Mark Zuckerberg has a problem that most CEOs would love to have: his company makes almost too much money from advertising. We’re talking $56 billion in a single quarter, with 98% of that coming from ads. That’s anenviable position, but also a precarious one. The rise of generative AI has everyone wondering whether the way people consume information online is about to fundamentally shift. If users start getting their answers from AI chatbots instead of scrolling through feeds filled with ads, Meta’s golden goose could start laying fewer eggs.
So naturally, Meta is trying to diversify. This week, the company announced it’s testing two paid subscription tiers for its Meta AI assistant, pricing them at $7.99 and $19.99 per month depending on features. These will roll out first in Singapore, Guatemala, and Bolivia, which is an interesting choice of markets. Simultaneously, Meta is launching premium subscription plans across Instagram, Facebook, and WhatsApp, plus a higher-tier verification service aimed at businesses.
The market responded favorably, with Meta’s stock climbing nearly 4% on the news. Analysts at Wolfe Research even projected that these subscriptions could generate up to $3 billion in revenue by 2027, growing to $16 billion by 2030. That’s tiny relative to Meta’s current advertising empire, but in an emerging market, it’s nothing to scoff at.
Here’s what makes this whole situation fascinating, though. Meta has tried this before. More than once. And it hasn’t exactly worked out.
The Long History of Meta’s Failed Diversification Attempts
Let’s look at the track record, because it’s quite the litany. The Portal video-calling device launched in 2018 and was quietly killed off four years later. Nobody really wanted a Facebook-branded smart display for their kitchen counter. Then there’s the Oculus acquisition back in 2014 for $2 billion. That was supposed to be Meta’s gateway to the VR future, but Reality Labs has now accumulated over $80 billion in operating losses since late 2020. The pivot to AI-powered smart glasses, leveraging that surprise hit with Ray-Ban Meta glasses, feels less like strategic foresight and more like scrambling for something that actually works.
Crypto was another swing and a miss. Zuckerberg’s Libra cryptocurrency initiative launched in 2019, faced intense regulatory pushback, and the last remnant of the project shut down in 2022. And let’s not forget Workplace, the business-focused chat product that debuted in 2016 and is now being shut down in 2024. That one never really gained traction against Slack, Microsoft Teams, and the rest.
The pattern is pretty clear. When Meta asks consumers or businesses to pay for something other than ads, the answer has usually been no.
Max Willens, an analyst at Emarketer, put it well: Meta is a victim of its own success. Its core advertising business is so massive that any new venture looks insignificant by comparison. “It can be very hard for a corporate parent to sustain enthusiasm for something that is naturally going to be much smaller, likely forever,” he said. That’s an honest assessment, and it explains why these efforts often feel half-hearted.
Could AI Be Different This Time?
There’s a reasonable argument that AI subscriptions are different from VR headsets or crypto. For one thing, AI is genuinely useful to a lot of people right now. The integration with existing products like Instagram and Facebook could make these subscriptions feel like natural upgrades rather than separate productsnobody asked for. Willens suggested that the real goal might not be building a whole new line of business, but rather using subscriptions to drive more engagement with the ad-supported platforms. Get creators and power users paying, keep them on the apps longer, and maybe that indirect advertising benefit is what actually matters.
But there’s a catch. The subscription prices are modest, and the addressable market for AI assistants is already getting crowded. OpenAI, Google, Anthropic, and everyone else with compute to spare is fighting for the same users. Meta’s advantage is its integration with billions of existing accounts, but that’s also its curse. People already use Facebook and Instagram for free. Asking them to pay for an AI assistant layered on top might be a hard sell.
Then there’s the cloud computing thing. At Meta’s annual shareholder meeting, Zuckerberg said a potential cloud computing business is “definitely on the table.” This would put him directly against Amazon, Microsoft, and Google, a trio that isn’t known for losing market share quietly. The logic would be that Meta’s massive AI infrastructure investments might leave it with excess capacity that could be rented out. But as analysts at Forrester noted, companies like Verizon and CenturyLink tried to leverage their data center resources into cloud businesses and got nowhere. Having capacity and having a competitive cloud stack are very different things.
The Bigger Picture
What we’re watching is a company that’s genuinely uncertain about its future, despite current record-breaking profitability. Meta’s ad business is as strong as it’s ever been, but the questions about AI-driven search and conversational interfaces aren’t going away. The company is betting billions on AI infrastructure, raising its 2026 capital expenditure guidance to between $125 billion and $145 billion. That’s an enormous wager.
Some analysts are bullish. The Wolfe Research team recommends buying the stock, arguing that Meta’s scale and AI investments will let it outgrow the digital advertising market and find new revenue streams. They’re probably right that Meta will remain dominant in ads for the foreseeable future. Whether the subscription play amounts to meaningful diversification or just another footnote in Meta’s long history of failed non-ad ventures is another question entirely.
The uncomfortable truth is that Meta has rarely succeeded at anything except advertising. Every attempt to branch out has stumbled, often because the company struggles to care about anything that doesn’t move the ad revenue needle. AI subscriptions might be the most promising alternative yet, simply because AI is genuinely relevant right now. But “promising” and “successful” are very different words, and Meta’s track record doesn’t inspire confidence that this time will be different.
The subscription push tells us something important about where Meta sees itself in five or ten years: uncertain, defensive, and scrambling to find a second act. Whether it finds one might determine whether this company remains a tech titan or gradually becomes a one-trick pony that eventually gets outmaneuvered by the next wave of innovation.
