Key points
- Meta adds three Muse models to Google Cloud
- Google Cloud offers OpenAI and Anthropic models alongside Gemini
- Meta is also a major Google Cloud customer
Meta Platforms (META) has made three Muse AI models available through Google Cloud, the cloud business of Alphabet (GOOGL). The models join Google's own Gemini and offerings from rival AI labs in its Model Garden catalog for business customers.
Meta CEO Mark Zuckerberg announced on Nov. 9 that Muse Ember, Muse Spark 1.3, and Muse Wonder were available on Google Cloud. Alphabet CEO Sundar Pichai welcomed the models to Model Garden less than an hour later.
The arrangement adds another link between the companies. Meta is already a Google Cloud customer: CNBC reported in August 2025 that it had agreed to spend more than $10 billion on Google Cloud services over six years, mainly for AI infrastructure. Google declined to comment, and neither company has publicly disclosed the terms.
Google Cloud is the only cloud listed for Muse Wonder
Muse Wonder is one of Meta's most capable models, released on Oct. 12. Meta's developer page lists three ways to build with it: Google Cloud, Meta's own Model API, and OpenRouter, a service that routes requests to different AI models through one account. Amazon Web Services and Microsoft Azure aren't listed.
The larger Muse Wonder Full wasn't among the models Zuckerberg named. Meta says it remains available only through a subscription to its Muse Code coding tool.
Meta's API charges $4 per million input tokens and $20 per million output tokens for Muse Wonder. Neither Zuckerberg nor Pichai disclosed the financial terms of the Google Cloud arrangement.
Google Cloud offers models from OpenAI, Anthropic, and Meta
With Meta's Muse models, Google Cloud offers models from Meta, OpenAI, and Anthropic alongside its own Gemini models. Google Cloud brought OpenAI's GPT-6 models to Model Garden on Oct. 21 in a private preview for select enterprise customers, and Anthropic's Claude models were already there. At the time, before Meta joined, Google said it was "the only hyperscale cloud provider offering access to all three leading closed source LLMs."
The billing is the part that matters to Google's numbers. In the OpenAI announcement, Google said that usage shows up on a single Google Cloud invoice and "draws down existing cloud commitments." So a company that has already promised to spend a set amount with Google can spend part of it on OpenAI's models. Google hasn't said whether the Muse models are sold the same way.
Google Cloud's backlog is $560.8 billion
Revenue backlog is money customers have committed under contract that a company hasn't recognized as revenue yet. Alphabet's third-quarter 10-Q puts its backlog at $567.7 billion as of Sept. 30, and $560.8 billion of that belongs to Google Cloud. Alphabet expects to recognize just over 55% of the total within 24 months.
That puts Google Cloud in the same range as the largest books on our AI backlog tracker: Microsoft's $677 billion of commercial remaining performance obligations, Oracle's $664 billion as of Aug. 31, and Amazon's $578 billion. The four aren't counted the same way. Alphabet began including contracts of a year or less in the first quarter of 2026, Amazon counts only contracts longer than a year, and Oracle's figure covers the whole company.
