TLDR
- Meta shares rose 6.88% as AI infrastructure costs came in below expectations
- Meta plans to add 14GW of AI computing capacity through the 2027 expansion
- Lower AI infrastructure costs support Meta’s long-term growth strategy
- Meta advances Iris AI chip production alongside Broadcom and TSMC partners
- Meta expands AI capacity while improving infrastructure deployment efficiency
Meta shares climbed 6.88% to $674.94 after breaking above the $670 level during mid-morning trading. The stock later traded near session highs, with $670 emerging as near-term support and $680 remaining immediate resistance.Attention shifted toward the company’s expansion of AI infrastructure and lower deployment costs.
Meta Expands AI Infrastructure at Lower Costs
Meta continues expanding its artificial intelligence infrastructure while reducing deployment costs across new computing capacity projects. Reuters reported that an internal company memo outlined faster infrastructure growth than previously disclosed. The update also showed that the company achieved stronger efficiency across its latest expansion plans.
The memo stated that Meta targets an additional 14 gigawatts of computing capacity across 2026 and 2027. The company already deployed one gigawatt during 2026 and expects another 5.5 gigawatts before year-end. As a result, total capacity growth continues accelerating alongside broader artificial intelligence development.
Meta also allocated approximately $145 billion for its capital expenditure programme supporting long-term infrastructure expansion. Based on the reported capacity targets, infrastructure costs appear lower than earlier external estimates.The reported figures indicate improved spending efficiency across large-scale computing projects.
Infrastructure Efficiency Supports Long-Term Expansion
Artificial intelligence requires large data centres, advanced processors, and substantial electricity to support model training and deployment.Infrastructure spending remains one of the largest commitments for technology companies developing advanced AI systems. Meta continues expanding its network while increasing available computing capacity across multiple facilities.
Reuters reported that the internal memo pointed to significant improvements in deployment efficiency during the current expansion cycle. The reported capacity additions exceeded earlier expectations outlined before the latest infrastructure update. The expansion schedule remains focused on increasing computing resources through 2027.
The reported cost reductions relate to existing infrastructure development rather than future hardware products. This distinction highlights improvements within Meta’s current deployment strategy and construction process.The company continues scaling its artificial intelligence infrastructure while maintaining lower reported development costs.
Custom AI Chips Strengthen Future Development Plans
Meta also plans to begin manufacturing its custom artificial intelligence chip, known internally as Iris, later this year. The programme follows successful testing and complements the company’s existing purchases of graphics processing units. Production will proceed alongside manufacturing partners Broadcom and TSMC.
The Iris chip forms part of Meta’s broader strategy to expand internal hardware capabilities over time. Reuters reported that the company intends to introduce updated custom chips about every six months through 2027.Meta secured multi-year supply agreements with key manufacturing partners supporting those plans.
The reported infrastructure improvements remain separate from the Iris chip programme and its future production schedule. Instead, Reuters indicated that existing deployment strategies produced the current efficiency gains across computing capacity projects.The infrastructure expansion and custom chip roadmap provide additional background for Meta’s long-term artificial intelligence development efforts.


