Meta launches Muse Image as it chases 3 revenue streams from AI

Meta released Muse Image on July 7, its first image generation model from Meta Superintelligence Labs, embedding AI creation tools across Instagram, WhatsApp, and its $155 billion advertising business — while letting users pull other people's public Instagram photos into generated images.

"With Muse Image, we're bringing native reasoning to the creative process so the model can adjust elements, swap styles, and create variations based on what the user provides," a Meta spokesperson said. "This is the first image model built entirely by our Superintelligence Labs team."

The model, internally code-named Mango, replaces third-party providers Midjourney and Black Forest Labs that Meta previously used for image generation in its Meta AI app. Muse Image is available for free in the Meta AI app, Instagram Stories, and WhatsApp direct messages, with Facebook and Messenger access coming later this year. Power users who exceed free limits must subscribe to Meta One, a monthly plan the company began testing in May at $7.99 per month.

The launch marks the second major release from Meta Superintelligence Labs, led by Alexandr Wang, who Meta hired last year to head its AI efforts. Wang previously oversaw the April debut of Muse Spark, the large language model that replaced Meta's Llama family. Meta said Muse Image works with Muse Spark "to reason through your prompt, search the web, and plan before it generates" — a capability the company calls agentic image generation.

The @mention feature is the most controversial element. Users can tag Instagram accounts in their prompts, allowing Muse Image to incorporate that person's likeness from public photos. Meta said "tagging a username lets Meta AI use public photos to build a visual," and users can control how their content is reused through Instagram's content reuse settings. The feature has drawn criticism from privacy advocates and users who say Meta opted them in by default.

Meta also plans to release Muse Video, an AI video generation model, at a later date. Wang said on Threads that it is "competitive on prompt adherence, visual fidelity, temporal consistency" — though the company has not disclosed benchmark comparisons against OpenAI's Sora or Google's Veo.

The advertising angle is where the revenue opportunity lies. Muse Image will power image-generation tools within Meta's Advantage Plus service, which lets brands automate ad creative development. Meta said it has been working with businesses and advertisers ahead of the rollout. "In the coming weeks, advertisers and agencies can expect to see image variants powered by Muse Image," the company said in a business-facing blog post.

Internal benchmark tests show Muse Image trailing OpenAI's latest GPT Image 2 model but beating Google's Nano Banana 2 model in tasks including single and multiple image editing. Meta did not disclose the test methodology or conditions for these comparisons.

For investors, the question is whether AI features can move Meta's revenue needle. Meta spent an estimated $35 billion to $40 billion on AI-related capital expenditures in 2025, according to company filings, and has signaled similar spending in 2026. The company's core advertising business generated $155 billion in revenue last year, and AI-powered creative tools represent a potential margin improvement rather than a new revenue line. Meta shares trade at roughly 23 times forward earnings. The subscription tier — Meta One at $7.99 per month — is still in testing and has not disclosed subscriber numbers.

Privacy concerns could complicate adoption. Public Citizen called the @mention feature "an egregious invasion of user privacy," and the BBC reported an "outcry" from users who discovered their public Instagram photos could be used in others' AI generations without explicit consent. Meta has not said how many users have opted out since the feature launched.

This article is for informational purposes only and does not constitute investment advice.