The conventional wisdom about AI goes like this: compute is scarce, it is expensive, it compounds, and therefore whoever controls the most of it will dominate the AI era. It's a compelling thesis. But it's wrong, or at least, profoundly incomplete.
What Wix actually is
Wix is not a compute company. We have never pretended to be. We are something more specific, and in the context of this moment, with a very different value proposition: we are the platform through which AI reaches the real economy. We are also something that is easy to underestimate if you look at only one dimension of what we do.
One website. Every part of your golf business: Bookings, memberships, events, locations, and a pro shop. Built with Wix Harmony.
One Wix dashboard. Every tool to run your business: payments, invoices, SEO & GEO, analytics, taxes, security, accessibility, automations and more.
Base44
Base44, our AI-native application builder represents a new frontier: the ability for anyone to build sophisticated, functional business applications without writing a single line of code. But Base44 and Wix are not competing products. They are complementary layers of the same platform. A Wix user running a growing business doesn't just need a website. They need a booking system, an inventory manager, a customer CRM, a marketing automation tool, and a payments dashboard. They need, in other words, ten different applications and increasingly, all of those will be built, customized, and managed within the Wix operating system.
Two custom Base44 apps built to support the business: Fairway Exchange, a golf gear marketplace, and Fairway Coach Hub, an AI-powered platform for swing analysis, coaching feedback, and student progress.
Two operational Base44 apps for Fairway Club: LiveMap, an interactive command center for locations and bays, and Leagueboard, a scoring and leaderboard app for leagues, tournaments, and group challenges.
Wix as the mechanism
There are hundreds of millions of small businesses, entrepreneurs, creators, and independent professionals in the world. A florist. A yoga instructor. A boutique hotel owner. A freelance photographer. A family restaurant. These are the people who constitute the long tail of the global economy generating a disproportionate share of employment, community, and human meaning, even if they rarely make the front pages of financial newspapers.
For most of the history of the internet, these people were promised access to the same tools as large enterprises and then handed a watered-down version that required technical expertise they didn't have, budgets they couldn't afford, and time they didn't have to spare.
AI changes that equation entirely. And Wix is the mechanism through which it changes.
When a small business owner uses Wix today, they aren't just getting a website builder. They are getting an AI system that understands their business, their customers, their industry, and their goals and actively works to help them succeed. AI-generated websites that would have required a designer. AI-written content that would have required a copywriter. AI-driven SEO recommendations that would have required a consultant. AI-powered eCommerce insights that would have required a data analyst.
Each of these capabilities, even five years ago, cost thousands of dollars a month and required specialized expertise. Today, they are inside Wix, accessible to anyone, everywhere, regardless of budget or technical background.
That is not a marginal product improvement. That is a civilizational shift in who gets access to the latest technology and economic tools. And Wix is at the center of it with integrated and functioning AI in the platform now.
The moat that matters
Here is the part of the compute narrative that I think gets systematically under appreciated: in a world where everyone has access to the same underlying AI infrastructure, the competitive advantage moves decisively to whoever has the best context. Let's be direct about the competitive landscape.
The "vibe coding" era means that increasingly, anyone can build anything. The question is no longer ‘Can you build it?’ It is ‘Do you understand the user well enough to build it right?’
That is where Wix's moat lives.
Generic AI knows a great deal about the world in the abstract. It has read the internet. It understands language, images, code. But it does not know what makes a small business's homepage convert. It does not know which product descriptions drive purchases in which categories. It does not know the seasonal patterns of a wedding photography business, or what pricing structures work for online tutors, or how a restaurant's social media presence correlates with reservation volume.
Wix knows all of this across hundreds of millions of sites, billions of user interactions, and years of real business outcomes.
This is the data moat. And in the AI era, data moats are the real oil wells.
Think of it this way: the same underlying AI model, fine-tuned on Wix's proprietary dataset of SMB behavior, produces recommendations, designs, and insights that a general-purpose model simply cannot replicate. The compute is the same. The intelligence is not. Because intelligence, at the application layer, is not just about processing power, it is about the richness and relevance of the information those processors have learned from.
Even if the world’s biggest AI companies decided tomorrow to build a website platform, they would face a formidable challenge: not the technology, but the accumulated understanding. Understanding of what a hair salon owner in São Paulo needs differently from a freelance consultant in Singapore. Understanding of what drives a first-time visitor to book an appointment versus bounce. Understanding is built from serving hundreds of millions of them, for decades, at every stage of their growth.
That understanding is not a feature. It is the product. And it cannot be downloaded.
But there is a second dimension to this moat that deserves equal attention, and it is one that the "just build it with AI" narrative consistently glosses over: the hidden complexity of running software that real businesses depend on.
