Meta stock surges 15% as company releases Muse Spark 1.1 agentic model and Muse Image generator
Meta's stock surged 15% this week following the release of two AI models: Muse Spark 1.1 for agentic and coding workloads on Thursday, and Muse Image for image generation on Tuesday. The releases come three months after Meta introduced its first foundation model, Muse Spark, as the company competes with OpenAI, Anthropic, and Google.
Muse Spark 1.1 — Quick Specs
Meta stock surges 15% as company releases Muse Spark 1.1 agentic model and Muse Image generator
Meta's stock rose 15% this week, marking its best weekly performance since early 2024, following the release of two new AI models aimed at competing with OpenAI, Anthropic, and Google.
On Thursday, Meta released Muse Spark 1.1, an updated version of its foundation model designed for agentic and coding workloads. Two days earlier, the company unveiled Muse Image, an AI model for image generation targeted at creators and advertisers on its subscription platform.
The releases follow Meta's introduction of Muse Spark three months ago, the company's first proprietary foundation AI model. CEO Mark Zuckerberg is leading an aggressive push into AI model development through Meta Superintelligence Labs, headed by Alexandr Wang.
Stock performance and market context
The 15% weekly gain erased Meta's year-to-date losses, bringing the stock up 2% for 2026. The Nasdaq, by comparison, has gained 13% this year.
Meta shares gained an additional 6% on Friday alone, according to CNBC reports.
Custom chip development underway
Meta received a further boost from reports of progress on its custom AI chips. According to Reuters, the company plans to begin manufacturing its first chip, code-named Iris, in September. The chip is part of Meta's data center expansion strategy announced in March, with a target of reaching 14 gigawatts of computing power in 2027.
Justin Post, a Bank of America analyst, noted that "Meta may have engineered significant cost savings to get capacity cost per MW well below our and Street expectations."
What this means
Meta is making a concentrated effort to establish itself as a major player in AI models, moving beyond its advertising business to compete directly with leading AI labs. The rapid release cadence—three major announcements in three months—signals urgency in catching up to competitors with established foundation models. The custom chip development, if successful, could significantly reduce Meta's reliance on external hardware providers and improve margins on AI infrastructure. However, Meta remains behind competitors in the foundation model race, and these releases represent catch-up moves rather than market leadership.
Related Articles
Alibaba Launches Wan3.0, Generating AI Videos Up to 30 Seconds From Text, Images, and Documents
Alibaba's Wan3.0 video generation model is now in beta, producing clips up to 30 seconds long and accepting text, images, video, audio, and documents like PDFs and PowerPoint files as input. Pricing runs per second of output across two tiers and three resolutions.
Mystery 'Stealth Model' Ox Alpha Appears on OpenRouter, Sparking Speculation Over Its Creator
A new AI model called Ox Alpha appeared on OpenRouter this week, listed only as being built by an anonymous third-party provider. Speculation about its creator has ranged from Z.ai's GLM models to an unreleased version of Microsoft's MAI, with no confirmation yet.
DeepSeek Releases Experimental V4-Flash-Vision-Exp, Claims Near-Parity With Opus 4.8 on Agent Benchmarks
DeepSeek has released V4-Flash-Vision-Exp, an experimental multimodal extension of V4-Flash that adds image understanding while preserving text reasoning capabilities. The company claims the model approaches or beats Anthropic's Opus 4.8 on its internal multimodal agent benchmarks.
DeepSeek Releases V4 Flash Vision Exp, an Experimental Multimodal MoE Model with 1M Context
DeepSeek has released V4 Flash Vision Exp, an experimental vision-enabled variant of DeepSeek V4 Flash 0731 that adds image understanding while matching the base model's text performance. The sparse mixture-of-experts model uses 13B active parameters out of 284B total and supports a 1M token context window.
Comments
Loading...