Meta Platforms Inc. has launched Muse Spark 1.1, introducing its first-ever paid developer tier via the Meta Model API. Priced aggressively at $1.25 per million input tokens—roughly 25% of the cost of rivals OpenAI and Anthropic—the model represents a major strategic pivot toward proprietary, closed AI. While CEO Mark Zuckerberg and AI Chief Alexandr Wang push to capture market share, Wall Street remains skeptical of Meta's record 2026 capital expenditures, warning that falling token prices and rising competition from cheap Chinese models could severely squeeze hyperscaler margins.
Muse Spark 1.1 pricing
- ▪Meta Platforms Inc. is charging developers $1.25 per million input tokens and $4.25 per million output tokens for Muse Spark 1.1 API access, with new accounts receiving $20 in free credits
- ▪Meta Platforms Inc. launched Muse Spark 1.1 on July 9, 2026, introducing a paid tier for developers to access the model via a new Meta Model API
Meta AI strategy pivot
- ▪Meta Platforms Inc. is currently training a more powerful AI model code-named Watermelon, following the release of Muse Spark 1.1, which was code-named Avocado
- ▪Meta Platforms Inc. has pivoted from its historical strategy of releasing open-source models for free toward prioritizing closed, paid models like Muse Spark 1.1
Competitive pricing pressure
- ▪The introduction of low-cost models from Meta Platforms Inc. and xAI is expected to intensify a domestic price war among frontier AI model providers
- ▪Mark Zuckerberg stated that Meta Platforms Inc. priced Muse Spark 1.1 aggressively at roughly 25% of the cost of top competing models from OpenAI and Anthropic PBC
AI infrastructure spending concerns
- ▪Meta Platforms Inc. is projecting record capital expenditures for 2026 to fund data centers, expensive AI chips, and infrastructure projects
- ▪Meta Platforms Inc. announced a new $10 billion data center investment in Canada as part of its aggressive infrastructure spending to develop superintelligence
Return on investment skepticism
- ▪A slower-than-expected payoff on AI investments could trigger a Magnificent 7 stock sell-off, margin squeezes, and increased credit risk for hyperscalers
- ▪Wall Street and critics like Apollo chief economist Torsten Slok are questioning whether Meta Platforms Inc.'s massive AI investments will yield expected cash flows
Chinese AI model competition
- ▪Chinese AI models are gaining global market share and leading their United States counterparts in token usage among the top 20 most-used models
- ▪Falling token prices and rising competition from cheaper Chinese AI models present a risk that hyperscaler cash flow projections may prove too optimistic
Story comments
Loading comments…