Meta Says AI Is Accelerating App Launches as New Consumer Products Near Release

Updated July 30, 2026 7:24 pm

Meta says artificial intelligence is now helping it build and launch new standalone apps faster, and the company expects more consumer products to arrive soon. The message came from CEO Mark Zuckerberg on the company’s second-quarter earnings call this week, where he linked AI-driven development tools and recommendation systems to a renewed push for fresh apps outside Meta’s core platforms.

The announcement matters because Meta has spent years trying to create successful new social products beyond Facebook, Instagram and WhatsApp, with limited results. This time, the company argues that large language models are changing the economics of experimentation by speeding up engineering work, improving ranking systems and helping new products find an audience more quickly.

Meta is betting AI can solve an old product problem

Meta’s latest app strategy is less about a single breakout launch and more about increasing the number of bets it can place. Zuckerberg told investors that AI is making it easier for the company to ship software and test ideas, which could let Meta move from occasional experiments to a more continuous stream of consumer releases.

During the call, he pointed to several recently launched products as evidence of the shift. Those include Instagram Instants, Forum, a standalone Groups app, and Seller, a standalone Marketplace app. He also said Meta is planning more ideas and intends to use its recommendation systems to help those products grow once they are live.

The company’s framing is significant. For years, Meta’s biggest challenge has not been invention in the abstract, but retention. It has repeatedly launched new social apps, only to watch many of them struggle to attract enough users to survive. AI, Meta now says, changes both the speed of development and the quality of the systems that support discovery and engagement.

What Zuckerberg said on the earnings call

Zuckerberg used the earnings call to argue that AI is becoming a practical product tool rather than only a research or infrastructure play. He said the company expects it to become “a lot easier” to launch new apps and that Meta plans to keep building new consumer products with help from its recommendation engines.

Zuckerberg said Meta is excited about how AI is helping teams move faster through product development, adding that the company expects shipping new apps to become easier and that it is preparing to release more consumer products soon.

He also said AI is already strengthening the company’s core business by making its apps more relevant and improving outcomes for advertisers and businesses. In other words, Meta is not presenting AI as a side project. It is positioning it as both a growth engine for existing products and a launchpad for future ones.

How AI is changing Meta’s product cycle

AI is changing Meta’s product cycle by reducing the amount of manual work needed to build, evaluate and iterate on app features. Instead of relying entirely on traditional engineering workflows and slow user testing, Meta can now use language models to assist with ranking, quality checks and experimental design.

That speed matters because social products often live or die based on how quickly they can improve the user experience and adjust to behavior patterns. If AI shortens the time between idea and feedback, Meta can test more concepts before rivals or internal priorities shift.

The company’s leadership is also suggesting that AI can help translate existing scale into new product momentum. Meta already has billions of users across Facebook and Instagram, and that distribution gives it a structural advantage when it wants to seed a new app. AI may make that advantage easier to exploit.

Meta has tried this before, and not always successfully

Meta’s renewed push for new apps comes with a long history of failed attempts. The company has repeatedly created internal groups designed to foster experimentation, only to shut down many of the resulting products after they failed to gain traction.

In its earlier Facebook era, the company ran Creative Labs, an internal incubator meant to produce fresh social ideas. That effort led to a series of launches that included Slingshot, Rooms, Paper, Moments and Riff. None became a lasting hit, and the program was eventually wound down in 2015.

Meta tried again in the early 2020s with NPE Team, another internal research-and-development effort focused on experimental apps. That group produced a long list of products, from Bump and Aux to Spark, CatchUp, E.gg, Venue, Hotline, Super, Tuned and BARS. Those, too, were eventually discontinued.

The history matters because Meta is once again arguing that the next wave of apps will be different. This time, it says the difference is not just strategy or branding. It is AI.

Why Threads is central to Meta’s AI argument

Threads is the clearest example Meta can point to when making the case that new apps can still scale inside its ecosystem. The text-based social platform has reached 500 million monthly active users, a milestone that gives Zuckerberg a stronger story than the company has had for many years when discussing standalone consumer launches.

Threads did not grow by accident. Meta used its existing audience across Facebook and Instagram to seed the app with users, then promoted it aggressively across its own services. That distribution strategy remains crucial, but Meta says AI has also contributed meaningfully to the app’s momentum.

The company says it has seen “significant gains” from AI-powered content recommendations, which help surface posts that are more likely to keep users engaged. That matters because recommendation quality is one of the biggest determinants of whether a social product feels useful or overwhelming.

Zuckerberg has repeatedly suggested that Threads could eventually become Meta’s next billion-user app. Whether that happens will depend on continued product development, strong retention and the company’s ability to turn initial attention into durable habit.

How Threads benefits from Meta’s recommendation systems

Threads benefits from Meta’s recommendation systems because those systems help the company better understand what users are likely to find relevant. The more accurate those systems become, the easier it is to keep people scrolling, posting and returning.

