Mark Zuckerberg, CEO of Meta Platforms, Inc., during the MetaConnect event on Wednesday, September 17, 2025 in Menlo Park, California, USA.
David Paul Morris | Bloomberg | Getty Images
meta is once again proving that you can make money doing more than just selling ads. It’s not a strategy that has worked in the past, but CEO Mark Zuckerberg is betting that artificial intelligence will yield better results.
The company announced this week that it will begin testing two subscription services for meta-AI apps and websites like ChatGPT. These paid services will be available first in Singapore, Guatemala and Bolivia and coincide with the official release of premium subscription plans for Instagram, Facebook and WhatsApp, as well as higher versions of verified subscription services designed to help businesses protect their brands.
Also this week, Mr. Zuckerberg said at Meta’s annual shareholder meeting that the possibility of a cloud computing business is “definitely on the table” and could eventually compete with Meta. Amazon, microsoft and google with cloud infrastructure.
Ever since Mr. Zuckerberg’s company, formerly known as Facebook, started selling digital ads nearly 20 years ago, its only real business has been advertising. Meta said in its earnings call last month that nearly 98% of its $56.3 billion in first-quarter sales came from advertising. This is one of the most profitable markets in the tech industry, it’s a very lucrative market, and it’s one of the meta companies that have long dominated in the US, along with Google.
Meta just posted its highest quarterly growth rate since 2021, showing that the online advertising market is now as strong as ever. But the rapid emergence of AI raises the question of what will happen when users turn to new interfaces for information and no longer spend so much time on screens that are constantly bombarded with links.
And when Meta asked consumers and businesses to open their wallets for anything other than advertising, the answer was usually no.
The Portal video calling device, which debuted in 2018, ultimately failed and was pulled from the market after four years. Meta acquired virtual reality hardware startup Oculus in 2014 for $2 billion, but the company’s Reality Labs division has yet to produce a breakthrough VR headset, resulting in more than $80 billion in operating losses since late 2020.
Reality Labs has shifted resources from VR to the more promising AI-powered smart glasses. The company is looking to capitalize on the surprising success of its Ray-Ban Meta glasses by partnering with EssilorLuxottica, Meta’s notable exception in hardware.
Next is cryptocurrency. Zuckerberg entered the field in 2019 with a proposal for a cryptocurrency initiative called Libra. The effort came under intense regulatory scrutiny, resulting in the last remaining cryptocurrency project being shut down in 2022.
Selling social media services to businesses has also been difficult. Facebook debuted its business-focused Workplace chat product in 2016, but ultimately announced it would shut down the service in 2024.
“New source of income”
Despite the uphill battle, some analysts are optimistic that Meta can find new avenues to revenue in the AI space. Meta AI subscriptions announced this week will cost between $7.99 and $19.99 per month, depending on specific features. Shares rose nearly 4% on Wednesday when the news broke.
Analysts at Wolf Research said in a note that day that subscriptions could contribute up to $3 billion to Meta’s total revenue in 2027, which could rise to $16 billion by 2030. That’s still a small number for a company with more than $200 billion in annual revenue, but it represents a huge opportunity for Meta in a fast-growing market.
Analysts at Wolfe said they recommend buying the stock “based on our long-term view that the company’s size, AI investments, category leadership position, and product catalysts will enable META to grow beyond the digital advertising market, scale, and generate new revenue streams.”
Mehta declined to comment.
EMarketer analyst Max Willens said Meta was a victim of its own success in online advertising. Because the company’s core business dwarfs any other efforts, “it’s very difficult for a parent company to maintain enthusiasm for something that is naturally going to be much smaller and probably last forever,” he said.
“Meta’s past efforts have all had different circumstances,” Willens said. “But I think it’s hard enough to be successful in one business, let alone two.”
Willens said the subscription push could be successful if viewed as a supplement to online advertising, rather than an entirely new line of business. Some of the new services are tailored toward creators and power users, so the end goal could be to get more content into Meta’s apps and services and keep people engaged with Facebook and Instagram longer, he said.
Selling technology to businesses can be an even bigger challenge.
Shashi Bellamkonda, research director at Info-Tech Research Group, said Mehta “has a lot of work to do to build an enterprise business from the ground up, as the company is solely focused on direct-to-consumer sales.”
Bellamkonda characterized Meta’s workplace offering as appearing “half-baked, with too much focus on the social aspect of Facebook, which makes money through advertising.”
To successfully compete within the enterprise, particularly in the cloud, Bellamkonda said Meta needs to “strengthen the processes, platforms, technology and, most importantly, the human resources” needed to operate, maintain and sell its products and services. When it comes to customer support, Meta is going in the opposite direction and reducing its headcount through layoffs.
Mr. Zuckerberg had no guarantee that Meta would enter the cloud computing market, which is led in the U.S. by Amazon Web Services, then Microsoft Azure, and then Google. The reason for doing so, he said, would be if the company found that it had excess production capacity after making significant investments in AI infrastructure.
In April, Meta raised its 2026 forecast for AI-related capital investment to $125 billion to $145 billion, from the previous $115 billion to $135 billion.
Forrester analyst Naveen Chhabra said that while current cloud computing leaders have been “winning because they’ve developed huge stacks” over the years, “Meta doesn’t have that, at least not yet.”
Chhabra pointed out that past efforts by companies like Verizon and CenturyLink to launch cloud businesses from their vast data center resources have failed.
“Historical evidence that carriers jump into cloud businesses hoping their capabilities and networks will bring them business has been proven wrong across regions and time,” Chhabra said.
Watch: Meta launches subscriptions for Instagram, Facebook, WhatsApp, with AI plans in the works.
