Meta Platforms is weighing sweeping layoffs that could affect 20% or more of its workforce as the social media giant attempts to manage the enormous costs of building artificial intelligence infrastructure.

The move is believed to be a part of the conglomerate’s preparation for a future in which AI systems perform many tasks once handled by large teams of employees.

Three people familiar with the discussions told Reuters that senior executives have begun signaling the potential cuts to other leaders inside the company and have asked them to examine ways to streamline their organizations. No final decision has been made on the timing of the layoffs or their precise magnitude, the sources said.

If implemented at the level being discussed, the move would represent Meta’s largest workforce reduction since the restructuring campaign in 2022 and 2023 that chief executive Mark Zuckerberg labeled the company’s “year of efficiency.”

Meta employed nearly 79,000 people as of December 31, according to its latest regulatory filing. A 20% reduction would therefore translate to more than 15,000 jobs — a scale that would mark one of the biggest technology-sector layoffs in recent years.

A company spokesperson, Andy Stone, responded cautiously when asked about the reported plans, saying: “This is speculative reporting about theoretical approaches.”

AI spending reshaping Meta’s priorities

The internal discussions point to the growing financial strain created by Meta’s massive push into artificial intelligence. The company has outlined plans to spend as much as $600 billion building new data centers by 2028 in order to support the computing demands of increasingly powerful AI models.

These facilities are designed to house vast clusters of specialized processors capable of training and running large-scale machine-learning systems that power chatbots, recommendation engines, and autonomous digital agents. Such investments are becoming a defining feature of the global technology race as companies compete to develop advanced AI capable of performing complex reasoning tasks and generating content across text, images, video, and audio.

However, the spending surge has dramatically altered Meta’s cost structure. Alongside infrastructure investments, the company has been offering unusually large compensation packages to recruit elite artificial intelligence researchers. Some offers reportedly reach hundreds of millions of dollars over several years, reflecting the fierce competition for talent in the AI sector.

Efficiency Gains From AI

The potential layoffs also align with Zuckerberg’s view that artificial intelligence will enable companies to operate with significantly smaller teams. He indicated earlier this year that the company was already seeing productivity improvements from AI-assisted development tools.

“I’m starting to see projects that used to require big teams now be accomplished by a single very talented person,” Zuckerberg said in January.

Such comments suggest that Meta expects automation tools — including AI capable of writing software code, generating marketing material, and performing data analysis — to reduce the need for large organizational structures over time. The shift mirrors a broader transformation underway across the technology industry, where companies increasingly view AI not only as a product opportunity but also as a tool for internal efficiency.

A broader wave of AI-driven job cuts

Meta’s plans are part of a pattern emerging across major U.S. companies as executives reassess staffing needs in light of rapid improvements in artificial intelligence systems.

Earlier this year, Amazon confirmed plans to eliminate roughly 16,000 jobs, representing close to 10% of its workforce. Fintech company Block also cut nearly half of its staff, with chief executive Jack Dorsey pointing explicitly to the growing capabilities of AI tools that allow businesses to “do more with smaller teams.”

Across Silicon Valley, the narrative is increasingly consistent: artificial intelligence is expected to boost productivity to the point where fewer employees can accomplish the same amount of work.

Meta’s aggressive spending on artificial intelligence comes after a series of challenges that raised questions about its progress in the AI race. Last year, the company’s Llama 4 models drew criticism after researchers said early benchmark results had been presented in a way that overstated their capabilities. Meta later scrapped the planned release of the largest version of the model, internally known as “Behemoth,” which had been scheduled for launch in the summer.

The company’s newly formed “superintelligence” team is now working to regain momentum with a new model code-named “Avocado.” However, people familiar with the project have said the model’s early performance has lagged expectations compared with competing systems.

Meta has also turned to acquisitions to accelerate its artificial intelligence strategy. The company recently bought Moltbook, a social networking platform designed specifically for AI agents, reflecting a broader vision in which autonomous digital assistants interact with users across Meta’s services.

In addition, Meta is spending at least $2 billion to acquire Chinese artificial intelligence startup Manus, according to earlier Reuters reporting. These deals highlight the company’s attempt to build a full AI ecosystem that extends beyond traditional social media into areas such as automated content generation, intelligent assistants, and digital commerce.

Balancing Investment With Investor Expectations

While Meta’s long-term bet on artificial intelligence has been welcomed by some investors, the scale of spending has also raised concerns about profitability. Data centers, advanced chips, and AI research require enormous capital investment, and the financial returns from such spending may take years to materialize.

Cutting operating costs through layoffs could therefore become a key part of Meta’s strategy to reassure investors that the company can fund its AI ambitions without undermining margins.