Meta CFO Susan Li said on the call that large language models are becoming increasingly useful for ranking and recommendation improvements. She explained that they help existing systems better understand content and generate better training data, while also supporting engineering teams that evaluate quality, detect trends and test ranking changes.

That is a major operational shift. Meta is not just using AI to generate content or automate support. It is using AI to shape the core logic of how content is organized, surfaced and measured across its products.

What exactly is Meta using LLMs for?

Meta is using large language models for both product development and ranking infrastructure. On the development side, the models help teams move faster. On the product side, they help determine what users see and how content is assessed at scale.

Li said every Reel and Feed post on Instagram is now automatically processed through an LLM and analyzed for subject matter and tone. That system gives Meta more structured data about the content flowing through its apps, which can improve recommendation quality and help the company tailor the experience more precisely.

Meta is also working on LLM-native recommendation systems. Those systems are designed to be built around language models from the ground up rather than layered onto older ranking architecture. If they perform well, they could become especially useful for newer apps that do not yet have the historical usage data Meta has accumulated on Instagram and Facebook.

Why Meta’s history makes this strategy riskier than it sounds

Meta’s optimism about AI must be measured against its history of false starts. The company has a well-documented pattern of creating promising internal prototypes, giving them resources and then abandoning them when user growth does not materialize.

That is not unusual in big tech, but Meta’s repeated attempts stand out because they have occurred in such a core area of the business. The company knows social apps better than almost anyone, yet it still has struggled to create successors to its main platforms without leaning on acquisition, replication or ecosystem gravity.

The challenge is partly behavioral. Users already spend most of their social time in a few dominant apps, so convincing them to open another one requires a clear and immediate value proposition. AI may make it easier to build products, but it does not guarantee people will use them.

There is also a branding question. Many of Meta’s previous experiments looked like responses to a trend rather than original categories with long-term staying power. If the new wave of apps feels too derivative, AI will not be enough to rescue them.

How Zuckerberg is reframing Meta’s growth story

Zuckerberg is reframing Meta’s growth story around two complementary ideas: AI makes existing products stronger, and AI makes new products easier to create. That gives investors a broader narrative than the familiar one centered only on ad efficiency, user engagement and Reality Labs spending.

In practical terms, this means Meta is trying to convince markets that it can do more than protect Facebook and Instagram. It wants to show that the company can also keep building new consumer surfaces, even in a saturated social market.

That message is especially important because Meta has faced recurring questions about whether its future lies mainly in AI infrastructure, advertising automation or long-shot bets like the metaverse. By pointing to app launches, Zuckerberg is making the case that consumer software remains central to the company’s identity.

The company also appears eager to make AI feel visible to users rather than only to developers. By connecting the technology to better recommendations, more relevant content and new app experiences, Meta is translating technical progress into product language investors can understand.

How much of this is about competition?

Much of this is about competition because Meta is operating in a market where faster product cycles can create strategic advantages. If rivals can launch AI-native apps quickly, Meta needs a way to match that pace without sacrificing quality or distribution.

At the same time, Meta’s scale means it can compete differently from smaller startups. Rather than building one app from scratch and hoping it catches on, Meta can seed ideas across its massive user base, use AI to optimize the experience and then funnel the winners into broader adoption.

That approach could make Meta more flexible as consumer behavior continues to fragment across formats such as short video, messaging, communities and AI-assisted content discovery.

What investors did and did not focus on

Investors on the earnings call did not spend much time pressing executives on the new apps themselves. Instead, they were more focused on AI capital spending and Meta’s enterprise ambitions, which have become larger themes in the company’s investor conversations.

That silence may be telling. It suggests the market still sees Meta’s AI story primarily through the lens of infrastructure, monetization and cost, rather than through the uncertain upside of experimental consumer software.

Still, Zuckerberg’s comments indicate that Meta sees the consumer side as an important part of the AI story. The company appears to believe that better model performance and better app-building tools can create a feedback loop: more launches, better recommendations, stronger engagement and, ultimately, more durable growth.

Timeline of Meta’s repeated app experiments

Meta’s push into new apps has unfolded in waves. The company has repeatedly tried to create fresh social products, but the latest wave differs because AI is now embedded in both development and ranking.

What happens next?

Meta says the next wave of consumer products is arriving soon, but the company has not yet detailed what those apps will be or how they will differ from earlier experiments. That leaves the market with more expectation than specifics.

The immediate question is whether Meta can turn AI-enabled development into a repeatable consumer formula. One successful app does not erase the company’s past disappointments, but it does give Meta a proof point it has lacked before.

If the company can combine its distribution advantage, recommendation improvements and faster product iteration, it may finally be able to do what it has struggled to do for years: build new standalone apps that matter at scale.

For now, the most important takeaway is that Meta no longer sees AI as only a backend advantage. It is treating AI as the engine of a broader app-launch strategy, one that could reshape how the company experiments with consumer software in the years ahead.