Meta CEO Mark Zuckerberg admitted that the company's AI agent development has progressed slower than planned over the past four months, despite a major reorganization and investment of up to $145 billion（約23兆円） in infrastructure this year. His AI chief offered a more optimistic view, claiming an upcoming model called "Watermelon" matches OpenAI's GPT-5.5 and that improvements to coding and agent capabilities are imminent.
Summaries like this, in your inbox every morning.
Sign up free →What happened
Mark Zuckerberg acknowledged at an internal town hall that AI agents—which Meta restructured around—have not advanced as fast as expected. The trajectory over at least the last four months hasn't accelerated the way the company anticipated, and the bets on the new structure haven't come to fruition yet.
Why it matters
Meta bet heavily on AI this year, moving roughly 7,000 employees into AI teams after a May layoff and planning to spend up to $145 billion（約23兆円） on AI infrastructure. Zuckerberg's admission signals that the company's reorganization and major talent investments have yet to deliver the efficiency gains and model breakthroughs it sought, even as competitors like OpenAI and Anthropic advance.
What to watch
Zuckerberg expects tangible results within the next three to six months. Meanwhile, AI chief Alexandr Wang countered that Meta's upcoming model code-named "Watermelon" has caught up with OpenAI's top model GPT-5.5, and that a Muse Spark update with major coding and agentic improvements is coming soon.
No comments yet. Be the first to share your thoughts!
Log in to join the discussion
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
1 minute a day. The AI essentials.
200+ sources · Email / LINE / Slack
