A full teardown of two go-to-market machines pointed at the same job (get a non-technical person from idea to live software): a 2.5-year-old private challenger valued at $6.6B (Lovable) and a 20-year-old public incumbent the market is repricing in real time (Wix). What's actually AI-native, what's an $80M acquisition wearing an AI costume, where the gap between the investor story and the GTM motion is widest, and the three moves I'd hand Lovable before its own moat starts to leak.

On May 28, 2026, Wix told its staff it was cutting about 20% of the company, roughly 1,000 people, and the reason it gave was AI. CEO Avishai Abrahami called the moment "the most significant shift in how companies are built since the invention of modern programming languages in the 1970s."

That was two weeks after Wix missed Q1 earnings and watched its stock fall 27% in a single session. A company that was worth roughly $8.8B last June, when it bought its way into AI, is worth around $2.3–2.6B today. The same AI that Wix sells the street as its growth engine is the AI it cited as the reason it could operate with a thousand fewer people.

Same twelve months, different company.

Lovable went from $100M ARR in July 2025 to $400M by February 2026. It did it with 146 employees and, until late 2025, essentially no sales team at all. Its $6.6B Series B valuation makes a 2.5-year-old startup worth more than the profitable public company that has been building websites since before the iPhone.

The number I keep staring at is the multiple. Wix has ten times Lovable's revenue, a real profit engine ($605M of free cash flow in 2025, 30% of revenue), and the market values its entire business at about 1.3x its recurring revenue. Lovable, with a fifth of the revenue and no profit history anyone has audited, is priced at sixteen.

Most of that gap is the nature of the two assets, not a verdict on either company. Private growth rounds price momentum and optionality; public markets price de-risked cash flow. A company doubling ARR gets a multiple a company growing 13% never will, and a 2.5-year-old in the hottest category in software carries a narrative premium a 20-year-old website builder gave back years ago. Fair enough. But strip the stage and the structure away and one question is left holding up Lovable's number: is the growth durable? That isn't a product question. The model underneath is a commodity anyone can call. It's a go-to-market question, and it's the one this teardown can actually answer.

Lovable's curve points up and to the right, so let me start there and explain why the shape is the motion, not the product.

Six dimensions, loosely graded, plus a composite for how AI-native each company's own go-to-market actually is, not how AI-native its marketing claims to be. (Loose on purpose. The grade is the hook; the qualitative read underneath is the point.)

The grades are the hook. The rest of this teardown is the evidence. The rows that separate them: demand creation (Lovable seeds open source and lets its users' apps do the advertising; Wix runs a $100M+ paid machine and buys Super Bowl spots); GTM org design (Lovable did $400M ARR with 146 people; Wix just cut a thousand); and own-machinery, where both are more modern than you'd guess.

Lovable's defensibility isn't the model. Anyone can call Claude. The defensibility is that the product is the distribution channel, and that took a specific set of decisions to build.

Start with the origin, because the "overnight $400M company" framing hides the lesson. Lovable is Anton Osika's second act. The first was GPT Engineer, an open-source project that hit roughly 50,000 GitHub stars and built a standing audience of hundreds of thousands of developers before Lovable existed as a company. By the time the commercial product launched in November 2024, the launchpad was already crowded. That is why $1M ARR landed eight days after launch. The viral moment wasn't luck. It was a warm list, lit on cue.

The free tier is a billboard network. Every project built on Lovable's free plan is public by default, and every public project carries an "Edit with Lovable" button that forks it into a new user's account in one click. So each free user isn't a cost center waiting to convert. They're a customer-acquisition surface. Lovable now points to 600M+ monthly visits to apps built on the platform. That is 600M monthly brand impressions Lovable does not pay for, because the product earns the reach.

Building in public is the entire demand-gen plan. Osika, on Lenny's Podcast: "For getting awareness we mainly been posting what we've shipped on social media. That's how people know about us." His Head of Growth, Elena Verna (who built PLG at Miro and Amplitude before this), says the quiet part out loud: "Traditional paid performance marketing isn't a meaningful lever for us at Lovable. Performance works when there's existing demand. We're building a new category, so the job isn't capturing demand, it's creating it."

The community is the retention layer. Lovable runs Lovable Launched, a Product-Hunt-style board where the week's best builds win free credits and get amplified on Lovable's own channels. It runs an ambassador program, an affiliate program, a Discord, and hackathons like "She Builds" (3,000 applicants, 200 builders, 48 hours on unlimited credits). None of this is a media buy. All of it manufactures the next cohort of evangelists.

Underneath the mechanics is the worldview that makes them work. Osika's thesis, in the version he gave Sequoia: "Point five percent of the world can even code. So I decided, let's build something for the 99% and not this productivity boost for developers." If you believe the bottleneck on software is talent scarcity, you build a product so simple it floods the market with builders, and the builders become the channel. The GTM is a byproduct of getting the product right. That belief is why the company could cross $400M ARR with 146 people. Look at the efficiency next to a labor-led incumbent and the gap is almost rude.

