Meta Launches AI Agent Subscription Product Hatch, Priced Up to $199.99/Month
chuong wang
Meta is considering pricing its consumer AI agent product Hatch at up to $199.99/month, matching OpenAI's and Anthropic's top tiers — the company's first attempt to turn its massive AI spending into subscription revenue beyond advertising.
What is Hatch, and what can it do?
Hatch is the consumer version of Meta's open-source agent project OpenClaw. Its core pitch: task automation. Describe what you need in plain language — "build me a fitness tracker" — and it generates a working tool.
Key features include "vibe coding" — writing code through everyday language instead of manual programming — plus calendar management, email drafting, and modular "skills" plugins that connect to third-party services.
This means → Hatch targets doing things for you, not chatting with you. The jump from "Q&A assistant" to "execution agent" is the central battleground in AI products right now.
$199/month — who is it competing with?
The paid tier, Hatch Plus, offers 5 to 10 times the daily usage quota of the free version. Token allowances reset each billing cycle and do not roll over — the same mechanic as ChatGPT Pro and Claude Max.
OpenAI and Anthropic both charge $200/month for their top subscriptions and have built large paying user bases. Meta's AI products have been almost entirely free until now — this is its first high-price subscription play.
In plain terms = Meta used to give away free AI tools to keep users on-platform and sell ads. Now it wants AI itself to generate direct revenue — a path OpenAI and Anthropic have already proven, with Meta arriving late.
How does Zuckerberg frame the direction?
On the Q2 earnings call in April, Zuckerberg said: "Our goal is not just to deliver Meta AI as an assistant, but to deliver agents that understand your goals and work toward them day and night."
He hinted users would create their own tools through vibe coding, predicting a "significant increase in entrepreneurial activity."
This reflects a strategic pivot: Meta's AI positioning is shifting from "a feature embedded in social platforms" to "a standalone, paid productivity product" — not just a new feature, but a change in business model.
Whose technology powers it, and when does it launch?
During development, Hatch runs on Anthropic's Claude Opus 4.6 and Sonnet 4.6. At launch, Meta plans to switch to its own model, Muse Spark.
In plain terms = Meta is borrowing Anthropic's models to build the prototype, but intends to replace them with its own — whether that switch goes smoothly will directly determine product quality and margin.
Timeline: an initial April U.S. launch was pushed back in May to a controlled pilot with roughly 10 companies, followed by branding finalization and a full rollout targeted for July. The schedule may still shift.
How much does this matter for Meta's finances?
Meta's 2025 revenue topped $200 billion, nearly all from advertising. Hatch is its most direct attempt to convert heavy AI capital expenditure into non-ad revenue.
Product marketing manager Jennifer Lin confirmed in an April internal memo that the name "Meta AI Skills" is not finalized; branding decisions are still underway.
This means → Whether Hatch can build a paying subscriber base in a market where OpenAI and Anthropic already have footholds is the key test of Meta's revenue diversification strategy — and with even the product name unsettled, that test is still far off.
Content is for reference only, not financial advice.