Author|Hua Lin Dance King

Editor| Jing Yu

Once upon a time, the model where users used internet products for free in exchange for watching 'ads' has finally come to an end in the AI era.

On May 27, Meta officially announced the global launch of its paid subscription plans. Instagram Plus is priced at $3.99 per month, Facebook Plus at $3.99 per month, and WhatsApp Plus at $2.99 per month.

At the same time, Meta is also testing a premium AI plan for heavy AI users (with two tiers: $7.99 and $19.99), as well as a professional package for creators ($49.99), all integrated under the 'Meta One' brand.

This is not a simple product update; it is a key card played by Meta in a larger strategic transformation. Behind this may signify the end of the 'free internet era' we were once familiar with.

01 Even the Landlord Is Running Out of Surplus

To understand the significance of today's subscription plans, we need to turn back the clock to a month ago.

On May 20, Meta initiated a massive round of layoffs that shocked Silicon Valley, cutting about 8,000 employees while freezing 6,000 open positions. At the same time, the company announced it would invest $125 billion to $145 billion in AI infrastructure. Laying off people is to concentrate money on AI.

Meta's CTO subsequently stated on May 25 that the company would use AI tools to implement a 'large-scale transformation' of its workforce. 7,000 employees were transferred to AI-related roles. The company's focus is visibly shifting toward AI.

This brings up a core contradiction: How can investors be convinced after pouring so much money into AI?

For Wall Street, the most concerning issue is not how much Meta is spending, but what kind of predictable returns this money can bring. Google has Cloud, Microsoft has Azure, Amazon has AWS—their AI investments can be measured by subscriptions and API calls. But what does Meta rely on?

Ad revenue fluctuates with the market and is not stable enough; the open-source large model Llama has enhanced technical reputation but does not directly generate revenue; AI glasses and AR devices are still in their early stages.

Thus, subscriptions entered Meta's field of vision.

This timing is not coincidental.

02 How to Convince Users to 'Pay'?

Meta's products have long operated under an implicit contract—you use my platform, and I sell your attention to advertisers. This logic has worked well for twenty years, with Facebook boasting over 3 billion monthly active users, Instagram over 2 billion, and WhatsApp users spanning the globe.

However, cracks are beginning to appear in this wall.

European regulators have been the biggest catalyst. To comply with the EU's data privacy regulations, Meta began testing an 'ad-free subscription' option in Europe as early as 2023, offering users a paid choice to avoid data tracking. The globally launched subscription plan is, in a way, an extension and deepening of this European experiment.

This time, however, Meta is playing by a different logic—not 'pay to remove ads,' but 'pay to unlock more.'

The core selling points of Instagram Plus include anonymous browsing of Stories, detailed replay data analysis, extended disappearing post duration, and custom themes and reactions. WhatsApp Plus focuses on enhanced privacy and expanded features.

The common characteristic of these features is that the free version is sufficient, but the paid version allows you to: 'have a little more control.'

From a product design perspective, this is more challenging than 'paying to remove ads.' Removing ads addresses a clear user pain point with a straightforward functional trade-off; but 'unlocking more' requires Meta to prove that these 'more' are truly worth the price.

Forrester's survey data pours cold water on this idea: 70% of respondents indicated they would 'definitely' or 'likely' not pay for a Meta subscription. Reasons vary—some feel the current free version is sufficient, others harbor long-standing resentment toward Meta's privacy practices, and some bluntly ask, 'Why should I give you more money?'

This resistance is real, but it is not insurmountable.

Snapchat+ serves as the best reference point. When Snap introduced its paid subscription in 2022, the general consensus was pessimistic, believing users wouldn't pay for a messaging app. However, to date, Snapchat+ has surpassed 15 million paying users. The key isn't 'whether users are willing to pay,' but whether the product delivers enough concrete, direct value.

X (formerly Twitter), Telegram, and Snap are all increasing their bets on subscriptions. Paid subscriptions are becoming an increasingly important part of social platforms' revenue mix.

03 AI Features: The Real Battlefield for Monetization

If Instagram Plus and WhatsApp Plus are merely test runs, then AI subscriptions are the core of Meta's ambition in this layout.

Meta announced it will test two tiers of AI subscription plans, priced at $7.99 and $19.99, with the main difference being the usage limits for advanced reasoning and 'thinking mode.' The basic version of Meta AI will remain free, but for faster response speeds, stronger reasoning capabilities, and higher usage limits, users will need to pay to unlock them.

This design logic is almost identical to the freemium models of OpenAI and Anthropic.

The difference lies in scale.

OpenAI's user base is in the hundreds of millions, while Meta's monthly active users number in the billions. Even with a conversion rate of just 1%, the numbers would be vastly different. An analyst from Seeking Alpha crunched the numbers: based on WhatsApp Plus's $2.99 price and an estimated 1.5% conversion rate, this single product alone could generate approximately $2 billion in annual revenue, with a gross margin close to 100%.

What excites investors even more is the predictability of such revenue. Ad revenue fluctuates with macroeconomic conditions and privacy regulations, but subscription revenue is predictable, recurring income. This is precisely what Meta has struggled to articulate regarding its AI investments—now, it has a story to tell investors.

On the day of the announcement, Meta's stock price rose nearly 3%, a straightforward and clear market response. Evercore ISI analyst Mark Mahaney gave a buy rating, expressing particular optimism about WhatsApp's long-term monetization potential. He projects that by 2030, WhatsApp alone could generate $40 billion in annual revenue.

This is, of course, the most optimistic scenario, and the real path is fraught with variables. But it at least shows that this road is not a fantasy; it is a business logic supported by numbers.

04 The 'Free Era' Has Ended

Remember that phrase long circulated in the tech world—'If the product is free, you are the product.'

Meta's business model has always been the quintessential footnote to this statement. Users exchange attention and data for free services, and Meta sells this data to advertisers. This logic ran fast during the smartphone era. The rise of Facebook, the explosion of Instagram, and the global expansion of WhatsApp were all built on this foundation.

But the definition of 'free' is quietly changing.

On one hand, heightened privacy awareness is making more and more users wary of the exchange of 'data for services.' EU regulations like GDPR and DMA are tightening step by step, costing Meta billions of dollars annually. On the other hand, competition in the AI era has made the cost of 'free' unprecedentedly high—training an advanced model and maintaining the computing power for an AI assistant is far more expensive than displaying a few ads.

Mark Zuckerberg needs a way to make users who derive real value from Meta AI pay directly for that value.

This is not a betrayal of the original intent of the 'free internet,' but an acknowledgment of a reality—in the AI era, 'free' requires someone else to foot the bill.

The payer can be advertisers, or it can be the users themselves. Meta now wants both to coexist.

The success or failure of the subscription plans ultimately depends on the answer to one question: Are features like anonymous Story browsing, advanced AI reasoning, and creator data analysis truly worth the few dollars you pay each month?

Twenty years ago, when Zuckerberg typed the first line of code in his Harvard dorm room, he probably didn't imagine charging users one day.

But that story is from twenty years ago.