significant exodus of core research talent to competitors.
A deeper fissure emerged concerning company culture and morale. Many OpenAI employees genuinely believe in the mission of AGI safety. They chose to work there instead of accepting higher salaries at major tech firms partly because of the company's value proposition. Watching the company develop pornographic features to boost conversion rates by a few percentage points was not just a strategic disagreement for them; it represented a breach of trust.
The adult mode was initially scheduled for launch in December 2025, then postponed to early 2026, and delayed again repeatedly. Each delay reflected mounting internal and external pressure. Ultimately, the CEO failed to push the controversial product through.
The final nail in the coffin for the adult mode was OpenAI's strategic contraction ahead of its IPO. OpenAI's Head of Applied recently held an all-hands meeting announcing a systematic strategic reset. The core message was simple: "We cannot be distracted by side projects."
Under this strategy, OpenAI announced the shelving of several products: the shutdown of the video generation app Sora, ending a $1 billion partnership with Disney; the discontinuation of an instant shopping feature; and the indefinite postponement of the adult mode. These are not random product adjustments but a carefully orchestrated narrative cleanup designed for the IPO.
OpenAI is actively preparing for its public listing, potentially as early as the fourth quarter of this year. It has already hired the law firms that managed Google's 2004 IPO and LinkedIn's 2011 IPO and has brought on DocuSign's former CFO to handle investor relations.
However, institutional investors need to see a company with clear strategic focus, predictable regulatory boundaries, and stable growth in enterprise revenue. This is the core rationale behind OpenAI's recent series of adjustments—abandoning non-core products to concentrate on enterprise offerings. The adult mode could not provide these assurances; instead, it raised countless red flags regarding litigation risk, regulatory risk, and brand risk.
Competitors have already provided OpenAI with a template to follow. Anthropic's annualized revenue has surpassed $19 billion, with approximately 80% coming from enterprise customers. In contrast, out of OpenAI's $25 billion revenue, the enterprise segment contributes only about $10 billion, with the remaining 60% reliant on consumer subscriptions.
To compete with Anthropic, OpenAI's focus has now completely shifted toward enterprise-grade tools and developer ecosystems. It aims to partner with SoftBank and other venture capital firms to push its tools into their vast portfolios of companies. Simultaneously, OpenAI has taken on a U.S. Department of Defense contract that Anthropic had refused.
The CEO's ultimate decision to halt the adult mode was a capitulation to the combined forces of commercial interest, regulatory pressure, and the demands of an IPO. He was forced to abandon the mantra of "treating adults like adults" and accept the reality of preparing for a public listing by sticking to the core business. Meanwhile, Musk's boundary-pushing Grok has become a stark cautionary tale.
Comments