Meta stock rises as AI chip plans and Muse Spark 1.1 rollout take focus

Meta Platforms Inc. META shares traded higher on Thursday, reversing early losses as investors assessed the company's latest artificial intelligence initiatives.

This includes plans to begin manufacturing an in-house AI chip and the public preview of its upgraded Muse Spark 1.1 model.

The developments show Meta's efforts to strengthen its AI infrastructure, reduce reliance on third-party chip suppliers, and expand its portfolio of proprietary AI products as competition with OpenAI, Anthropic and Google intensifies.

Meta advances custom AI chip strategy

According to an internal memo reviewed by Reuters, Meta plans to begin manufacturing its in-house artificial intelligence chip, code-named Iris, from September.

The chip is part of the company's four-generation Meta Training and Inference Accelerators (MTIA) program and is designed to improve the AI systems powering Facebook and Instagram.

Reuters reported that testing of the chip took just six weeks and uncovered no major issues, marking progress for an in-house chip initiative that has faced challenges since it began more than five years ago.

Meta is working with Broadcom on the chip's design, while Taiwan Semiconductor Manufacturing Co. (TSMC) will manufacture it.

The custom silicon is intended to complement the large number of graphics processing units (GPUs) Meta purchases from Nvidia and Advanced Micro Devices (AMD) while helping the company lower computing costs and reduce dependence on external suppliers.

Meta plans to deploy seven gigawatts of computing infrastructure this year before doubling capacity to 14 gigawatts in 2027.

To support its AI expansion, the company has secured long-term supply agreements with Samsung Electronics for memory chips, Sandisk for flash storage and Sumitomo Electric for fiber-optic equipment.

Muse Spark 1.1 expands Meta's AI ambitions

Alongside its infrastructure investments, Meta introduced Muse Spark 1.1, describing it as the latest version of its AI model focused on coding and agentic AI capabilities.

The company is making the model available through a public developer preview, allowing developers to join a waitlist for API access through Meta's developer portal.

"This is going to be served on top of the computer infrastructure that we’ve built," said Meta's AI chief Alexandr Wang.

Wang said the updated model is Meta's "strongest model for agentic and coding work yet."

He also said the pricing is "very aggressive and attractive" compared with competing offerings from OpenAI and Anthropic. New API users will receive $20 in free credits before usage-based pricing begins.

"The goal is to really have attractive pricing that scales with immense consumption usage," Wang said.

According to Wang, Meta trained Muse Spark 1.1 to strengthen coding capabilities because "You kind of have to build coding capabilities as part of that in service of overall agentic capabilities."

Meta said it is currently limiting API access to its own ecosystem rather than making the model available through third-party AI marketplaces.

AI investments continue to expand

The latest announcements come as Meta continues to increase spending on AI infrastructure.

Reuters reported that the company expects to spend as much as $145 billion on AI infrastructure this year, accounting for a significant share of Big Tech's projected AI investment.

The company also announced plans to build its first Canadian data center in Sturgeon County, Alberta.

The AI-optimized facility represents an investment of more than CAD $13 billion, is expected to support about 3,000 construction jobs at peak and more than 300 operational positions, and will run on 100% clean and renewable energy.