Wix's motion is excellent. It's also the exact motion AI is built to compress, which is why a great machine is currently being repriced as a liability.

Read it straight off the company's own surfaces. Wix spends more than $100M a year on marketing, $90M in Q1 2026 alone, across roughly ten channels. CMO Omer Shai, an 18-year veteran, runs it on a metric he calls TROI, time to return on investment: "If I'm keeping the ROI at 11 months, I have unlimited budgets." Wix has run six Super Bowl ads over 18 years and returned for Super Bowl LX in February 2026 to launch its AI builder. Shai's logic for the spot is honest about what it is: "The only day that people are focusing on commercials is around the Super Bowl. And the day after, people keep talking about what was the ad of who."

This is a demand-capture machine of real craft. The problem is structural: it is a machine that converts spend into signups, and AI is collapsing the cost of the thing those signups were buying. When a non-technical person can describe a site and get one in ninety seconds, the moat was never the builder. It was everything after.

To Wix's credit, its leadership knows this and says it well. President Nir Zohar gave the single best steelman of the incumbent case on the Q1 2026 call: "AI has made building online simple and anyone can generate a simple good-looking website in minutes. But that's as far as it goes. The real complexity begins the moment you hit published. How do you host it, get found on search engines, run your storefront, secure your customers' data and actually operate the business day to day? These are the hard problems, and we've been solving them for 20 years."

He's right that the hard part is after publish. The question the market is asking is whether owning the hard part is worth what it used to be, when the easy part (the part Wix spent two decades and a Super Bowl budget teaching the world it owned) just went to zero.

Look at how the relationship is monetized once you're in, and you can see why retention shows up in the reviews as resentment. Customers report renewals jumping from $210 to $620, being charged $900 after turning off auto-renew three months early, and, my favorite detail, a 30% price increase "attributed to AI features being added, with no option to opt out." The AI narrative lands on customers too, as a line item they're paying whether they wanted it or not. The most common complaint is the oldest one in the incumbent playbook: "Won't allow you to move to another platform, locks customers and keeps increasing price with every renewal."

This is the section a normal tools roundup never writes, because it means reading earnings calls. It's also where the real insight lives. Four disconnects between the story and the motion.

One: "AI-native" is a story; the architecture was bought. Wix markets Wix Harmony as "first-of-its-kind." The reality is that Wix's entire vibe-coding capability is acquired. In June 2025 it paid $80M for Base44, a six-month-old, eight-person, bootstrapped startup, and added $25M in retention bonuses on top. You don't pay nine figures for an eight-person team you could have built internally. The "AI-native" company had to buy AI-native off the shelf.

Two: the market is pricing zero AI re-rating. Every Wix call now leads with AI driving "improved conversion and monetization." The market's verdict: a 27% single-session drop on the Q1 miss, a multiple compressed to ~1.3x ARR, and a $1.6B Dutch-auction buyback that didn't defend the price. Then the tell: the same AI Wix calls its growth engine became the rationale for cutting 20% of staff. AI can be your growth story or your headcount-reduction story. Telling both in the same quarter is how you teach the market not to pay for either.

Three: the growth is agencies, not AI self-creators. The whole AI narrative is about empowering a new class of non-technical builders, "significantly expanding our TAM," per Abrahami. The segment data says the actual engine is Partners (agencies and web pros): roughly $750M in FY2025, up 23%, while the legacy self-creator business grows around 11%. And the Q1 miss that broke the stock? It centered on a slower-than-hoped Partners quarter. The AI-self-creator story is the marketing; the agency channel is the business, and the business is the part that wobbled.

Four: Base44 may be worth more than all of Wix. Here's the arithmetic that should keep Wix's board up at night. Base44 went from a few million in ARR at acquisition to $100M by early March 2026 and $150M by mid-May, under eighteen months from founding. Price a standalone vibe-coding company at $150M ARR using the private multiples the category trades at (Lovable is at ~16x even after compression), and Base44 alone implies something in the $2–4B range. Wix paid $80M for it a year ago. The entire public parent, Base44 included, trades around $2.4B. The legacy platform's 1.3x multiple is burying its single best AI asset, and investors who want Base44 exposure can't buy it cleanly. (To be precise: that standalone figure is an inference from comparable private multiples, not a market quote. But the gap is wide enough that the direction holds even if you halve the multiple.)

The fair read on Wix isn't "dead company." It's a genuinely profitable business with a real cash engine and a legitimately fast-growing AI asset, narrating a transformation to a market that has decided the legacy motion underneath gets a legacy multiple. The cash is real. The AI is real. The motion the cash came from is the part being repriced.

Lovable is winning, which is exactly why the analysis worth doing is on where it's exposed. Three real cracks, and they connect.

