Meta's AI Bet: Winning Over 250 Million Small Merchants

Opening

Reader, on March 25th, Mark Zuckerberg sent out a company-wide announcement. He was launching a new company-wide initiative called “Meta Small Business.” He told product managers, designers, and engineers alike: “If you want to work on this, reach out.” Just from that, you can tell this isn’t a small effort.

But there’s one line in Zuckerberg’s internal memo that stands out. In the middle of talking about supporting small businesses, he suddenly drops the word superintelligence. When “small business” and “superintelligence” show up in the same sentence, it tells you Meta isn’t treating this as just another small-business support program. Today I want to walk through both the surface and the substance of this initiative.

Why Meta Is Going All In on Small Business

Let’s start with the numbers. Meta’s 2025 annual revenue was roughly $201 billion (~₩280 trillion). Of that, advertising accounts for 97%.¹​ And globally, more than 250 million small businesses use Facebook, Instagram, and WhatsApp. That figure comes directly from Meta itself. I recently met a subscriber of this newsletter — a friend, actually — who told me I cover too much negative stuff about Meta, which honestly surprised me. I’m a Meta shareholder and, if anything, one of the more bullish people on its outlook.

The things I’ve covered before — Meta’s almost unsettlingly precise ad targeting, its aggressive AI investment — might rub consumers the wrong way, but for advertisers and business owners, they’re genuinely compelling. Here’s where it gets interesting: according to GroupM’s analysis, about 55% of Meta’s ad revenue comes from large advertisers in sectors like tech, retail, and finance. Small businesses vastly outnumber them, but because each one spends so little individually, their overall contribution to revenue is smaller. In short: a huge headcount, but thin wallets.

The core goal of Meta Small Business is to close that gap with AI — to raise ARPU (average revenue per user)²​ for each individual small business’s ad spend. Up to now, the reason small businesses haven’t fully utilized Meta’s ads is simple: creating ad creative, setting targeting, and optimizing budget — doing all three alone is hard. If AI automates these three steps, the barrier to entry drops dramatically.

Meta already runs an AI ad automation tool called Advantage+. As of 2024, more than 4 million advertisers had used it, and Meta’s internal data shows AI-driven campaigns saw an average 22% improvement in ROAS (return on ad spend). By the end of 2026, Meta’s roadmap calls for a system where a merchant just uploads one image and sets a budget, and AI handles everything else — creative production, targeting, and optimization — automatically.

The Personnel Choices Reveal the Strategy

Who leads this project matters too. Meta has put two senior executives front and center.

First is Dina Powell McCormick, Meta’s President and Vice Chairman. She spent 16 years as a partner at Goldman Sachs, and what catches my attention most is that she led Goldman’s “10,000 Small Businesses” program — a flagship economic development initiative that provided management education and access to capital for small businesses across the US. Add in her political network from serving as Deputy National Security Advisor under the Trump administration, and this hire looks like a deliberate move to capture three things at once: government relations, small-business experience, and global finance.

Second is Naomi Gleit, Head of Product. An early Facebook employee, she’s one of the longest-tenured executives at Meta. She’s the one who will actually design and ship the small-business tools from a product standpoint.

Someone who once supported 10,000 small businesses at Goldman Sachs is now in a position to deliver AI tools to 250 million small businesses. That’s a 25,000x difference in scale. This personnel choice alone tells you how strategically Meta views this project.

Another Front in the AI War: Why Small Business?

Let’s step back and look at the bigger picture. Looking at where the major players in the AI war are currently focused helps explain why Meta’s choice is so interesting.

OpenAI and Google are building out “agentic commerce”³​ — a world where AI shops on your behalf. You can already check out Shopify products directly within ChatGPT, or have a Google shopping agent check real-time inventory. Shopify itself is evolving into a “commerce operating system” built around its AI assistant, Sidekick, aiming to automate the entire operation of a store.

But there’s a common weakness in all these approaches: they lack an existing customer relationship. No matter how good OpenAI’s shopping agent gets, it has no way of knowing “who this shop’s regular customers are.” Google has search data, but no social relationship data.

Meta, on the other hand, is starting from a completely different position. Through Facebook Pages, Instagram Business accounts, and WhatsApp Business chats, small businesses and their customers are already communicating on a daily basis. Meta’s daily active people (DAP) stood at 3.58 billion as of December 2025. That means the channel for deploying AI tools is already in place.

