On March 24, OpenAI announced the discontinuation of the video generation tool Sora, which had been online for less than six months, while also terminating its $1 billion content licensing agreement with Disney (according to reports from Variety and Bloomberg). Data shows that Sora generated monthly revenue of $367,000 with daily operating costs of $15 million. The ratio of revenue to costs is 0.08%.

The day before, documents obtained by CNBC indicated that Microsoft was listed as the number one risk factor in OpenAI's IPO preparation paperwork sent to investors. Three weeks prior, Awesome Agents reported that ChatGPT's built-in e-commerce feature, Instant Checkout, was quietly shut down due to a conversion rate close to zero.

A company valued at $730 billion is not showcasing growth stories during its IPO sprint but is instead cutting product lines, severing partnerships, and delineating boundaries with its largest shareholder. This does not resemble a contraction; it feels more like a planned narrative cleanup.

Sora: Technological Marvel, Commercial Disaster

When Sora launched in October last year, it was a phenomenal product in the App Store, surpassing 1 million downloads within five days and totaling over 3 million downloads. However, according to third-party data tracking platforms Appfigures and Similarweb, the 30-day retention rate was only about 1%, while TikTok's 30-day retention rate during the same period was 32%.

According to Appfigures data, downloads fell by 45% month-over-month in January, dropping to approximately 1.2 million. January revenue was $367,000, down 32% from a peak of $540,000 in December. Analyst Deepak Mathivanan from Cantor Fitzgerald estimated that Sora generated a peak of 11.3 million videos daily, with a generation cost of about $1.30 per video, leading to daily operating costs of about $15 million and an annual burn rate of approximately $5.4 billion.

For OpenAI, the issue with Sora is not that the technology is flawed, but that the business model does not hold up. Spending $5.4 billion annually for less than $5 million in revenue is poison in any IPO prospectus.

The associated cost of shutting down Sora is the cancellation of the $1 billion investment agreement with Disney. According to Variety, this agreement was originally set for three years, allowing OpenAI to use over 200 characters from Marvel, Pixar, and Star Wars to train and generate content, excluding characters associated with live actors' likeness and voices. Bloomberg reported that this money has actually not yet been delivered.

De-Microsoftization: A 16-Month Systemic Project

Viewing Sora and the Disney agreement as isolated incidents misses a more significant clue. Since Altman was fired and then returned in November 2023, OpenAI has completed a six-step operation over 16 months, systematically downgrading Microsoft from controller to minority shareholder.

The restructuring in October 2025 at PBC was a critical turning point. According to Fortune, Microsoft's stake was diluted from 32.5% to 27%, and exclusive cloud priority rights were simultaneously canceled. Six days after the restructuring was completed, OpenAI signed a $38 billion cloud computing agreement with Amazon, with ESM China reporting that all deployment objectives are expected to be completed by the end of 2026. Amazon's total commitment amounts to $50 billion, of which $15 billion will be paid upfront, while the remaining $35 billion is tied to milestones like the IPO process.

More crucially is the separation of cloud architecture. Microsoft’s Azure retained OpenAI’s stateless API calls (i.e., the basic reasoning services of ChatGPT and API), but Frontier, OpenAI’s stateful enterprise-level agent platform, is exclusively deployed on AWS. According to Windows Central, Microsoft considers this a violation of the original exclusivity clause and is contemplating legal action.

According to analysis from Next Platform, about $281.3 billion of Microsoft's $625 billion revenue backlog comes from OpenAI's Azure procurement commitments, making up 45%. Microsoft's fiscal year 2026 capital expenditure is estimated at $100 to $125 billion, while AI revenue is only about $13 billion annually. This indicates a counterintuitive situation: as OpenAI is busy with its "de-Microsoftization," Microsoft's financial dependency on OpenAI may have deepened.

The Zero Conversion of E-commerce and Narrative Focus

The shutdown of ChatGPT Instant Checkout did not attract much attention, but it points to the same logic as Sora's discontinuation. This feature, launched last September, originally promised integration with over 1 million Shopify merchants, but only about 12 merchants were actually integrated. According to Awesome Agents, by the time of its shutdown, the purchase conversion rate was close to zero, and a state sales tax collection system was never established.

After the shutdown announcement, public market data showed Expedia's stock rose by 13.69%, Booking by 8.46%, and Shopify by 3.96%. The market's interpretation is clear: OpenAI exiting e-commerce is a boon for existing players.

The simultaneous contraction of three product lines points to a common goal: to focus the IPO valuation narrative on the core AI model. According to OpenAI CFO Sarah Friar and official OpenAI data, a revenue target of $20 billion annually, 80 million weekly active users, and 1 million business customers are enough to support a clean growth story. Sora's annualized cost of $5.4 billion, the zero conversion of e-commerce, the legal risks of the Disney agreement, and the uncertainty of Microsoft relations—all of these represent noise on the IPO prospectus.

Pre-IPO "Slimming" is not Unique to OpenAI

In 2020, Uber sold its self-driving division ATG to Aurora for $4 billion in exchange for a 26% stake. ATG had annual R&D expenses of $457 million. After cutting this division, Uber achieved its first profitable quarter post-IPO. Similarly, WeWork also cut non-core businesses before its IPO in 2019, but it did so too late, with its valuation crashing from $47 billion to $8 billion, leading to a failed IPO and bankruptcy in 2023.

OpenAI's actions are more aligned with Uber's: proactively cutting high-cost, low-return business lines and cleaning up its financial structure before going public. The difference lies in scale. Uber's ATG burned $457 million annually, while OpenAI's Sora burned $5.4 billion annually, a difference of one order of magnitude. Added to the canceled $1 billion deal with Disney, the dissolution of the e-commerce team, and the public downgrade of Microsoft relations, OpenAI is conducting the largest business line cleanup in tech history during its IPO sprint.

Altman is telling a growth story through subtraction.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。