Mark Zuckerberg told investors on January 28, 2026 that Meta will begin shipping a new generation of AI models and products in the coming months after a year of internal restructuring, and he flagged AI-driven commerce tools as a key focus. The executive said the company rebuilt the foundations of its AI program in 2025 and expects steady technical progress through 2026. He linked the product timetable to recent hires, acquisitions and upgraded infrastructure spending, and said the work will surface in consumer-facing services soon. The comments were made during an earnings-related investor call held the same week Meta published its quarterly results.
Key Takeaways
- Meta announced a planned rollout of rebuilt AI models and products beginning within months of January 28, 2026, following a 2025 reorganization of its AI lab.
- Zuckerberg singled out “agentic” shopping tools that can recommend products from Meta’s business catalog by using personal context.
- Meta confirmed the December 2025 acquisition of Manus, a general-purpose agent developer, and said Manus will both continue as a service and be integrated into Meta products.
- The company raised its 2026 capital expenditure outlook to between $115 billion and $135 billion, up from $72 billion in 2025, citing support for Meta Superintelligence Labs and core business needs.
- The cited infrastructure plan still falls short of an earlier, reportedly projected $600 billion figure for spend through 2028.
- Competitors including Google and OpenAI are already pursuing agent-enabled commerce and partnerships with payment and delivery platforms such as Stripe and Uber.
- Meta argues its access to personal context — history, interests, content and relationships — will differentiate its agents, though how that advantage will be realized commercially remains to be shown.
Background
Meta’s public AI ambitions follow several years of large-scale investment in models, data centers and talent. In 2025 the company reorganized its internal AI teams into a restructured lab and said it would focus on building more integrated, product-ready systems. The reorganization aimed to concentrate research, engineering and product teams around shared model infrastructure to accelerate deployment.
Agentic AI refers to systems that perform multi-step tasks on behalf of users, often by combining language models, retrieval of personal or contextual signals, and transactional connectors. Across the industry, firms from Google to OpenAI have pursued agent frameworks and commercial partnerships to enable payments, bookings and recommendations. This competitive context helps explain why Meta is highlighting both infrastructure and commerce use cases.
Main Event
On the investor call on January 28, 2026, Zuckerberg said Meta would begin shipping new models and products over the coming months and expected steady progress through the year. He emphasized the company rebuilt the foundations of its AI program in 2025 and that the current phase is focused on productization. He declined to give exact release dates or enumerate specific features beyond a focus on commerce tools.
Meta framed commerce as a priority area for agentic systems, saying new shopping assistants will help people find the right products among businesses listed in its catalog. Zuckerberg argued that the agents’ value comes from personal context — data about users’ history, interests, content and relationships — which Meta believes it can leverage to deliver more personalized recommendations.
In December 2025 Meta acquired Manus, a general-purpose agent developer, and pledged to operate and sell Manus as a standalone service while integrating its technology into Meta offerings. The Manus deal is positioned as both an immediate capability boost and a longer-term integration target for Meta’s agent roadmap.
The earnings release that week showed a marked increase in planned infrastructure spending. Meta now expects to spend between $115 billion and $135 billion in capital expenditures in 2026, versus $72 billion in 2025, attributing the rise to investments supporting Meta Superintelligence Labs and core business operations.
Analysis & Implications
If Meta executes on a quick public rollout of agentic models, the product move could accelerate competition over transactional AI features across platforms. Agent-enabled commerce ties natural-language assistance to payments, inventory and fulfillment systems, and firms that can stitch these elements together at scale may capture more direct commercial value from AI interactions.
Meta’s argument that personal context will differentiate its agents highlights a tension between personalization and privacy. Access to richer user data can improve relevance but raises regulatory and reputational risk, especially in jurisdictions with strict data-protection rules. How Meta chooses to surface and control data usage will shape both product uptake and regulatory scrutiny.
The jump in 2026 capex signals a heavy hardware and infrastructure commitment that could strengthen Meta’s long-term ML capacity but will heighten investor pressure for clear monetization paths. Historically, shareholders have demanded more clarity about how AI investments translate to revenue; delivering tangible product wins this year will be important to justify elevated spending guidance.
Strategically, capturing commerce flows within Meta’s apps would create new revenue funnels beyond advertising, but it depends on partner integrations, merchant adoption and consumer trust. Competitors are already forming payment and logistics partnerships, so speed of deployment and developer ecosystem traction will matter.
Comparison & Data
| Period | CapEx (reported or projected) |
|---|---|
| 2025 | $72 billion |
| 2026 (guidance) | $115–135 billion |
| Through 2028 (reported projection) | ~$600 billion (reported earlier) |
The table captures the scale shift Meta outlined around its infrastructure program. The firm attributes the 2026 increase to support for Meta Superintelligence Labs and core operations. Analysts will watch whether capital intensity leads to proportional gains in model capability, user engagement or new commerce revenue streams over the next two to three years.
Reactions & Quotes
In 2025, we rebuilt the foundations of our AI program. Over the coming months, we’re going to start shipping our new models and products
Mark Zuckerberg, Meta (investor call, January 28, 2026)
Increased investment to support our Meta Superintelligence Labs efforts and core business
Meta filing (official earnings release)
Meta acquired Manus in December and plans to both operate the service and weave its agent tech into Meta products
TechCrunch report (industry media)
These statements show the company is signaling both short-term product expectations and long-term infrastructure commitments. Outside observers noted the comments reaffirm prior messages about building product-ready models while shifting emphasis toward commercial applications like shopping. Investors have responded by asking for clearer metrics tying AI work to revenue growth.
Unconfirmed
- Exact timelines and product definitions for the models and agentic shopping tools remain unspecified and were not provided on the call.
- How Manus technology will be integrated into consumer products versus sold as a standalone service is not yet detailed.
- Whether Meta’s claimed advantage from personal context will translate into higher conversion rates or new revenue streams has not been independently verified.
- The reportedly projected $600 billion infrastructure figure through 2028 was referenced earlier and remains an aggregate estimate without an updated breakdown from Meta.
Bottom Line
Meta has signaled an aggressive push to move rebuilt AI capabilities into consumer-facing products in 2026, with agentic commerce positioned as a priority use case. The company is backing that ambition with materially larger infrastructure spending and the Manus acquisition, but key execution details and timelines are still vague.
For regulators, partners and investors, the coming months will be a test of whether Meta can turn heavy capital investment and richer personal context into trustworthy, monetizable products. Observers should watch product launches, partner integrations and any disclosures about data controls closely for evidence the strategy is translating into sustainable business outcomes.