Loading page…
The product is designed to turn existing material such as presentations and spreadsheets into video, while Alibaba is funding a more capital-intensive AI push.
Listen to this story
Alibaba is expanding access to Wan3.0, a file-to-video system that can turn PDFs, slide decks, spreadsheets, web pages, and other media into clips rather than requiring a blank prompt. It supports up to 30 seconds—double Wan 2.7’s stated limit—and up to 1080p, with usage priced at $0.05–$0.20 per second. The rollout offers teams a low-cost way to repurpose existing business materials, but Alibaba has not provided independent evidence on how accurately the model extracts and structures source information.
Wan3.0 accepts text, images, audio, video, PDFs, PowerPoint files, spreadsheets, and web pages.
Pricing ranges from $0.05 per second at 480p to $0.20 at 1080p.
Alibaba claims consistent characters, objects, backgrounds, spatial layouts, micro-expressions, and multilingual voice output; these claims are not independently evaluated here.
Alibaba has formally rolled out Wan3.0, a video-generation model built to turn business source material into clips of up to 30 seconds. The differentiator is the input: documents, spreadsheets, slides, and web pages can become video rather than merely serving as background material for a human editor.
The workflow Alibaba describes is a conversion step: feed Wan3.0 an existing document or digital asset, and the model generates a corresponding video. Alibaba Cloud says the model can work from text, images, audio, video, PDFs, PowerPoint presentations, and spreadsheets; its other stated inputs include web pages.
That design shifts the prompt from an open-ended instruction toward an organization’s pre-existing information. A presentation can supply the subject matter, while a spreadsheet or web page can supply the raw material for a video asset. The evidence does not specify how Wan3.0 selects, summarizes, or structures information from any one file type.
The company says Wan3.0 maintains consistency in characters, objects, backgrounds, and spatial layouts across a video. It also says the model supports faces with synchronized micro-expressions and multilingual voice output. Those are company-stated capabilities, not independently evaluated performance measures in the supplied evidence.
Wan3.0 entered public beta on August 6 through Alibaba Cloud’s Model Studio and Qwen Cloud, and Alibaba is now rolling out wider access. Alibaba says users had already applied the beta to short dramas and film production, advertising and marketing, tourism promotion, and music-video creation.
The early use cases map to work where a team already has a campaign brief, destination information, product material, or a presentation to adapt. A 30-second ceiling still sets a clear boundary: Wan3.0 is positioned around short-form video units, not a single-pass substitute for a longer production.
The rollout arrived one day after Alibaba announced a $10 billion share placement to fund rising AI spending. The company also reported that quarterly earnings fell 75% from a year earlier, attributing the decline to soaring AI-related capital expenditure.
That financial context makes Wan3.0 more than a feature update. Alibaba is pairing a practical, priced video service with an announced effort to finance higher AI investment. Whether file-to-video generation becomes a durable business workflow will depend on how the wider rollout performs for users, not only on the capabilities Alibaba has described.
Editorial analysis
Wan3.0’s more consequential feature is not simply a longer generated clip. It is the attempt to make existing business material—slides, spreadsheets, web pages, and documents—the starting point for video production. That could lower the work needed to turn routine marketing and informational assets into visual campaigns. The next evidence to watch is whether wider access produces sustained use beyond Alibaba’s stated early applications in advertising, tourism, film, and music-video work—and how users judge the company’s claims of visual consistency.
Loading discussion...
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.