Meta’s AI agent development has not advanced as rapidly as anticipated, CEO Mark Zuckerberg revealed, highlighting a slower pace despite the company’s aggressive investments and restructuring. This slower progress challenges expectations held by technology and crypto firms banking on autonomous AI agents to transform digital interactions and blockchain payments.
During a recent company meeting, Zuckerberg noted that the development trajectory over the past four months did not accelerate as predicted. Meta’s strategic shift began early this year with a strong focus on agentic AI infrastructure, driven by concerns the company was lagging in the space. Zuckerberg confirmed that, while adoption of AI agents has yet to reach projected levels, Meta expects to see returns on its AI investments within the next three to six months.
The company reorganized its workforce in May, cutting about 10% of jobs and reallocating roughly 7,000 employees to AI-centric teams. Zuckerberg admitted this transition was imperfect, as leadership misjudged the timing of the push for AI capabilities. Despite the challenges, Meta expanded its AI Business Agent globally across Instagram, Messenger, and WhatsApp platforms. This AI feature enables businesses to automate customer interactions, including answering inquiries, recommending products, and closing sales without human involvement.
Zuckerberg also disclosed plans to create a personal AI assistant aimed at enhancing his leadership decisions. Meanwhile, the crypto sector remains highly invested in AI agent technology. CEOs of major crypto companies have predicted AI agents will soon dominate blockchain payment usage, leading to several experimental integrations involving AI-driven stablecoin transactions.
Notably, Amazon Web Services integrated Coinbase’s x402 payment protocol into its AI services, allowing agents to execute transactions in USD Coin (USDC). Crypto startup Oobit introduced a Visa-supported virtual card for AI agents to complete online purchases in USD Tether (USDT) on behalf of businesses. However, real-world adoption is limited, with transaction volumes through AI agent-enabled blockchain protocols remaining in the low millions over recent months.