Key Highlights
- Shares of Meta gained 2% Tuesday morning before settling at a 0.75% increase following reports that the company will introduce proprietary AI chips to its data centers in early 2027.
- The processor, known as MTIA 450 or Arke, represents Meta’s third-generation custom silicon, built in collaboration with Broadcom and TSMC.
- A subsequent fourth-generation chip named Astrid is expected to complete design phases within approximately 30 days and arrive in data centers by late 2027.
- This strategic initiative targets decreased dependence on Nvidia hardware while lowering energy consumption and infrastructure expenses.
- Wall Street analysts maintain a Strong Buy rating on META with a consensus 12-month price target of $758.03, suggesting 13.5% potential upside.
Meta Platforms (META) shares experienced upward movement Tuesday morning, climbing as high as 2% following a Bloomberg report indicating the social media giant intends to deploy proprietary AI processors in its data facilities during the first half of 2027. The stock subsequently moderated its gains, trading approximately 0.75% higher.
The processor generating attention is designated as MTIA 450, alternatively referred to as Arke. This represents the third iteration in Meta’s proprietary silicon initiative, which the company initially unveiled in 2023.
Meta is presently conducting testing on the Arke processor. A subsequent fourth-generation model, internally code-named Astrid, is anticipated to complete design activities within the next month and become operational in data centers by late 2027.
Yee Jiun Song, Meta’s vice president of engineering overseeing the custom chip initiative, informed Bloomberg that each successive chip generation provides enhanced performance per watt of energy consumed and per dollar invested relative to commercially available alternatives.
Song further indicated that these processors can execute AI models with greater efficiency than existing Nvidia processors, especially for inference operationsāthe phase where AI models produce real-time responses.
Decreasing Reliance on Nvidia Hardware
Meta engineered these chips through a partnership with Broadcom handling design aspects and Taiwan Semiconductor Manufacturing (TSMC) managing production. The objective centers on diminishing the company’s dependence on Nvidia equipment for executing AI workloads.
Meta has pledged to deploy more than a gigawatt’s capacity of its proprietary chips across a 12-month timeframe. Song emphasized that at such magnitude, absorbing a 30% cost escalation becomes unfeasible, underscoring the importance of the internal development approach.
The initial batch of 12 test processors arrived from TSMC on September 1. Performance metrics aligned within 2% to 3% of design projections. Immediately, engineers executed Meta’s proprietary AI models alongside tools from DeepSeek and Alibaba.
Prioritizing Inference Over Training
Meta had previously pursued development of a chip called Olympus, scheduled for 2028 or 2029, which was intended to manage both model training and inference functions. The company discontinued that initiative citing cost considerations.
The present chip roadmap concentrates exclusively on inference capabilities, maintaining reduced costs while satisfying Meta’s requirements for operating AI systems at scale.
Meta Superintelligence Labs contributes to the effort, communicating forthcoming model specifications with the engineering division to ensure chips are tailored around practical deployment scenarios.
Tuesday’s stock appreciation follows recent positive momentum stemming from Meta’s introduction of the Muse AI agent.
According to TipRanks, META presently carries a Strong Buy consensus rating derived from 44 analyst evaluations conducted over the previous three months. The distribution includes 38 Buy recommendations and six Hold recommendations.
The consensus 12-month price target registers at $758.03, suggesting 13.5% upside potential from present trading levels.
Get 3 Free Stock Ebooks
Discover top-performing stocks in AI, Crypto, and Technology with expert analysis.
- Top 10 AI Stocks - Leading AI companies
- Top 10 Crypto Stocks - Blockchain leaders
- Top 10 Tech Stocks - Tech giants
