OpenAI expects to generate $2.5 billion in ad revenue this year and $100 billion by 2030, according to a source familiar with recent presentations to investors.

Why it matters: It's the clearest sign yet that OpenAI sees advertising as central to its future business.

Driving the news: OpenAI's ad pilot generated $100 million in annual recurring revenue in under two months and the company says it's projecting steep growth.

• The company has told investors to expect the $2.5 billion in 2026 ad revenue to grow to $11 billion in 2027, $25 billion in 2028 and $53 billion by 2029, per the source.
• The projections assume OpenAI's products reach 2.75 billion weekly users by 2030 and capture a share of the global ad market dominated by Google, Meta, Amazon and TikTok.

Context: Chatbot ads could be unusually lucrative because users volunteer exactly what they want.

• Meta and Google have built an effective ad business by both knowing and inferring what their users like, care about and want to buy.
• With chatbots, users often directly state their goals.

The big picture: OpenAI is trying to convince investors and potential backers of a future IPO — that it has multiple revenue streams that scale alongside the company's ever-increasing spending on compute.

Yes, but: The introduction of advertising, while a potentially lucrative way to offset its massive costs, risks upending one of the key selling points of AI chatbots — that they work for the users, not for the advertiser.

Advertising has been the primary revenue stream for the consumer internet, funding services such as email and social networking without directly charging users.

• Critics say that has created a system where tech companies are incentivized to keep people spending as much time as possible on their services.

Between the lines: OpenAI and rival Anthropic have diverged over the issue, with Anthropic using a Super Bowl commercial to declare that Claude will remain ad-free.

• OpenAI, for its part, has said that advertising will allow its products to reach more people and has laid out a series of principles designed to clearly separate any advertising and lay out the kind of data that will be shared with marketers.

Meta yesterday debuted Muse Spark, a homegrown AI model it says significantly narrows the performance gap with models from OpenAI and Anthropic.

Why it matters: The model — code-named Avocado and built over the past nine months by a team led by Alexandr Wang — is a major upgrade over its Llama 4 models, Meta tells Axios.

Driving the news: Muse Spark will power queries in the Meta AI app and Meta.ai website immediately, with plans to expand across Facebook, Instagram and WhatsApp.

• The model accepts voice, text and image inputs, but produces text-only output.
• As Axios first reported, Meta plans to release a version of Muse Spark under an open-source license.

Zoom in: The model uses a fast mode for casual queries and several reasoning modes.

• A "shopping mode" highlights how Meta hopes to differentiate itself. It combines large language models with data on user interests and behavior.
• Over time, the model will also power "features that cite recommendations and content people share across Instagram, Facebook, and Threads," Meta said in a blog post.

Zoom out: A Meta executive told Axios that Muse Spark doesn't mark a new state of the art, but is competitive with the latest models from leading labs at certain tasks, including multimodal understanding and processing health information.

• In other areas, including coding, the company acknowledges that there is a gap between Muse Spark and the models already available.
• All flavors of the model are free to use, though Meta may impose rate limits.

Yes, but: Consumers should be aware that Meta's privacy policy sets few limits on how the company can use any data shared with its AI system.

The big picture: Meta sees the release of Muse Spark as just one step toward its broader vision of personal superintelligence.

OpenAI is finalizing a product with advanced cybersecurity capabilities that it plans to release to a small set of partners, a source familiar told Axios.

Why it matters: AI capabilities have reached a tipping point, at least in terms of autonomy and hacking capabilities. Model-makers are now so worried about the havoc their own tools could cause that they're reluctant to release them into the wild.

• Anthropic is also planning a limited rollout of Mythos, its new model.

Driving the news: Anthropic announced plans Tuesday to limit access of its new Mythos Preview model to a handpicked group of technology and cybersecurity companies over fears of its advanced hacking capabilities.

• At the time, it was the first AI company to take such an approach with a new model.

Zoom in: OpenAI introduced its "Trusted Access for Cyber" pilot program in February after rolling out GPT-5.3-Codex, the company's most cyber-capable reasoning model.

• Organizations in the invite-only program are given access to "even more cyber capable or permissive models to accelerate legitimate defensive work," according to a blog post.
• At the time, OpenAI committed $10 million in API credits to participants.

The big picture: Former government officials and top security leaders have been ringing alarm bells over the past year about AI models that — in the wrong hands — could autonomously disrupt water utilities, the electric grid, or financial systems.

• Those capabilities now appear to be here.

Threat level: Even if AI companies hold back their models for limited releases, top security experts all have the same message: There's no going back.

• "You can't stop models from doing code enumeration or finding flaws in older codebases," said Rob T. Lee, chief AI officer at the SANS Institute. "That capability exists now."
• It's only a matter of weeks or months before there's a new model with similar capabilities out in the wild, Wendi Whitmore, chief security intelligence officer at Palo Alto Networks, told Axios during a panel at the HumanX conference in San Francisco on Tuesday.

• Anthropic lost its bid to pause Pentagon blacklisting in a D.C. appeals court. (Axios)
• Meta reportedly shut down an internal leaderboard — called Claudeonomics — that tracked how many tokens employees were using. (The Information)
• Visa is rolling out a new platform where AI agents do the shopping. (Axios)
• Brands are learning how to reach consumers using AI tools in place of search engines. (Axios)

Spot & Tango is hiring a professional dog breath sniffer to test out pup dental products. Maybe this is the AI-proof job we should all be chasing.