B2B Growth Moved to the Ends of the Barbell, and the Middle Is Paying for It
Lovable runs on 146 employees with revenue reported around $500 million, while Shopify grew revenue 34% and Samsara grew ARR 30%. The median public B2B software company is growing about 13%.
Lovable employs 146 full-time people. This week it raised $400 million at a $13.3 billion valuation, double its mark from earlier this year, on revenue reported around $500 million. Divide the revenue by the headcount and it comes to roughly $3.4 million per employee, a ratio without an obvious precedent in software at this scale.
At the other end of the same market, the standout earnings this season belong to companies nobody markets as AI stories. SaaStr's read across the quarter: Shopify grew revenue 34 percent, Toast grew 23 percent while adding a record 9,500 net new restaurant locations, and Samsara grew ARR 30 percent. These are software companies attached to shelves, kitchens, and trucks. The median public B2B software company, for comparison, is growing about 13 percent, and the classic horizontal categories, CRM and sales and marketing and collaboration tools, are growing in single digits.
Two poles, then. At one end, a product that distributes itself so completely that half a billion dollars of revenue runs on a team that would fit in a restaurant. At the other, products metered to physical throughput, growing at twice the market median. The barbell is the story, and the middle is paying for both ends.
Seat pricing is where the squeeze lands
The middle looks like this. HubSpot grew ARR 20.1 percent to $3.58 billion and still cut full-year revenue guidance, citing larger buying committees and deals that now wait for C-suite or board sign-off. Its CFO said the "headwinds we saw in Q2 will persist throughout the remainder of the year." Salesforce's most recent quarter grew 13 percent. This is the profile of the squeezed asset: horizontal software, priced per seat, sold to knowledge workers, through a motion designed when a deal needed two signatures instead of five.
The squeeze has a mechanism, and it runs through the pricing unit. SaaStr's observation about the physical-economy winners is that they price on transactions, locations, and assets rather than seats. "AI does not reduce the number of meals served," as the piece puts it. When AI makes a restaurant more productive, Toast keeps its meter. When AI makes a sales team more productive, the seat vendor loses seats. The same technology that expands one company's billable base contracts the other's, and both effects compound quarterly.
Lovable keeps the margin others spend on sales
Lovable escapes the squeeze from the opposite direction. Its product is the funnel: users arrive, build something, and pay, with almost no go-to-market payroll in between. The 13 percent growers carry thousands of quota-carrying salespeople as a cost of doing business. Lovable's entire company is smaller than most mid-market sales teams, which means the margin the middle spends acquiring customers is margin Lovable keeps. Anton Osika announced the round as capital "to build the business that builds businesses," with Menlo Ventures returning to lead alongside EQT's Scaleup Europe Fund.
The honest test of that pole is already scheduled. The raise funds 68 open roles, roughly half again the current headcount, and every previous generation of hyper-efficient company got less efficient the moment it had $400 million of encouragement. Whether Lovable's curve survives its own funding is the most instructive experiment in GTM right now, and it will run in public.
The comparison pressure lands on everyone else either way. Lovable's curve is already the new denominator in board conversations: a GTM leader defending a plan that spends 40 percent of revenue on sales and marketing is now defending it to an investor who just read about a company spending almost nothing. The comparison is unfair, and the unfairness doesn't matter. Benchmarks set narratives, and narratives set budgets. That is the quiet tax the barbell levies on the middle: not lost deals, but lost benefit of the doubt.
Procore also sells software attached to physical work and grew 16 percent, so atoms guarantee nothing; vertical dynamics and execution still decide individual outcomes. And one earnings season is weather, not climate. Fair on both counts. But the direction of the repricing shows in who has to explain themselves this quarter. The guidance cuts and the lengthening sales cycles concentrate in horizontal seat-land. The raises, the multiples, and the beat-and-raise quarters concentrate at the poles.
The middle pays for one pole's distribution and the other's meter
The sequence runs through the pricing unit. A customer's AI rollout shrinks a team, the seat vendor loses seats, and the larger buying committee HubSpot described stretches the deals that remain. Toast's meter counts meals, which AI does not reduce. If the pattern holds, every vendor pricing a seat in front of knowledge work carries the exposure HubSpot guided to for the rest of the year.
Lovable's 68 open roles will test the other pole. It runs revenue reported around $500 million on 146 people, roughly $3.4 million per employee, and the hiring adds about half again its headcount.
