For nearly two decades, Meta has operated one of the most profitable advertising businesses in the world, and it has done almost nothing else. The company originally known as Facebook reported that roughly 98% of its $56.3 billion in first quarter 2026 revenue came from selling digital ads. That extraordinary concentration makes it one of the most single-revenue-dependent major tech companies in existence.
Now, CEO Mark Zuckerberg is making his most ambitious push yet to change that. The company announced this week that it will begin testing subscription services for its ChatGPT style Meta AI app and website, rolling out first in Singapore, Guatemala and Bolivia. The plans are priced at $7.99 and $19.99 per month, depending on features. At the same time, premium subscription tiers for Instagram, Facebook and WhatsApp went live, alongside upgraded versions of its brand-verification service aimed at businesses.
At Meta’s annual shareholder meeting this week, Zuckerberg also said that entering the cloud computing market is something the company is actively considering, a move that would place it in direct competition with Amazon Web Services, Microsoft Azure and Google Cloud.
The 5 ventures that did not work out
The challenge is that Meta has tried to build outside advertising before, repeatedly, and the results have not been encouraging.
Portal (2018): The video calling home device launched with considerable fanfare but found little consumer enthusiasm. Meta pulled it from the market in 2022, four years after its debut.
Oculus and virtual reality (2014): Meta acquired VR hardware startup Oculus for $2 billion over a decade ago. Despite years of investment and rebranding its entire corporate identity around the metaverse concept, its Reality Labs division has piled up more than $80 billion in operating losses since late 2020, with no breakout product to show for it.
Libra cryptocurrency (2019): Zuckerberg unveiled an ambitious cryptocurrency initiative called Libra, which drew swift and intense regulatory backlash from governments around the world. The last remnants of the project were shut down in 2022.
Workplace (2016): Facebook launched a business focused team communication tool called Workplace to compete in the enterprise software space. The company announced in 2024 that it would discontinue the service entirely.
Metaverse hardware (ongoing pivot): After betting heavily on virtual reality, Reality Labs has quietly shifted much of its focus toward AI-powered smart glasses, developed in partnership with EssilorLuxottica under the Ray-Ban Meta brand. That product has been one of the company’s rare hardware bright spots.
Why analysts are divided on what comes next
Despite the long list of setbacks, some on Wall Street are cautiously optimistic about the AI subscription push. Analysts at Wolfe Research projected this week that subscription revenue could contribute as much as $3 billion to Meta’s total for 2027, potentially growing to $16 billion by 2030. Meta’s stock climbed nearly 4% on Wednesday after the subscription announcements were made public.
Those figures, while promising in isolation, would still represent a relatively small slice of a company that generates more than $200 billion in annual revenue. The core ad business remains dominant, and analysts note that Meta’s very success in advertising may work against it when trying to cultivate something new.
Observers point out that divisions with enormous margins tend to crowd out internal enthusiasm for smaller, slower growth efforts. Some industry watchers argue that the subscription push could still succeed not by replacing advertising, but by feeding it. If subscription plans attract more creators and power users to Meta’s platforms, the result could be more content, longer engagement sessions and, ultimately, stronger advertising performance.
The cloud challenge is a different beast entirely
The potential cloud computing play presents a far steeper climb. Meta has been pouring money into AI infrastructure at a staggering rate, raising its 2026 capital expenditure guidance to between $125 billion and $145 billion. Zuckerberg has suggested that if the company ends up with excess data center capacity, offering cloud services could be a logical next step.
Industry analysts are skeptical. The established cloud leaders have spent years building out deep technical ecosystems, enterprise relationships and service portfolios that cannot be assembled quickly. Telecom companies that once attempted similar pivots, betting their vast network and data center resources would translate into cloud businesses, were largely unsuccessful across multiple markets and time periods.
Meta declined to comment on its plans.
