Meta Platforms' AI Lab Delivers Models
Analysis based on 6 articles · First reported Jan 21, 2026 · Last updated Jan 22, 2026
The internal delivery of Meta Platforms' first high-profile AI models signals positive progress in its AI strategy, potentially boosting investor confidence and its stock price. This development positions Meta Platforms more competitively against rivals like Alphabet Inc. and Microsoft in the lucrative AI market.
Meta Platforms' new artificial intelligence lab, Meta Superintelligence Labs, has delivered its first high-profile AI models internally. Andrew Bosworth, CTO of Meta Platforms, announced this development at the World Economic Forum in Davos, highlighting the models' promise despite being only six months into development. This progress follows significant strategic changes led by CEO Mark Zuckerberg, including an overhaul of Meta Platforms' AI leadership and aggressive talent recruitment. While specific models were not confirmed, earlier reports suggested a text AI model codenamed 'Avocado' and an image/video model codenamed 'Mango' were in development for a first-quarter launch. Meta Platforms has faced criticism regarding its Llama 4 model's performance, especially compared to rivals like Alphabet Inc.'s Alphabet Inc., Amazon (company), and Microsoft, which have established clear revenue paths for their AI investments. Bosworth emphasized that significant post-training work is still required to make the AI usable for consumers, but noted that 2025's investments are starting to show returns, with 2026 and 2027 expected to be pivotal years for consumer AI.
Set up alerts, explore entity relationships, search across thousands of events, and build custom intelligence feeds.
Open Dashboard