Stock Markets June 4, 2026 11:30 AM

Meta Developing 'Agentic' AI Assistant Built on Muse Spark, Internal Tests Underway

Company reportedly trials advanced digital assistant and plans Instagram shopping integration as it increases AI infrastructure spending

By Derek Hwang META

Meta is reported to be building a highly personalized, agentic AI assistant using its Muse Spark model and is testing internal prototypes. The initiative includes work inspired by the OpenClaw open-source project and a separate internal agent codenamed "Hatch," with a planned Instagram shopping integration before the fourth quarter. The moves come as Meta raises capital expenditure plans to expand AI infrastructure amid concerns over user engagement among younger demographics.

Meta Developing 'Agentic' AI Assistant Built on Muse Spark, Internal Tests Underway
META

Key Points

  • Meta is developing an agentic, personalized AI assistant powered by its Muse Spark model and testing it with staff.
  • OpenClaw is an open-source project supported by OpenAI that links hardware and software and can learn from generated data with less human oversight; Meta’s work aims to be similar.
  • Meta is training an internal agent called "Hatch" with internal testing targeted by the end of June and plans an Instagram shopping agent ahead of Q4; the company has raised its annual capital spending forecast to expand AI infrastructure.

Meta is working on a highly personalized artificial intelligence assistant intended to perform routine tasks for its billions of users, according to recent reporting. The company has been developing agentic AI tools - systems that can take multi-step actions with reduced human oversight - and is building an advanced digital assistant powered by its Muse Spark AI model.

Sources familiar with the work say the assistant is being trialed internally by a select group of employees. The stated objective is to produce a product broadly similar to OpenClaw. In a correction issued alongside coverage of the program, OpenClaw is described as an open-source project supported by OpenAI and is not owned by OpenAI. OpenClaw is designed to bridge multiple hardware and software tools, and to learn from the data those connections generate with considerably less human intervention than conventional chatbots.

Separately, reporting indicates Meta is training an internal AI agent with the codename "Hatch," which draws inspiration from the OpenClaw approach. The company aims to complete internal testing of Hatch by the end of June. In parallel work, Meta plans to incorporate a distinct agentic shopping tool into its Instagram service, with a target launch prior to the fourth quarter of this year.

Meta has not immediately provided comment on these initiatives in response to a request for comment. The developments come after the company raised its annual capital spending forecast late last month, signaling intentions to invest billions more in AI infrastructure. That planned increase in AI-related capital expenditure coincides with concerns that Meta could face losses related to a global youth backlash against social media, a factor the company is confronting as it expands investment.


Summary of developments

  • Meta is building an advanced, agentic AI assistant running on its Muse Spark model and testing it internally.
  • OpenClaw is an open-source project supported by OpenAI that can connect hardware and software and learn with reduced human oversight; Meta aims to develop a similar product.
  • Meta is also developing an internal agent codenamed "Hatch" and plans an Instagram shopping agent targeted for rollout before Q4.

Context on corporate positioning

The company's reported agentic ambitions are occurring alongside a decision to raise annual capital spending guidance to support AI infrastructure build-out. Meta's reported efforts span assistant capabilities, internal productivity agents and consumer-facing shopping tools integrated into its social services.

Risks

  • Increased capital spending on AI infrastructure could heighten financial pressure on Meta if the investments do not yield expected returns; this affects technology and infrastructure capital markets.
  • Potential losses tied to a global youth backlash against social media present user engagement and advertising revenue uncertainty for Meta and the broader social media sector.
  • The internal testing timeline and product integration targets (end of June for "Hatch" testing and pre-Q4 shopping launch) present execution risks if development or rollout schedules slip; this impacts e-commerce and social platforms planning.

More from Stock Markets

Toronto market ends at fresh record as healthcare, financials and materials lead gains Jun 4, 2026 After-Hours Movers: Lululemon Dips on Guidance as Software and Data Names Show Mixed Reactions Jun 4, 2026 Lululemon Lowers Fiscal 2026 Revenue and EPS Guidance as U.S. Demand Softens Jun 4, 2026 Anthropic Places Engineers Inside NSA to Support Mythos AI for Offensive Cyber Tasks Jun 4, 2026 Trump Directs $700M Toward Coal Industry, Lifting Peabody Shares Jun 4, 2026