The Morning Download: Tubi Builds AI-Driven Media Model

Good morning, CIOs. Tubi, the ad-supported streaming service, is putting AI at the center of its strategy to take on bigger players, the WSJ Leadership Institute's Belle Lin reports.

The free streamer this week announced that its app is integrated into OpenAI’s ChatGPT, letting viewers search for what to watch inside the chatbot.

“Our bet is that the future of television should be as easy and personalized as Instagram,” said Tubi Chief Executive Anjali Sud. That turns on the use of AI to keep its Gen Z-heavy audience hooked.

That age group, which is largely part of Gen Z, is where the company thinks AI can help. The streamer owned by Fox Corp. declined to share the dollar amount of its investment in AI. Fox and The Wall Street Journal’s parent company, News Corp, share common ownership.

Tubi combined large language models with deep learning techniques—enabling it to better understand “the nature of the user’s intent,” Mike Bidgoli, Tubi’s chief product and technology officer, says. An LLM can predict a user’s viewing patterns the way it predicts text from a few letters or words. The company uses a variety of commercial LLMs and its own custom model.

• Tubi can create a level of “hyper-personalization”—suggesting more personalized shows and movies, as well as ads, based on data like a user’s location and time of viewing, as well as watch history and even what’s in the content.
• The company automatically plays AI-selected scenes from movies, individualized to viewers as they scroll, and it can create a trail of digital breadcrumbs for them to follow as they delve deeper into subgenres.
• One of the company’s ad products, called Tubi Moments, uses AI to tag scenes within movies and shows, reflecting the scene’s tone, sentiment and visual cues. A scene featuring a drink gets labeled as a category a beverage maker can buy—allowing the beverage maker to display an ad during the movie’s ad breaks.

Tubi, which has about 700 employees, generated more than $1.1 billion in fiscal 2025 revenue. Fox said Tubi turned profitable in the first quarter of fiscal 2026, and Tubi finished fiscal 2025 with about a 2.2% share of all television viewing, according to Nielsen

On Tuesday, AI giant Anthropic unveiled a partnership with cybersecurity companies that raised big questions about how parts of the security industry may be disrupted by artificial intelligence, WSJ Pro’s James Rundle reports.

The company said its new Project Glasswing initiative allows select companies access to its Claude Mythos2 Preview frontier model, specifically for defensive cybersecurity work. Participants include CrowdStrike, Palo Alto Networks, Microsoft, Apple, Amazon’s AWS cloud business, JPMorgan Chase, Google, Broadcom, Nvidia and the Linux Foundation.

Anthropic said its new model already has found thousands of high-severity vulnerabilities, including some in every major operating system and web browser.

For technology leaders, many of whom say cybersecurity is the single biggest issue that keeps them up at night, it’s a sign that traditional vendors—even those hit by the recent SaaS selloff—still have a role to play in securing enterprise data.

Cyber shares rose as some investors were encouraged by the companies’ inclusion in the Anthropic project.

Meta Platforms announced a new large language model Wednesday, its first major new AI model in more than a year, WSJ’s Meghan Bobrowsky reports.

The rollout of the model, called Muse Spark, is a critical moment for Meta, which has spent billions of dollars hiring AI talent in a bid to catch up to OpenAI, Anthropic and Google DeepMind. The leading labs have been putting out models at an accelerating pace.

In a departure from its previous models, which were open-source, Muse Spark is a closed model that will power Meta’s AI chatbot and AI features within it. The company said it planned to release a private preview of the model to a few partners via an application programming interface, or API, which allows developers to build on top of existing software, and at some later point might open-source some versions of the model.

Open-sourcing a model gives the public access to some parts of the code and other architecture behind it, and some tech leaders say open-source models are critical to effectively building some AI apps inside the enterprise.

According to Meta’s internal benchmark tests, Muse Spark outscored Google’s Gemini on some tests and was competitive with models from OpenAI and Anthropic on others.

Locals are, ironically, using AI to fight the development of power hungry data centers in their backyards. Community activists and residents in rural Ohio are tapping AI tools to push back against data-center builds, citing concerns about electricity prices and environmental damage.

But the buildout of data centers seems as bullish as ever.

Applied Digital, the Dallas-based AI data-center developer, reported fiscal 2026 third-quarter results that topped Wall Street estimates after the closing bell on Wednesday, as the company said it sees accelerating demand for AI data-center capacity.

And specialist investor NOVA Infrastructure, just raised $1.45 billion to back assets like data centers, battery projects and water-management systems—looking to benefit both from rising demand for infrastructure tied to artificial intelligence and the industry’s relative immunity to AI-driven disruptions.

What about software?

Neo, the venture firm, led by CEO Ali Partovi, is showing returns that leave the rest of the crowd in the dust, illustrating the growing gap between high performers and everyone else in today’s venture market. The San Francisco-based firm turned a $150 million fund into $1.2 billion, after it made early bets on coding tools company Cursor, prediction-markets operator Kalshi, and other startups whose valuations have been on a tear lately.

CoreWeave expanded its long-term agreement with Meta Platforms, providing the company with artificial-intelligence cloud capacity through December 2032 for about $21 billion.

The companies said Thursday the deal expands their existing relationship, increasing support for Meta’s development and deployment of AI, the WSJ reports.

The dedicated capacity will be deployed across multiple locations and include some of the initial deployments of Nvidia’s Vera Rubin platform, they said. The companies previously entered an AI cloud infrastructure deal last fall, worth up to $14.2 billion.

A federal appeals court denied Anthropic’s request for relief from the Defense Department’s supply-chain risk designation. (WSJ)