Meta Platforms (NASDAQ:META) is heading into one of its most important product events of the year. Meta Connect takes place September 23–24, with AI technologies, AI glasses, and virtual reality officially on the agenda. Meta Connect AI will therefore be a major focus for investors watching the company’s next product cycle. The timing matters. Meta shares have climbed nearly 20% during September, helped by enthusiasm around its Muse AI agent and custom-chip strategy. At the same time, investors are digesting an enormous AI infrastructure commitment. Meta expects $130 billion to $145 billion of capital expenditures in 2026.
That creates a simple question heading into Connect. Can Meta show enough useful AI products to make that spending easier to understand? Muse gives Zuckerberg a consumer product to build around. AI glasses provide a hardware distribution layer. Custom silicon could improve economics over time. A reported next-generation model, codenamed Watermelon, adds another possible catalyst. Connect now brings all four pieces together.
Muse Turns Personal Agents Into A Real Meta Connect AI Test
Muse gives Meta something it badly needs: a visible consumer product tied directly to its AI investment cycle. Meta launched the personal AI agent on September 8. Unlike a standard chatbot, Muse can browse the web and complete multi-step tasks. It can book appointments, fill forms, make purchases, and work across connected apps. Meta built it around a dedicated virtual machine called Muse Secure VM.
That makes Connect an early checkpoint for adoption and product expansion. Meta says Muse is rolling out across the U.S. on mobile and the web. It is also coming to AI glasses. Most basic usage is free, while heavier usage can move into subscription plans. Meta Connect AI could provide a clearer look at how those pieces fit together.
The broader foundation is already showing traction. On Meta’s Q2 call, Zuckerberg said daily interactions with Meta AI had risen 60% after the assistant was rebuilt around Muse Spark. More than one million businesses were also using Meta business agents each week.
That matters because Muse is not operating in isolation. Meta can distribute AI through WhatsApp, Instagram, Facebook, and its glasses. Connect should offer a clearer view of how Meta plans to turn that distribution into regular agent usage.
Meta’s AI Products Are Converging The Spending Case Still Needs Proof
Meta Connect arrives after a nearly 20% September share rally and against a 2026 capital-expenditure plan of $130–145 billion. Muse, AI glasses, model development, and custom silicon are converging into a broader AI platform, but the market debate is increasingly financial. The key question is whether product adoption and monetization can scale quickly enough to make Meta’s expanding infrastructure commitment easier to justify.
Meta can distribute Muse and related AI capabilities across WhatsApp, Instagram, Facebook, glasses, business agents, and advertising, broadening potential monetization paths.
Capital intensity is rising sharply, with 2026 capex expected at $130–145 billion while near-term free cash flow faces pressure.
Watch Meta Connect on September 23–24 for evidence of Muse adoption, useful AI glasses experiences, model progress, and clearer monetization paths.
Meta’s opportunity is increasingly visible across agents, models, glasses, and advertising, but the investment case now hinges on translating that product breadth into durable usage and monetization while capital spending remains exceptionally high.
Watermelon Could Raise The Stakes For Meta’s Model Strategy
The biggest unknown may be the model investors have not officially seen yet. Investor’s Business Daily reports that attention is building around a next-generation Meta model reportedly codenamed Watermelon. The model could become one of the closely watched announcements around Connect. Meta has not officially confirmed its launch, so expectations should remain separated from facts.
What Meta has confirmed is a rapid pace of model development. The company introduced Muse Spark 1.3 on September 2. That followed Spark 1.1 and Spark 1.2 earlier this year. Meta is positioning these models around agentic tasks, coding, multimodal understanding, and tool use.
Zuckerberg also made Meta’s longer-term strategy clear during the Q2 call. The company wants both efficient models for billions of consumer requests and more advanced models for difficult problems. It also plans a mixture of closed and open models.
