Key Takeaways

  • Meta is preparing to release Hatch within weeks, with a premium tier under consideration at up to $199.99 per month.
  • Hatch is designed to complete tasks across DoorDash, Etsy, Reddit, Yelp and Microsoft Outlook without requiring users to switch apps.
  • Meta’s larger opportunity may be turning WhatsApp into a distribution channel for its own agent and agents from other developers.

Meta is preparing to release Hatch, a consumer AI agent intended to complete multistep tasks across third-party services, within the next several weeks. Internal plans reviewed by The Information also show that Meta has considered charging as much as $199.99 per month for a premium version.

That price would be roughly 25 times the $7.99 monthly subscription Meta introduced for its AI chatbot three months ago. Final pricing has not been set, however, and the reported figure comes from an internal proposal rather than a public price list.

The distinction matters. Hatch is not simply another conversational chatbot. Meta has trained the agent to interact with services including DoorDash, Etsy, Reddit, Yelp and Microsoft Outlook. A user could ask it to order food, research a purchase, assess reviews or organize an inbox without leaving the interface.

Early prototypes also include a dashboard for tools created on a user’s behalf, according to RuntimeWire. Examples include fitness trackers and travel itineraries. These capabilities push Hatch toward software that acts rather than merely answers.

Autonomy requires significant computational resources. The open-source agent OpenClaw provides a free route to a similar personal-agent experience, but users run it on their own machines and handle command-line setup, model selection, credentials and permissions. Meta’s commercial proposition is convenience backed by centrally managed infrastructure.

OpenClaw's creator reportedly accumulated a $1.3 million monthly token bill while operating agents at scale. That provides some context for the proposed Hatch Plus price. Internal plans describe the premium tier as providing five to ten times the free product’s daily capacity, with allowances resetting during each billing cycle.

In that model, Meta would not really be selling a digital assistant for $199.99. It would be packaging access to inference capacity, external services and agent execution. Whether consumers view those resources as worth more than many productivity-software bundles is another matter.

A product marketing professional with experience across major enterprise tech firms argued on social media that the market for a $199.99 consumer version of OpenClaw could be nonexistent. They also suggested a different interpretation: the figure might reflect how much heavy users could cost Meta in external API charges.

The economics extend well beyond Hatch. Gartner projects overall AI spending will reach roughly $2.6 trillion in 2026, with infrastructure representing more than 45% of the total. Global end-user spending on AI-optimized IaaS is expected to rise from about $21.5 billion in 2025 to more than $42 billion in 2026. Agents that repeatedly call models, APIs and external tools help explain that growth.

Meta is also developing a WhatsApp feature through which users could connect with agents created by other developers. A limited rollout could begin shortly. That gives Meta two routes into the market: selling Hatch directly and positioning WhatsApp as the consumer-facing venue where multiple agents operate.

Positioning WhatsApp as an agent marketplace or access layer allows Meta to benefit even when users choose services such as OpenClaw or alternatives including Google’s Remy. Distribution is a powerful advantage, especially when an agent needs ongoing access to conversations, identity and third-party accounts.

Hatch was initially developed using external large language models. Meta reportedly planned to shift it to its own Muse Spark models before release, while another model called Watermelon is targeted for October. Muse Spark 1.2 and the Muse Code terminal agent arrived on August 5, followed by the 30-billion-parameter, open-weight Muse Glimmer five days later.

Owning the model stack could reduce Meta’s exposure to external API charges. It could also provide tighter control over latency, usage limits and product integration. Still, connecting Hatch to email, shopping and booking services introduces less visible challenges around OAuth 2.0 permissions, data retention, erroneous transactions and user approval. How much freedom should a consumer agent receive before every consequential action requires confirmation?

Those questions will be particularly sensitive in Europe, where WhatsApp is widely used and regulators have previously scrutinized Meta’s combination of data across services. The reporting does not establish whether Hatch will launch in the EU or how Meta will separate information gathered from Outlook, commerce services and its social platforms.

The commercial motivation is clear. Meta generated $60.8 billion in second-quarter revenue, with advertising contributing $59.36 billion, or roughly 98%. Meanwhile, expected capital spending for the year stands at $130 billion to $145 billion, including finance lease payments. Hatch gives Meta a way to connect customer payments to compute consumption. The launch will test whether consumers are prepared to pay for that connection, and whether WhatsApp can become the operating layer for the broader agent economy.