The ARR is growing faster than the paying base that's supposed to back it. At $100M ARR in mid-2025, Lovable had roughly 8M users but only about 180,000 paying, a 2.25% conversion rate. To support $400M from subscriptions, that paying base has to roughly triple or prices have to climb hard. Osika says net revenue retention is around 100%, but it's unaudited, and a credit-metered product has a structural retention problem: when the meter is the relationship, every renewal is a referendum on the bill.

The credit model is making Lovable feel like the thing it replaced. This is the crack that rhymes with the Wix section, and Lovable should be nervous about it. Its own reviews are bimodal: 3.9 on Trustpilot across ~1,250 reviews with a visible one-star tail, and the one-stars are almost all about money. "I encountered a frustrating cycle where the AI repeated the same coding errors, forcing me to spend thousands of dollars on credits. An expensive dead-end." A survey of 200+ developers found 65–75% hit a bug-loop on complex builds, where the AI burns paid credits failing to fix its own errors. Support is bot-only at a company doing $400M ARR. Read those complaints next to the Wix renewal complaints and they're the same emotion: I'm paying for a meter I can't predict and a company I can't reach. That's the exact resentment Lovable's challenger story is supposed to escape.

The valuation prices a moat the category may not grant. Lovable's $6.6B prices inevitability. But vibe coding is converging from every direction: Wix's own Base44, Replit, Bolt, Vercel's v0, Cursor, and the labs themselves shipping agentic coding. Independent comparisons already treat Lovable, Base44, and Bolt as direct substitutes. When the underlying capability is a model call anyone can make, the durable moat has to be the distribution flywheel and the workflow lock-in, not the generation itself. Lovable has the best flywheel in the category. It does not yet have the workflow gravity (the hosting, the data, the operational stickiness) that Zohar correctly named as the hard part. The race is whether Lovable builds the second moat before the category commoditizes the first.

Three moves, each chained to a specific finding above. No item without a trigger.

1. Ship a cost-preview before every prompt. Move → Trigger → Why.

Move: Show the estimated credit cost of an action before the user runs it, and a running "this session" meter, with a one-click cap.

Trigger: the one-star tail is dominated by credit-burn shock and bug-loops that spend "thousands of dollars" failing, and 65–75% of complex builds hit that loop.

Why it moves the needle: the resentment isn't the price, it's the unpredictability. A meter you can see is a meter you trust, and trust is the entire difference between Lovable's retention story and Wix's. This defuses the single complaint most likely to cap NDR.

2. Productize the land-and-expand the enterprise is already doing for you. Move → Trigger → Why.

Move: Package the bottom-up enterprise pattern (one builder replaces a SaaS contract, then the team piles in) into a self-serve "team workspace + SSO + spend controls" tier with an in-product upgrade path, before the new sales org tries to top-down it.

Trigger: eXp Realty replaced $1M+ in SaaS and Atonom replaced a $40K Salesforce contract with a $1,200 Lovable CRM, both bottom-up, while Lovable is only now hiring its first sellers.

Why it moves the needle: the expansion is happening without a sales motion. Productizing it protects the PLG margin that justifies the multiple, instead of importing the expensive enterprise motion that broke the incumbent.

3. Put a human on the highest-value accounts. Move → Trigger → Why.

Move: A human-in-the-loop support and solutions tier for the top paying cohort (Business/Enterprise), staffed by the GTM engineers Lovable is already hiring.

Trigger: bot-only support is the second-most-cited complaint, and the accounts feeling it most are the enterprise teams running real workloads, the exact accounts that drive expansion revenue.

Why it moves the needle: it's the cheapest possible insurance on the disconnect that actually threatens the valuation, ARR quality. You don't need humans for 8M free users. You need them for the few thousand accounts that make the NDR number real.

The lesson isn't "be an AI startup." Wix uses more AI in its own stack than most companies that call themselves AI-native, and it's the one getting repriced. The lesson is about where the motion lives. Lovable's growth is inside the product: every free build is a billboard, every public app is a fork button, every shipped feature is the marketing. Wix's growth is outside the product, in a $100M budget and a Super Bowl slot, a machine that has to be re-funded every quarter to keep spinning. One of those compounds for free while you sleep. The other stops the day you stop paying.

That's the whole teardown in a sentence: Lovable grows by being shared, Wix grows by being advertised.

The 16x-versus-1.3x gap between them is mostly stage and structure, private hypergrowth against public maturity. What the gap doesn't settle, and what actually decides whether Lovable grows into that $6.6B, is whether a motion built on a commodity model stays durable once everyone can call the same model. The answer is the boring moat: the hard part after publish, the hosting and the data and the operational gravity Zohar named. Lovable has the best distribution flywheel in the category and not yet that second moat. Wix already owns the boring moat. It just forgot to make anyone want it.