E-commerce

Meta tees up agentic shopping tools as 2026 AI capex rises to 115–135 billion dollars

Mark Zuckerberg says Meta will ship new AI models and products within months, with agentic shopping and personalized context as a key commerce wedge.

Meta tees up agentic shopping tools as 2026 AI capex rises to 115–135 billion dollars
Jan 30, 2026
2 min read
By Sarah Chen

Key Takeaways

  • Meta says new AI models and products will begin shipping within months, with commerce a named priority.
  • Agentic shopping implies multi-step, action-taking assistants, which could shift catalog optimization and attribution on Meta platforms.
  • Meta forecasts 2026 capex of 115–135 billion dollars versus 72 billion dollars in 2025, signaling sustained infrastructure buildout.

Meta is preparing to roll out new AI models and consumer-facing products “in the coming months,” with commerce positioned as a near-term payoff for its retooled research org and massive infrastructure ramp.

Meta’s agentic shopping tools target e-commerce discovery

On Meta’s Q4 2025 earnings call, CEO Mark Zuckerberg pointed to “agentic shopping tools” designed to help people find the right products from businesses in Meta’s catalog. In practice, “agentic” implies software agents that can take multi-step actions on a user’s behalf (for example: narrowing options, checking constraints, and completing a purchase flow) rather than just answering prompts.

For B2B marketers and e-commerce operators, the key shift is discovery. If Meta embeds agents into Facebook, Instagram, and WhatsApp surfaces, product feeds and catalogs may be queried conversationally, with personalization based on signals Meta already holds. That could change how brands think about catalog hygiene, creative testing, and funnel attribution on Meta properties.

The strategic context: AI-driven commerce is becoming a platform race. Google has introduced a protocol for agent-driven purchases, and OpenAI has launched an agentic shopping system with partners including Stripe and Uber, according to TechCrunch reporting that cites those announcements (Google, OpenAI).

Meta’s bet: personal context plus major infrastructure spend

Zuckerberg argued that the differentiator for Meta agents will be personal context—history, interests, content, and relationships—enabling more tailored recommendations and actions. That advantage also raises obvious questions for advertisers about privacy, user consent, and how personalization will be governed across consumer and business tools.

Meta is also backing this direction with capital. The company now expects 2026 capital expenditures of 115 billion to 135 billion dollars, up from 72 billion dollars in 2025, per its earnings release (Meta investor relations).

For operators, the near-term watch item is product packaging: which agent features ship inside ad tools, shops, and messaging flows first—and what new measurement primitives Meta introduces to prove ROI.

Stay Informed

Weekly AI marketing insights

Join 5,000+ marketers. Unsubscribe anytime.

Related Topics

MetaMark Zuckerbergagentic shoppingAI commerceInstagramFacebookWhatsApp