That makes any Watermelon announcement more than a benchmark story. The bigger issue is whether a stronger model creates better products across Meta’s existing network. Personal agents, business agents, recommendation systems, developer APIs, and glasses can all benefit from the same underlying intelligence. Meta Connect AI could show how quickly those model improvements are reaching real products.
Connect could therefore provide evidence on how quickly Meta’s model research is moving into products people actually use.
AI Glasses Put Meta’s Hardware Bet In The Spotlight
AI glasses may be the most tangible part of the Connect story. Meta has already confirmed that AI glasses will be a major focus of the September 23–24 event. Its developer platform is also pushing new tools for building hands-free experiences across camera, audio, and display capabilities.
The business is beginning to show financial traction too. Reality Labs generated $431 million of Q2 revenue, up 16% year over year. Meta said strong growth in AI glasses helped offset weaker Quest headset sales. Zuckerberg also said early sales of newer glasses exceeded company expectations. He specifically promised more information about the lineup at Connect. Meta Connect AI will therefore put the relationship between Meta’s hardware and its personal-agent strategy directly in focus.
The strategic connection to Muse is important. Meta says Muse will eventually arrive on its AI glasses. That could turn glasses from a camera-and-audio accessory into a physical interface for an always-available personal agent.
There are risks alongside that opportunity. French prosecutors and regulators recently increased scrutiny of smart glasses following privacy and recording concerns. Similar questions are emerging elsewhere.
So Connect needs to show more than new hardware. The useful question is whether Meta can make glasses practical enough for everyday AI use while handling privacy concerns that come with wearable cameras.
Custom Silicon & $145 Billion Capex Raise The Bar
The product announcements matter because Meta is spending extraordinary amounts to support them. The company expects 2026 capital expenditures of $130 billion to $145 billion. Q2 alone included $31.1 billion of capital expenditures, while free cash flow fell to $784 million.
Meta is trying to improve those economics through custom hardware. It plans to develop and deploy four new generations of MTIA chips within two years. Hundreds of thousands of MTIA chips are already deployed for inference workloads. MTIA 300 is in production, while MTIA 400, 450, and 500 are expected to support generative-AI inference through 2027.
Meta is also finding outside capital for parts of the buildout. In July, it announced a BlackRock venture for a 1-gigawatt data center campus in El Paso, Texas. The facility is expected to begin adding capacity in 2028.
This gives Connect a financial dimension beyond new gadgets. Meta needs enough high-value AI workloads to keep enormous amounts of compute productive. Advertising already provides one path. Advantage+ exceeded a $75 billion annual revenue run rate in Q2, while more than nine million small businesses used Meta AI creative tools. Meta Connect AI could help investors judge whether newer products can broaden those monetization paths.
New agents, glasses, APIs, and enterprise services would broaden those paths further.
Final Thoughts
Meta Connect arrives with unusually high expectations because several parts of the company’s AI strategy are converging at once. Muse has moved personal agents from a future concept into a real consumer product. Meta’s model family is advancing quickly. AI glasses are gaining commercial relevance. Meanwhile, custom silicon and new data centers are supporting a far larger infrastructure footprint.
The valuation adds another layer to the setup. Meta currently trades at 7.71x LTM enterprise value to revenue, 15.71x LTM EV/EBITDA, 19.71x LTM EV/EBIT, and 25.71x LTM diluted earnings. Its LTM price-to-sales multiple stands at 7.62x. Those figures have expanded from June levels, when LTM EV/EBITDA was 13.13x and LTM P/E was 20.49x.
In other words, the market is now paying more for Meta’s trailing earnings and operating performance than it was only a few months ago. At the same time, heavy AI investment is putting pressure on near-term free cash flow.
That does not make Connect a simple pass-or-fail event. It does make the details important. Investors can watch whether Meta shows stronger consumer adoption, better AI models, useful glasses experiences, and clearer monetization paths. September 23 should provide another important data point on whether Meta’s AI products are developing quickly enough to match the scale of its spending.
Disclaimer: We do not hold any positions in the above stock(s). Read our full disclaimer here.
