Meta Launches Muse Spark 1.1 at Industry-Low Prices and Plans AI Data Center Rental to Rival AWS
Summary
- • Meta stock surged 5%+ and turned positive year-to-date after CEO Zuckerberg revealed plans to rent AI computing capacity to third parties
- • Muse Spark 1.1 priced at $1.25/M input tokens and $4.25/M output tokens — roughly 80% cheaper than Anthropic's comparable Opus 4.8 at $5/$25
- • Meta is exploring both AWS-style AI model hosting and neocloud chip/server access as new revenue streams beyond its core advertising business
- • A new data center in Canada was announced as Meta's 33rd global facility, underscoring its massive AI infrastructure buildout
Details
Meta stock +5%, turns YTD positive
Shares climbed more than 5% at midday, erasing year-to-date losses that had accumulated on concerns about rising AI capex and uncertain returns
Muse Spark 1.1 released
Meta launched a new AI model Thursday, positioned for developers as a capable and affordable alternative to frontier models from Anthropic and OpenAI
Aggressive token pricing announced
Meta charges $1.25/M input tokens and $4.25/M output tokens vs. Anthropic's $5/$25 for comparable Opus 4.8 — approximately 75–80% cheaper on both input and output
AI compute rental plans revealed
Zuckerberg told Bloomberg Meta is exploring renting AI compute capacity to third parties — either serving competitors' models like AWS, or selling raw chip and server access as a neocloud
33rd data center announced in Canada
Meta revealed a new facility in Canada, announced Wednesday, expanding its global AI infrastructure footprint to support both internal workloads and potential external cloud customers
Developer pricing war accelerated
Meta's pricing undercuts rivals sharply; in a post-tokenmaxxing world, capable lower-cost models may shift adoption away from premium frontier offerings and compress rivals' margins
Rows track Meta's concurrent financial, product, infrastructure, and strategic moves and their market impact.
What This Means
Meta's aggressive Muse Spark 1.1 pricing — at roughly one-fifth the cost of Anthropic's comparable model — signals a deliberate strategy to win cost-sensitive developers and erode rivals' market share. Combined with plans to monetize its massive data center infrastructure through third-party compute rental, Meta is positioning itself not just as an AI model provider but as a full cloud computing competitor, potentially challenging the AWS/Azure/GCP hyperscaler oligopoly. For pure-play AI model companies, a sustained race to the bottom on token pricing could significantly compress margins and shift the competitive dynamic toward companies with diversified revenue. Investors clearly welcomed the dual announcements, but Meta still faces the fundamental challenge of converting its AI infrastructure bet into durable, high-margin revenue streams.
Sentiment
Broadly excited about aggressive pricing and cloud ambitions, with some notes on coding gaps
“Alexandr Wang (Meta's Chief AI Officer): 'The goal is to really have attractive pricing that scales with immense consumption usage.' Here is what that pricing actually looks like against the field: Meta Muse Spark 1.1: $1.25 per million input tokens, $4.25 per million output — $20 in free credits on every new account. positioning: called 'very aggressive and attractive' compared directly to anthropic and openai pricing.”
“Meta just dropped Muse Spark 1.1, it's actually insane. Agentic performance rivaling GPT-5.5 + Opus 4.8, 1M context, parallel sub-agents, real computer use that scripts OR clicks intelligently. This + the new public Model API at crazy low prices? The agent wars just went nuclear.”
“meta muse spark 1.1 is more about cheap agents than model hype. $1.25 input $4.25 output 1m context parallel subagents strong tool use scores. coding is still opus and gpt territory but agents that use tools, click through apps, run research and call APIs are the real play here.”
“Meta just detonated the AI pricing war - Muse Spark 1.1 is live, dirt-cheap, and explicitly built to run autonomous Agents.”
“From testing it today, I am very bullish on Muse Spark 1.1. It's nothing like the Llama Models and I would say comparable to OpenAI/Anthropic for a decent sized subset of my usage. Overall, the model accuracy is frontier worthy and the RL on it is almost magical as it's very intuitive akin to Fable 5. One surprising thing is that Muse Spark is blazing fast... This makes me wonder if the best AI Models will all become their own Clouds.”
Split
~75/25 positive/negative — most excited by pricing/agent focus and Meta's cloud entry; minority notes coding still lags competitors and questions long-term monetization.
Sources
- Meta lifts capital expenditure forecast, doubling down on AI push - Yahoo FinanceFinance
- Google-parent Alphabet soars as rivals stumble over AI costs - Yahoo FinanceFinance
- Meta is still burning money on AR/VRTechCrunch
- Meta shares slide as investors weigh Big Tech's AI spending spree - BBCBbc
- Big Tech’s massive spending is back in focus on Wall Street - CNNEdition
- Meta, like SpaceX, looks to turn excess AI compute into cashTechCrunch
- Meta falls as massive AI infrastructure spending plans overshadow chip progress - Yahoo FinanceFinance
- Meta enters the crowded AI coding battle with Muse Spark 1.1TechCrunch
- Meta’s new AI chips will begin production in SeptemberTechCrunch
- Meta launches low-cost Muse Spark 1.1 as enterprise AI spending comes under scrutinyComputerworld
- Meta's new AI chips will begin production in September (4 minute read)Techcrunch
- Meta stock turns positive on the year on data center plans, new AI model pricing - Yahoo FinanceFinance
- Why Meta Platforms Stock Surged This Week - Yahoo FinanceFinance