Meta made two significant moves in recent weeks that are worth understanding together.
First, the company launched Muse Spark, its first major proprietary AI model developed under Meta Superintelligence Labs. This is the unit built around Alexandr Wang following his $14.3 billion departure from Scale AI in 2025. Second, Meta raised its 2026 capital expenditure guidance to between $125 billion and $145 billion, up from the already aggressive $115 to $135 billion range it announced in January.
Both moves represent a meaningful shift in Meta's AI strategy.
Muse Spark is a deliberate break from what made Meta's AI reputation. The company built that reputation on open-source models. Its Llama family drew over 1.2 billion cumulative downloads, and open-source access became a core part of Meta's developer relationship. Muse Spark is proprietary. Meta says it hopes to open-source future versions of the model, but there is no timeline or commitment attached to that.
Technically, Meta describes Muse Spark as delivering competitive performance in multimodal perception, reasoning, health, and agentic tasks at significantly lower compute cost than its older Llama 4 midsize variant. The company credits improved training techniques and rebuilt infrastructure for the efficiency gain. The model launched roughly ten months after Wang joined and after a period that produced few visible model releases. Morningstar analyst Malik Ahmed Khan put it plainly: "Meta had to show investors and operators they have been working on something of substance. That's the first step."
Step two is monetization. Muse Spark needs to compete with Claude, GPT-5, and Gemini in enterprise settings. That question remains open.
On the spending side, Meta's first quarter 2026 revenue reached $56.3 billion, up 33% year over year, with a 41% operating margin. The company held $81.2 billion in cash at quarter's end. These are not the financials of a company funding AI through desperation. The capex increase was driven by higher memory prices and new data center costs, according to management. Multi-year infrastructure contract commitments jumped by $107 billion in a single quarter. The company also locked in a $21 billion AI infrastructure agreement with CoreWeave through 2032.
Meta Compute, a new internal division, will manage this data center capacity buildout going forward, with plans for tens of gigawatts of capacity this decade.
For enterprise buyers, the Muse Spark launch changes a specific and practical thing. Open-source Llama gave developers and companies a free, customizable foundation. Muse Spark removes that option for Meta's most capable models, at least for now. Organizations that built AI workflows on Llama's open-source base are now looking at a future where Meta's leading models require paid access.
Meta's 1 billion monthly AI users give it distribution no other AI lab can match. If Muse Spark performs as advertised and Meta builds a coherent monetization strategy around it, that scale advantage could matter a great deal. Whether the $145 billion bet translates into AI market leadership is the question every enterprise technology buyer is now watching.
