Meta Issues Non-Voting Share Class to Accelerate AI Infrastructure Spending
Meta announces a new class of non-voting shares to fund its accelerating AI initiatives, with capital expenditure guidance climbing to $135–$145 billion for 2026.
Last verified:
Meta’s Non-Voting Share Strategy
According to The Verge AI, Meta announced a new class of non-voting shares as part of its third-quarter 2026 earnings report, filed on October 22. The issuance will fund artificial intelligence initiatives, potential acquisitions, and employee stock compensation, with trading to commence in 2027. Meta Chief Financial Officer Susan Li characterized the move as essential to “expand our funding tools and give us flexibility to invest behind the unprecedented momentum we’re seeing across every part of our business driven by AI.”
The non-voting structure is strategically significant: it allows Meta to raise capital without dispersing voting authority, preserving founder control while demonstrating commitment to sustained AI spending to the investment community.
Escalating Capital Expenditure Targets
Meta’s capex guidance now ranges from $135 billion to $145 billion for 2026, up from the prior forecast of $130 billion to $145 billion. This upward revision reflects intensifying competition in large-language models and data-center infrastructure. The company has been recruiting engineering talent from competitors including OpenAI while building proprietary foundation models to challenge offerings from Google DeepMind, Anthropic, and others.
The $135–$145 billion annual burn rate positions Meta among the highest-capex technology companies globally, signaling an all-in bet on AI capability and competitive positioning.
Muse Agent Traction and Early Friction
Meta launched its Muse AI agent in September 2026, enabling automation of tasks such as email triage, restaurant reservations, and travel logistics. According to The Verge AI, Muse reached 10 million downloads in the United States within its first month, and a companion model called Muse Wonder rolled out concurrently.
However, the agent has encountered early operational friction. A security researcher reported that Muse disclosed a user’s home address to a stranger via Facebook Marketplace, raising privacy concerns. Additionally, the company suspended the conversation-sharing feature after discovering that chat logs were publicly accessible on the web. These incidents represent typical growing pains for agentic systems, but they underscore the tension between feature velocity and user trust.
Why This Matters
The non-voting share issuance signals Meta’s conviction that the AI infrastructure race requires sustained, multi-year capital deployment at scales competitors may struggle to match. For investors, this structure preserves the company’s governance while signaling disciplined long-term planning. For teams evaluating AI partnerships, Meta’s capex commitment and Muse’s early momentum suggest the company is serious about building end-user-facing agents—though reliability and privacy hardening will determine whether adoption sustains beyond early novelty.
If Meta’s capex translates to measurable model-quality improvements and agentic reliability over the next 12 months, the share issuance could prove a watershed moment in the company’s pivot from a social network to an AI infrastructure and applications provider.
Frequently Asked Questions
Why is Meta creating a new share class instead of using existing shares?
The non-voting structure allows Meta to raise capital without diluting voting control for existing shareholders. It also provides flexibility for acquisitions and employee compensation while signaling long-term AI commitment to investors.
How much will the capex increase impact Meta's profitability?
The guidance of $135–$145 billion represents an increase from the prior $130–$145 billion range. The long-term profitability impact depends on the return on those infrastructure investments and competitive pricing in AI services.
What is Meta's Muse agent and how is it performing?
Muse is Meta's AI agent launched in September 2026, capable of tasks like email management and trip planning. It reached 10 million US downloads in its first month but has faced privacy concerns and feature rollbacks.