AIToday
Large Language ModelsOpen-Source AIAI Business & IndustryYahoo Finance AIPublished: Aug 29, 2026, 19:02 JST2 min read

Meta open-sources Muse Glimmer, keeps Muse Spark 1.2 closed

Meta open-sources Muse Glimmer, keeps Muse Spark 1.2 closed

Key takeaway

  • Meta released its open-weights AI model, Muse Glimmer, on Aug. 10.

  • It has not yet released the weights for its flagship, Muse Spark 1.2.

  • Meta's AI investments are boosting ad performance on Facebook and Instagram.

3 Key Points

  1. What happened

    Meta Platforms released Muse Glimmer, a 30-billion-parameter AI model with open weights, on Aug. 10. As of Aug. 25, it had not released the weights for its more powerful flagship model, Muse Spark 1.2.

  2. Why it matters

    Unlike OpenAI and Anthropic, Meta doesn't rely on subscriptions or APIs to monetize AI models. Instead, it uses AI to improve ads and recommendations on Facebook and Instagram, which already showed results: in Q2, AI improvements boosted Facebook ad clicks by 8.3% and ad conversions by 15.7%.

  3. What to watch

    Meta's capital spending for 2026 is now expected to be $130 billion to $145 billion, up from the initial $115 billion to $135 billion. The company also has $81.6 billion in noncancelable commitments due in 2027 and another $68 billion in data center leases starting in 2027-2028.

Ask the AI about this article →

Context & Analysis

Meta's decision to open-source a less powerful model while keeping its flagship closed suggests a careful balance between transparency and competitive advantage. The company's core monetization is not through selling AI models, but through enhancing its ad and recommendation systems on Facebook and Instagram. Early results are encouraging, with AI improvements leading to an 8.3% increase in ad clicks and a 15.7% rise in ad conversions in Q2. This strategy could justify heavy spending on AI infrastructure, but the company's capital expenditures are rising sharply, with 2026 expectations now at $130 billion to $145 billion. Depreciation is also climbing, up 40% year over year, which could pressure future margins. Meta's valuation appears reasonable at around 16.8 times expected 2027 earnings, but the company faces the challenge of ensuring that AI investments continue to lift revenue per user, especially in lower-monetization markets like Asia-Pacific, where much of its ad growth is coming from. If spending remains high without corresponding revenue gains, the stock may struggle.

FAQ

What is the difference between Muse Glimmer and Muse Spark 1.2?
Muse Glimmer is a 30-billion-parameter model that was released with open weights on Aug. 10. Muse Spark 1.2 is a more powerful flagship model, but as of Aug. 25, its weights had not been released.
How is Meta's AI strategy different from OpenAI's and Anthropic's?
Meta doesn't rely primarily on subscriptions, APIs, or enterprise software to monetize its AI models. Instead, it uses AI to improve recommendations and advertising across Facebook and Instagram, and it is developing business messaging and AI-agent opportunities through WhatsApp and Messenger.
What is Meta's planned capital spending for 2026?
Meta now expects $130 billion to $145 billion in capital spending in 2026, up from its initial range of $115 billion to $135 billion.
Yahoo Finance AIRead Original Article

Get the latest Large Language Models news every morning

For example, today's edition would include:

  • Google's WikiSkill gives AI agents a memory of past mistakesTHE DECODER · 19m ago
  • AI job losses so far: only 3% of US workersFortune AI · 19m ago
  • Chatbot on Your Own PC: Privacy, Offline, FreeWIRED AI · 19m ago

AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.

Free · takes 30 seconds · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. Q&As are published on this page for other readers too.

Related Articles

Next articleOpenAI cuts off Cursor after SpaceX acquisition