Ask yourself this: what does it actually cost to build a fully functional booking system from scratch, even with the best AI tools available today? The initial build might look deceptively simple. A few prompts, a working prototype, a demo that impresses in a meeting. But a system with real depth, one that a business can actually rely on, carries an enormous ongoing burden that rarely shows up in the demo.
Bugs surface in production. Browsers update and break things. Security vulnerabilities emerge that require immediate patches. Privacy regulations change, accessibility standards, payment compliance and so on. Someone has to keep up with all of it. And critically, how do you even know if the AI-generated code got the compliance right? You can ask the LLM to handle privacy regulations. But can you be certain it did or that it will maintain it on an ongoing basis as compliance regulations change?
Consider a booking system for a business with multiple staff members, complex scheduling rules, and real customers making real appointments. One bug, a double-booking, a missed notification, a payment that doesn't process and you've lost revenue, lost trust, and potentially lost customers permanently. The stakes are not hypothetical. They are existential for a small business.
A large company absorbs these costs with dedicated engineering teams, security departments, and legal counsel. A small business like a yoga studio, a hair salon, a photography business has none of that. They cannot staff a security team. They cannot audit their own code compliance. They cannot be on call at 2am when a scheduling conflict corrupts their weekend bookings.
And yet, the problem doesn't go away just because they can't afford to solve it.
This is precisely what Wix has been solving quietly, continuously, at massive scale for 2 decades. Every security patch. Every browser compatibility update. Every regulatory change. Every infrastructure upgrade. Handled, invisibly, so that the business owner never has to think about it. The question for any small business evaluating a DIY AI-built alternative is not "what does it cost to build?" It is "what does it cost to run and what happens when something goes wrong?”
That question has a very uncomfortable answer when the alternative is a bespoke system assembled with an AI tool by someone who is not a professional engineer, maintained by no one, and insured by nothing.
Wix's value is not just what it builds. It is what it absorbs so its users never have to.
When you own your model, you own your improvement curve
Being an AI-native, for us, means something more than just how we use AI. It means we build it too. Wix Harmony, our AI-powered website creation experience, now runs on a model we trained ourselves, purpose-built for the specific task of creating websites, optimized on two decades of real user behavior, and continuously improved on our own terms. It is faster, more accurate, and structurally less expensive than relying on external providers. The cost savings alone represent a meaningful structural shift in how we operate.
But that is only half the story. When Base44 joined Wix in 2025, something important happened beyond the product combination. The same data science team that built and trained the Wix Harmony model turned its attention to Base44's platform and together, they have made Base44 the first app-creation platform in the world to launch its own proprietary large language model “Base 1”. This wasn't Base44 building in isolation. It was Base44 gaining immediate access to years of accumulated expertise, proven training infrastructure, and hard-won institutional knowledge the moment they became part of Wix. That acceleration matters. It is what a real platform advantage looks like in practice.
The Base44 model was trained on tens of millions of real interactions from the platform, the kind of dense, domain-specific signal that external models simply don't have access to. The result is a production-grade AI capable of managing complex, multi-turn conversations, generating and running code, handling backend operations, and operating as a true development agent. It is not a prototype. It is live, and it is serving users today.
There is a clear logic to why both teams chose to go down this path. When you own your model, you own your improvement curve. You are not waiting on a third party's release cycle to fix what matters to your users. You act on feedback daily, not quarterly. You optimize for your specific domain rather than accepting a general-purpose compromise. And you control one of the largest cost lines in any AI-native business, compute and inference, which translates directly into healthier margins and a more sustainable path to scale.
Building at this level is genuinely hard. It demands rare expertise spanning machine learning research, large-scale systems engineering, and advanced mathematical optimization. It requires the kind of depth that takes years to develop and cannot simply be hired overnight. Most companies never attempt it. We did. And the early results validate the conviction.
The bottom line
The AI era will produce infrastructure giants. It will also produce application-layer companies that translate that infrastructure into real human value at massive scale. History tells us, consistently and clearly, that both can be enormously successful and that the application layer often ends up generating more enduring value than the infrastructure beneath it.
Wix is not in the compute race burning valuable capital on infrastructure. We are in the economic empowerment business. We are taking the most powerful technological capabilities in history and putting them in the hands of the people that need them.
We have hundreds of millions of users, we have cash, and we have a product story that once you see it demonstrated concretely is genuinely hard to replicate. Not because of any single feature, but because of the accumulated depth of understanding that sits behind every feature, understanding of users, of businesses, and of what it actually takes to succeed online.
The infrastructure will be built. The compute will scale. The AI will improve. And when it does, the hundreds of millions of small businesses who need a partner to help them navigate all of it will know exactly where to turn.