However, the potential workforce reduction highlights how dramatically Meta’s priorities have shifted over the past decade. The company once expanded rapidly to support the growth of its core platforms — Facebook, Instagram, and WhatsApp — as well as ambitious projects such as virtual reality and the metaverse.

Now the focus has turned to building the infrastructure and research capacity needed to compete in the global race for artificial intelligence leadership.

If Meta ultimately proceeds with layoffs on the scale being discussed internally, it would mean not only a major restructuring for the company but also a broader turning point for the technology industry. It would broaden a trajectory in which AI is beginning to reshape not just the products companies build, but also the size and structure of the organizations behind them.

Google has invested $1 million in an animation startup developing artificial intelligence-powered children’s content, a move that underscores growing concern inside YouTube over the explosion of low-quality AI-generated videos on the platform.

The funding, provided through Google’s AI Futures Fund, will support Animaj, a digital animation studio that focuses on children’s programming. In addition to the capital injection, Animaj will receive early access to several of Google’s latest generative AI tools, including the video-generation system Veo, the multimodal AI model Gemini, and the image-generation technology Imagen.

The investment is another attempt by Google to shape how generative AI is used on YouTube at a time when the technology is rapidly transforming online video creation.

Over the past year, generative AI tools have dramatically lowered the barrier to producing video content. With automated scriptwriting, synthetic voice narration, and AI-generated visuals, creators can produce large volumes of videos in minutes. While the technology has enabled new forms of creativity, it has also triggered a surge in what many creators call “AI slop”—mass-produced, low-effort videos designed primarily to attract views rather than deliver meaningful content.

The trend has become particularly visible on YouTube’s short-form video platform, where automated channels can publish dozens or even hundreds of clips per day.

Neal Mohan, chief executive of YouTube, has acknowledged the challenge, saying the platform has a “responsibility to maintain the high-quality viewing experience that people want.”

Mohan’s comments suggest that YouTube is trying to strike a balance between embracing generative AI tools and preventing a flood of automated content from degrading the platform’s ecosystem.

Animaj’s founders say the company intends to show that artificial intelligence can be used to produce polished, story-driven children’s programming rather than the chaotic and often nonsensical clips that have circulated widely online.

Co-founder Sixte de Vauplane said Google’s investment reflects the platform’s recognition that AI-generated content needs better standards.

“Google knows the problem and the issue of AI slop that is happening right now on YouTube,” Vauplane said. “They know that right now, you don’t have a lot of people and a lot of players in the kids media industry that have really proven their ability to use AI in a very good way.”

Animaj already runs several children’s entertainment channels that collectively generated about 22 billion views last year, a scale that illustrates both the massive demand for kids’ content and the influence such channels can have on YouTube’s viewing ecosystem.

For Google, backing studios capable of producing higher-quality AI-driven programming could help set benchmarks for responsible AI content creation while maintaining the efficiency benefits of automation.

Children’s Programming Becomes A Critical Battleground

Children’s entertainment has long been one of the most dominant categories on YouTube, with animated series, nursery rhymes, and educational videos attracting billions of views globally. That popularity has also made the segment particularly vulnerable to algorithmic manipulation and low-quality automated uploads designed to capture advertising revenue.

Investigations have repeatedly uncovered strange or disturbing AI-generated videos targeting children—sometimes featuring distorted characters, incoherent storylines, or misleading educational material. Following reporting by The New York Times on unusual AI-generated children’s videos circulating widely on the platform, YouTube removed a number of channels that had amassed billions of views.

The crackdown highlighted the platform’s growing concern that uncontrolled AI content could undermine trust among parents, educators, and advertisers. A YouTube spokesperson, Boot Bullwinkle, said creators are required to disclose when artificial intelligence is used to generate realistic or potentially misleading content.

“We require creators to disclose when they’ve used A.I. to create realistic content, meaning things a viewer could easily mistake for a real person, place, or event,” Bullwinkle said.

The investment is also seen as part of Google’s broader push to integrate generative AI across its ecosystem, including search, advertising, creative tools, and digital media platforms. Tools like Veo, Gemini, and Imagen are designed to automate key parts of the creative process—from generating images and animations to producing scripts and voiceovers.

For content creators, the technology could significantly reduce production costs while accelerating the pace of video creation. But for platforms such as YouTube, the same efficiency also raises the risk of overwhelming audiences with low-quality material.

By supporting studios like Animaj, Google appears to be experimenting with a hybrid model—pairing human storytelling and editorial oversight with AI-driven production tools.

