This week the New York Times reported on Meta’s intended entry into the prediction markets business. Interestingly, many in the media efforted to point out that Meta, and CEO, Mark Zuckerberg, were once again jumping in late to a new high growth interactive social media play, similarly to how it reacted “late” when it launched Reels after yielding valuable user engagement time to TikTok. Another element of this week’s reporting on Meta’s announcement focused on how Meta would monetize its new Arena app given that the stated business model, at least to start, would operate on a free, gamified play-money model, in stark contrast to the two dominant, for profit predictive market platforms Polymarket and Kalshi.
This latter question (or criticism) is a fair one indeed. However, with a little creativity, one can imagine that the move is highly strategic for a reason that goes well beyond user engagement, data mining and the attendant advertising revenue. In a world racing toward agentic commerce, Meta (along with Google and its soon to be released search agents) may be positioning itself to front run the largest LLM concerns (OpenAI, Anthropic, Grok) and card networks (Mastercard, VISA) who are all in a full-blown, knock-down drag-out competition to provide the agents of choice for the dawning agentic commerce era.
Ostensibly, Meta entering the prediction market space to keep existing users in its ecosystem, as well as attract a whole slew of new ones, is a no brainer.
Prediction markets are tracking to be the most explosive, high growth, social engagement platforms since the launch of TikTok. The global prediction market industry grew 4X in 2025, hitting $63.5 billion in total notional trading volume, up from roughly $15.8 billion in 2024. From 2022 to 2025, the sector grew 127X (Trade the Outcome). Bernstein brokerage projects a compound annual growth rate of roughly 80% between 2025 and 2030, with prediction market trading volume reaching $1 Trillion annually by the start of the next decade, driven by regulatory clarity, blockchain tokenization, and institutional interest shifting toward economics and macro contracts.
Additionally, the nature of consumer interaction with prediction markets exploits the same driver of user engagement that Meta has been employing across its platforms since its inception – addiction. Perhaps an extremely cynical take, but in truth, across all its platforms – Facebook, Reels, Instagram – Meta’s business model trades and depends on human physiology – the brain’s reward and pleasure system fueled by dopamine release. When someone gets a notification – a like or mention – the brain receives a rush of dopamine sent along reward pathways, causing a feeling of pleasure. What was once merely speculation has now been borne out in scientific research. Furthermore, Harvard found that the mere act of self-disclosure on social networking sites (posting about oneself and sharing with others) lights up the same brain region that is activated when taking an addictive substance.
Most social media platforms – Meta being the “OG” in this category and the largest across all platforms totaling almost 4 billion users – exploit what’s known as this reward prediction error system: unexpected likes and notifications maximize dopamine release. So not only does Meta’s entry into the prediction market space make simple business sense – for maintaining and expanding its user base. It’s also adding a new business unit whose core activity exploits the same physiological mechanism that makes all of its other platforms so addictive. Gambling. Gambling, whether in a sports book or through a prediction market, avails itself of the same neurochemical process.
Now, this could easily be an op-ed on business ethics but that’s not the intention here. The prima facie business case for Meta’s entry into prediction markets is pretty straight forward. What is of import here is that the above belies a far shrewder long-term strategy by Meta to become the agent provider of choice in agentic commerce.
The foundations of success in agentic commerce are two-fold: 1) capturing consumer intent, and 2) capturing the beginning of the consumer buying journey.
Think about it. By developing its own prediction market app, Meta will be adding a slew of new informational signals from consumers in a market segment that’s expected to be a trillion dollars in just a few years. Prediction markets offer insights not just into what consumers are interested in, but also how they feel about those same things, and most importantly, how strongly they feel about those things as quantified by what side of a given prediction market a consumer puts his/her money (play money) behind. This information is extremely valuable to any company attempting to capture consumer intent.
Now add to this new data trove of intent the ability to more effectively capture the beginning of consumer buying journeys. Prediction markets give Meta yet another product to wield against consumers to keep them captive to Meta’s ecosystem. More consumers, enhanced engagement (it’s addictive, remember), and perhaps most importantly, longer engagement, puts Meta in a very strong position to capture this critical aspect of agentic commerce.
Agentic commerce doesn’t exist in the physical world, only the virtual. And there’s only so much time that consumers can spend online in a single day. Thus, competition for eyeballs (online engagement) is a zero sum game, and Meta’s move into prediction markets gives it another powerful weapon to capture that consumer screen time.
