Meta Superintelligence Labs delivers first internal AI models, CTO Andrew Bosworth reveals

Meta's New AI Lab Delivers First Internal Models, CTO Reveals | AIM

Meta's AI unit has produced its first internal models as Mark Zuckerberg bids to compete at the frontier. Meta Platforms' newly formed artificial intelligence lab has delivered its first high-profile AI models internally, the company's chief technology officer has said, according to Reuters. Speaking at a press briefing on the sidelines of the World Economic Forum's annual meeting in Davos, Switzerland, CTO Andrew Bosworth said the models developed by Meta Superintelligence Labs -- established last year -- showed significant promise. "They're basically six months into the work, not quite even," Bosworth said, adding that the team's AI models were "very good". According to WSJ, in December, Meta was developing a text-based AI model codenamed 'Avocado', expected to launch in the first quarter, alongside an image- and video-focused model codenamed 'Mango'. Bosworth did not confirm which of these models had been delivered internally. Meta's progress is being closely watched after chief executive Mark Zuckerberg moved to overhaul the company's AI leadership, set up a new lab and aggressively recruit top talent with lucrative compensation packages, as he seeks to position Meta at the forefront of AI development. Bosworth cautioned that building usable AI systems involves far more than training models alone. "There's a tremendous amount of work to do post-training to actually deliver the model in a way that's usable internally and by consumers," he said. He described 2025 as a "tremendously chaotic year" for Meta, marked by rapid infrastructure build-out, expanded computing capacity and efforts to secure sufficient power to support its AI ambitions. However, he added that the company is beginning to see returns on those investments. Looking ahead, Bosworth said 2026 and 2027 would be pivotal years for consumer AI. Recent advances, he noted, have already produced systems capable of handling everyday questions from users and families, with the next challenge being to translate those capabilities into mass-market products. Meta has already begun rolling out AI-powered consumer hardware, including smart glasses developed with Ray-Ban. The company recently paused international expansion of the device, however, to focus on meeting strong demand in the US.

Meta's AI Lab Delivers First Models In-House | PYMNTS.com

Andrew Bosworth, the company's chief technology officer, told Reuters Wednesday (Jan. 21) that the models created by Meta's Superintelligence Labs team showed promise. "They're basically six months into the work, not quite even," said Bosworth, who spoke with Reuters at the World Economic Forum in Davos, Switzerland. He added that the team's AI models were "very good." The news follows reports from last month that Meta was developing a text AI model codenamed Avocado scheduled for a first-quarter debut, as well as a n image and video-focused model codenamed Mango. Bosworth did not specify which models were delivered, Reuters added. The report notes that Meta's AI efforts are being watched closely in the wake of CEO Mark Zuckerberg's moves in the last year, which saw the company overhaul its AI leadership, start a new lab and recruit talent with multi-million dollar compensation offers. Meta, Reuters added, has been criticized for the performance of its Llama 4 model amid launches by rivals such as Google. Bosworth told the news outlet that the technology was still being worked on. "There's a tremendous amount of work to do post-training" for AI, "to actually deliver the model in a way that's usable internally and by consumers," he said. Bosworth's comments came a little more than a week after reports that Meta was planning to cut 10% of the jobs in its 15,000-person Reality Labs unit, part of a larger plan to shift its focus on virtual reality products as it concentrates on other AI wearables. Writing about Meta's AI projects last year, PYMNTS contrasted the company's efforts with those of rivals Amazon, Google and Microsoft. Where those companies pair their AI investments with clear paths to revenue, Meta has concentrated on building for internal use. The report cited the example of Microsoft monetizing OpenAI's models via subscriptions to Azure and Copilot, and Google selling Gemini and Vertex AI access through its cloud division, while also building a burgeoning TPU business. And Amazon's Bedrock and SageMaker platforms transform infrastructure into recurring enterprise revenue. "Meta, by contrast, uses its AI for engagement, recommendation engines, ad ranking, and tools like Meta AI and Reels," PYMNTS wrote. "Those may improve user metrics, but it's not clear how they will contribute to the bottom line."

Meta's new AI team delivered first key models internally this month, CTO says

DAVOS, Switzerland, Jan 21 (Reuters) - Meta Platforms' new artificial intelligence lab has delivered its first high-profile AI models internally this month, the company's chief technology officer said on Wednesday. At a press briefing on the sidelines of the World Economic Forum's annual meeting in Davos, CTO Andrew Bosworth said the models built by its Meta Superintelligence Labs team, formed last year, showed a lot of promise. "They're basically six months into the work, not quite even," Bosworth said, adding that the team's AI models were "very good". Media outlets reported in December that Meta was working on a text AI model codenamed Avocado slated for a first-quarter launch, along with an image and video-focused model codenamed Mango. Bosworth, known as "Boz," did not specify which models were delivered internally. Meta's efforts are being closely followed after major moves by CEO Mark Zuckerberg to shake up its AI leadership, form a new lab and poach talent with sky-high offers, hoping the company can win in the highly competitive technology frontier. (Reporting by Jeffrey Dastin in Davos, Switzerland. Editing by Paritosh Bansal and Mark Potter)

Meta's new Superintelligence Labs team delivers its first AI models internally, CTO reveals

Bosworth said the models built by the Meta Superintelligence Labs team showed a lot of promise. For months now, Meta has been on a hiring spree for its new Superintelligence Labs, bringing in top AI talent with high compensation packages to compete in the AI race. Now, it seems like the company's efforts are starting to show results. Meta's Chief Technology Officer, Andrew Bosworth, has revealed that the lab has delivered its first high-profile AI models internally. Speaking at a press briefing during the World Economic Forum in Davos, Bosworth said the models built by the Meta Superintelligence Labs team showed a lot of promise, reports Reuters. "They're basically six months into the work, not quite even," he said, describing the models as "very good." Last month, it was reported that Meta is developing a text-based AI model codenamed Avocado, expected to launch in the first quarter of 2026, and a model focused on images and video called Mango. Bosworth did not confirm which specific models had been delivered internally. Also read: Google DeepMind CEO responds to OpenAI testing ads in ChatGPT: Here's what he said Bosworth emphasised that the technology is still in development. "There's a tremendous amount of work to do post-training" for AI "to actually deliver the model in a way that's usable internally and by consumers," he said. Also read: Samsung accidentally reveals smarter Bixby powered by Perplexity AI: Check details Reflecting on 2025, Bosworth called it a "tremendously chaotic year" for building the lab, setting up infrastructure, and securing computing power. Despite the challenges, he said Meta is already seeing positive returns from its investments. Looking ahead, he said 2026 and 2027 will be critical years for consumer AI. He added that recent advances have created models that answer "the kinds of things that you ask every day with your family, your kids," and further improvements are expected to handle more complex questions.