Shaping The Future Of AI-Powered Entertainment

The battle against “AI slop” is emerging as one of the defining challenges for platforms built around user-generated content. As generative AI becomes more powerful and widely accessible, the volume of automated content online is expected to rise sharply.

Technology companies are therefore under increasing pressure to develop new rules, moderation tools, and creator partnerships that ensure AI enhances creativity rather than eroding quality.

For YouTube, where billions of people consume video daily, the stakes are particularly high.

Google’s investment in Animaj suggests the company is not only trying to control the spread of low-quality AI videos but also to cultivate a new generation of creators capable of using artificial intelligence responsibly—especially in sensitive categories such as children’s entertainment, where trust, safety, and quality remain big issues.

Tesla Chief Executive Elon Musk said the company’s long-discussed “Terafab” project aimed at producing artificial intelligence chips will launch within seven days, signaling a major step in the electric vehicle maker’s effort to secure the computing power needed for its autonomous driving ambitions.

Musk disclosed the timeline on Saturday, offering one of the clearest indications yet that Tesla is moving closer to vertically integrating a critical part of its artificial intelligence infrastructure — the semiconductor hardware that powers its self-driving systems.

The Terafab initiative is tied to Tesla’s development of its fifth-generation AI chip, commonly referred to as AI5, which is designed to support the company’s next wave of autonomous vehicle technology and large-scale AI training systems.

The chips power Tesla’s driver-assistance and autonomy software, including its Full Self-Driving system, which relies heavily on advanced neural networks trained on massive volumes of real-world driving data collected from Tesla vehicles.

Musk has repeatedly warned that the global supply of advanced chips is insufficient to meet Tesla’s rapidly expanding demand for artificial intelligence computing.

“Even when we extrapolate the best-case scenario for chip production from our suppliers, it’s still not enough,” Musk said during Tesla’s annual shareholder meeting last year.

That shortage of high-performance processors has pushed Tesla to explore building its own massive semiconductor fabrication facility.

“So I think we may have to do a Tesla terafab. It’s like giga but way bigger. I can’t see any other way to get to the volume of chips that we’re looking for,” Musk said at the time.

“I think we’re probably going to have to build a gigantic chip fab. It’s got to be done.”

The strategy mirrors Tesla’s broader approach of controlling critical parts of its supply chain — a philosophy the company previously applied to battery manufacturing through its “Gigafactory” network.

Potential Partnerships With Chipmakers

Although Tesla is designing its own AI processors, Musk has suggested that the company could collaborate with existing semiconductor manufacturers. He said last year that Tesla might work with Intel on the manufacturing side, although no formal agreement has been reached.

“We haven’t signed any deal, but it’s probably worth having discussions with Intel,” Musk said at the time.

Tesla is also partnering with two of the world’s most advanced chip foundries — Taiwan Semiconductor Manufacturing Company and Samsung Electronics — to produce versions of its AI processors. These chips are used in Tesla’s data centers as well as inside its vehicles, enabling real-time processing of camera feeds, sensor inputs, and complex machine-learning models that guide driving decisions.

The Importance of Autonomous Vehicles

Tesla’s push to expand chip production highlights how computing power has become the central bottleneck in the race to develop fully autonomous vehicles. Modern self-driving systems rely on enormous neural networks that must be trained using billions of miles of driving data.

That process requires vast computing clusters running specialized chips optimized for artificial intelligence workloads.

Tesla already operates a large AI training system known as Dojo, designed to accelerate development of its autonomy software. The AI5 processor is expected to deliver significantly greater performance than the company’s current generation chips, enabling more advanced perception models and decision-making algorithms.

Industry analysts say Tesla’s Terafab plan is a typical example of the broader shift among major technology companies toward designing and controlling their own AI hardware. Companies building large artificial intelligence systems increasingly require specialized chips tailored to their specific software architectures and data processing needs.

Relying solely on external suppliers can limit both performance and scale, particularly as demand for AI processors surges globally.

So, Tesla could gain greater control over costs, supply, and performance — factors that are becoming decisive in the AI arms race, by designing its own chips and potentially building dedicated fabrication capacity.

However, the EV maker’s investment in AI hardware is closely tied to Musk’s long-standing view that the company is fundamentally an artificial intelligence and robotics company rather than just an automaker. Autonomous driving remains a central pillar of that vision, with Musk arguing that large fleets of self-driving vehicles could eventually form a global robotaxi network.

To achieve that goal, Tesla must train increasingly sophisticated neural networks capable of handling the complexity of real-world driving environments.

The launch of the Terafab initiative, therefore, represents more than a manufacturing project. It marks another step in Tesla’s effort to build the vast computing infrastructure required to turn autonomous driving from a technological ambition into a scalable business.