The enterprise AI video generation market just lost two major players in less than three months, creating an unexpected opening for a Chinese competitor that most teams have never heard of. OpenAI discontinued Sora on March 24, 2026, and ByteDance froze Seedance 2.0's international rollout on March 16, 2026. By June 22, 2026, Alibaba Cloud launched HappyHorse 1.1 into that gap, positioning itself as the primary global alternative for teams building AI video workflows.

The two exits happened for entirely different reasons, but both reveal hard truths about the economics and legal challenges of scaling AI video technology. Understanding what went wrong with Sora and Seedance matters because it shapes which tools are actually available to you if you're building video generation into your product or workflow.

Why Did OpenAI Shut Down Sora?

Sora's discontinuation was not about quality. The model could generate convincing video. The problem was cost. Generating video requires orders of magnitude more computing power than generating text or images. A single high-quality video clip involves rendering hundreds of frames at high resolution while maintaining temporal coherence, all faster than a human waits. The hardware cost per output second is vastly higher than language generation.

OpenAI's pricing at the consumer subscription tier simply did not recover those costs. The company announced the shutdown on March 24, 2026, citing the need to focus on sustainable products. The consumer app went dark on April 26, 2026. The API continues through September 24, 2026, giving existing integrations a wind-down window, but no new development is planned.

What makes this significant is that Sora was not a marginal product. It was the most publicly visible AI video model in the world, the model that generated headlines for two years before launch, and that shaped how people thought about AI-generated video. Its exit demonstrates how difficult the unit economics of this category are to get right.

What Happened to ByteDance's Seedance?

Seedance 2.0's situation was structurally different. ByteDance launched the model inside China on February 12, 2026 to strong reception. The plan was a global rollout in mid-March. That plan ended on March 16.

The trigger was a viral clip generated with Seedance 2.0 showing Brad Pitt and Tom Cruise in a rooftop fight. The clip spread widely, and Hollywood's legal teams moved quickly. The Motion Picture Association issued a public statement accusing ByteDance of "large-scale unprotected copyright use." Disney's lawyers reportedly described the model's outputs as a "virtual smash-and-grab of Disney's IP." Warner Bros. Discovery, Paramount Skydance, Netflix, and Sony Pictures all sent separate cease-and-desist letters. US senators Marsha Blackburn and Peter Welch wrote to ByteDance calling Seedance "the most glaring example of copyright infringement from a ByteDance product to date" and demanded an immediate shutdown.

ByteDance responded by freezing the international rollout and promising to implement stronger intellectual property safeguards. As of now, Seedance 2.0 operates only within China's domestic short-drama market at pricing of 28 to 46 yuan per million tokens. A global API is not available.

Who Is HappyHorse and Why Should You Care?

HappyHorse 1.0 arrived in early April 2026 under a name with no corporate affiliation. It entered the Artificial Analysis leaderboard anonymously, climbed to number two in text-to-video with a 1,444 Elo score, and attracted significant attention from the AI video research community before anyone knew who made it. On April 10, 2026, Alibaba confirmed it was the creator. The model came from Alibaba's Taotian Future Life Lab, led by Zhang Di, formerly VP of Kuaishou and the head of the Kling AI technology team, who joined Alibaba at the end of 2025.

The anonymous launch was deliberate. Alibaba wanted benchmark validation before attaching its corporate identity to the product. HappyHorse 1.0 is a 15-billion-parameter unified Transformer that generates synchronized video and audio in a single forward pass. It outputs 1080p video in approximately 38 seconds on a single H100 GPU. Native lip sync works across seven languages out of the box. The model became available via FAL.ai on April 27, 2026, with early-access pricing and a 10% discount for API integrators.

What's New in HappyHorse 1.1?

HappyHorse 1.1, released June 22, 2026, is an enterprise-focused upgrade that addresses three specific pain points teams face when generating video at scale.

• Zero-drift lip sync for dialogue: This addresses one of the most visible failure modes in AI video, where avatar or talking-head content shows lip movement that diverges from the audio over the course of a sentence or paragraph. Version 1.1 specifically targets extended dialogue rather than short clips, which matters for product demos, training content, and anything longer than 10 to 15 seconds.
• Video editing modality: HappyHorse 1.0 supports text-to-video, image-to-video, and reference-video generation. Version 1.1 adds a fourth capability: video editing mode, where the model takes an existing clip as input and modifies it per natural-language instructions. This means HappyHorse 1.1 can now sit inside editing workflows, not just generation workflows.
• Context-aware speech pacing: Generated voiceover has traditionally produced speech at a fixed or manually-specified pace. Context-aware pacing adapts delivery speed based on the content's emotional register, faster for energetic content, slower for explanatory or instructional scenes. For teams generating training videos or product explainers at scale, this reduces the manual voiceover correction pass that often follows AI generation.

HappyHorse 1.1 is available now on Alibaba Cloud Model Studio with full API access. A 40% launch discount applies for the first two weeks.

How to Evaluate AI Video Models for Your Workflow

• Benchmark performance: Check where the model ranks on the Artificial Analysis Video Arena, which uses blind human preference votes to score models on Elo, the same method used in competitive chess. A higher Elo means more head-to-head preference wins against other models. HappyHorse 1.0 holds 1,444 Elo in both text-to-video and image-to-video as of June 2026, ranking number two globally and ahead of Google Veo 3.1 with audio by 69 points in the text-to-video category.
• Compute efficiency: Consider how much computing power the model requires per output second. HappyHorse 1.0 outputs 1080p video in approximately 38 seconds on a single H100 GPU, which is relevant if you're calculating infrastructure costs for your workflow.
• Feature completeness: Evaluate whether the model supports the specific modalities you need, such as text-to-video, image-to-video, video editing, or synchronized audio generation. Different models excel at different tasks.
• Language and localization support: If you're generating content in multiple languages, verify native support. HappyHorse 1.0 includes native lip sync across seven languages out of the box.
• Global API availability: Confirm that the model is available through a global API with clear pricing and support. Sora's exit and Seedance's regional restriction demonstrate that global availability is not guaranteed.

The compressed competitive field matters because it reduces your options. Teams that were planning to build on Sora or Seedance now need to migrate to alternatives. HappyHorse 1.1's combination of strong benchmark performance, enterprise-focused features, and global availability positions it as the primary option for teams that need a reliable, scalable AI video API.

The real lesson from Sora and Seedance is that the AI video market is still finding its footing. Economics and legal liability are both real constraints on which models survive at global scale. For teams building video workflows, that means choosing a provider with sustainable unit economics and robust intellectual property safeguards is as important as raw model quality.