While competitors have to tell small businesses “come to our platform,” Meta can say, “we’ll make the tool you’re already using smarter.” That difference matters more than it might seem.

Oz’s Lens

Honestly, I think this announcement is one of the smartest platform strategies of the AI era. From my own experience building go-to-market strategies, the biggest bottleneck for new technology penetrating a market usually isn’t the technology itself — it’s adoption cost. For a small business owner, “we need to adopt AI” really means “we need to learn a new tool,” and that’s an enormous hurdle for someone short on both time and money.

Meta’s approach reduces that hurdle to almost zero. If an owner who’s already running a Facebook Page can use an AI ad tool immediately, with no separate signup or learning curve, adoption rates are bound to surge. And once you start depending on an AI tool, it becomes much harder to leave the platform. This is a classic switching-cost strategy.

That said, there’s something to watch out for. Zuckerberg wrote in his memo that he wants “the prosperity created by superintelligence to be widely shared” — but flip that sentence around, and it could just as easily describe a structure where Meta captures a large share of the value AI creates, sitting in the middle. The moment small businesses can’t run ads without AI, pricing power shifts entirely to Meta. In fact, Meta’s ad prices (CPM) have been steadily rising — up 9% year-over-year in 2025.

Somewhere between “free” and dependency, that’s where the truth probably lies.

Closing

First, Meta Small Business isn’t charity. It’s a growth strategy to lift ARPU across 250 million small businesses using AI.

Second, Meta’s differentiation in this competitive landscape is clear: layering AI on top of an already-established merchant-customer relationship is an asset neither Google nor OpenAI has.

Third, the structure of giving small businesses free (or cheap) AI tools and then driving up their ad spend will continuously test the line between convenience and dependency.

If you run a small business or use Meta ads, I’d recommend checking out the Advantage+ tool first. You’ll get a real sense of which direction these tools are heading. Personally? In Korea, I think the closest equivalents to this are Danggeun Market (a Korean neighborhood marketplace app) and Naver Place.

References & Further Reading

• Axios, “Exclusive: Zuckerberg launches Meta Small Business”, March 25, 2026. — The exclusive report that sparked today’s issue.
• Meta Investor Relations, “Meta Reports Fourth Quarter and Full Year 2025 Results”, January 29, 2026. — The source for key figures like $201 billion in 2025 annual revenue and 3.58 billion DAP.
• Meta, “Dina Powell McCormick Joins Meta as President and Vice Chairman”, January 12, 2026. — Confirms Dina Powell McCormick’s background and role.
• CNBC, “OpenAI’s first crack at online shopping stumbled. It’s preparing for the next wave”, March 20, 2026. — A well-organized rundown of the agentic commerce race among OpenAI, Google, and Shopify.
• Axios, “Zuckerberg says AI will dramatically change how Meta employees work in 2026”, January 29, 2026. — Details how AI is being used internally at Meta, including a 30% engineering productivity boost.
• Marketing Dive, “Meta streamlines AI use for brands with new business agent, creative tools”, October 2, 2025. — Useful for understanding Meta’s AI tool strategy for SMBs, including Business AI and the expansion of Advantage+.

The author, Kwangseob Ahn, is a professor of business administration at Sejong University and lead consultant at OBF (Oswarld Boutique Consulting Firm). He teaches statistics and data analysis — business data management and business analytics — while leading GTM and AI strategy consulting in the field, designing the seam between technology and business. He has published academic research on a memory architecture for AI dialogue systems (HEMA) and runs Daily Arxiv, a daily curation of global AI papers. He holds a master’s from Korea University’s Graduate School of Technology Management and a KMBA. He is the author of Homo Brainless: The People Who Outsource Their Thinking.

Footnotes

1. 97% advertising revenue share: This means Meta’s revenue structure is almost entirely dependent on advertising. The remaining 3% comes from Reality Labs (VR headsets, etc.). ↩

2. ARPU (Average Revenue Per User): The average revenue a company earns per user. When this number rises, it means the company can earn more money from the same number of customers. ↩

3. Agentic Commerce: A new form of shopping in which an AI agent searches for, compares, and even pays for products on the user’s behalf. Think of it as “having AI do your shopping for you.” ↩