Sora is dead. Not winding down, not pivoting — discontinued. OpenAI's help center has carried the notice for months now: the consumer web and app closed in April, and the API followed on September 24. The flagship that was going to remake Hollywood, that had directors and agencies and a PR campaign worthy of a studio launch, is simply gone.
What happened. The timeline is quieter than the story deserves. April 2026: Sora's web and app shut, with export windows and refund notices. September 24: the API closed, leaving businesses that had built on it scrambling — migration guides appeared within days. OpenAI never gave it a funeral; the help-center article did the talking. That silence is itself the first angle: this wasn't announced as a strategy, it was discovered as a fact.
Why it matters more than one product. Sora's death is the closing bracket on the AI-video land rush. Read the pricing record alongside it: Google raised Veo 3.1 to $0.40 per second — double its launch price — and pushed the cheap tiers to 720p only. Pika raised prices ~25% this cycle. Runway held. The pattern of a market shaking out is unmistakable: the subsidized era of consumer AI video is over, the survivors are charging like incumbents, and the most-hyped entrant didn't survive to see it. Consolidation, in one sentence: fewer players, higher prices, and one very large grave.
The angle beneath. The same week OpenAI buried Sora, it closed $11 billion at a $1.25 trillion valuation — while Ramp's card data showed AI spending flatlining and Anthropic's IPO priced with public-market skepticism. The signal isn't that AI is failing; it's that focus has arrived. Compute, enterprise revenue, and agents won; generative-video demo glory lost. Even OpenAI's record raise reads correction-shaped — an initial portion of a round, scaled-back goals, belief rationed. Belief and receipts have decoupled, and Sora is what the gap looks like when it closes on something.
Where the users went. That's the open question worth tracking. Sora's refugees — creators, small studios, API integrators — are choosing between Veo's raised prices, Runway's steady subscriptions, Kling and Pika's tiers, and open models catching up fast. The first migration guides (Seedance's among them) are already competing for them. Whichever platform absorbs Sora's audience inherits the market's second act.
Takeaway. For builders: single-vendor dependency on frontier media models is now a documented risk — architect for migration before the help-center article appears. For everyone else: the AI-video correction is not a crash, it's a sorting. The gold-rush shovels are being inventoried; the towns that survive are the ones that charge rent.
Story also referenced in today's AI & Tech News roundup. Sources: OpenAI Help Center — what to know about the Sora discontinuation, migration analysis.