The useful part of Arthur Alvarenga's first startup was the part that failed. In 2014, while studying computer science in San Francisco, he encountered Nextdoor, the neighborhood social network. He and his collaborators liked the premise enough to try a Brazilian version. The project became Behive. They spent roughly five years working on community software, only to discover that the local market had already made its choice. Brazilian neighbors did not need a dedicated network. They had WhatsApp.
There are graceful versions of this story in which Behive looks like a planned apprenticeship. It was not. The company did not find a durable market. Other startups founded later appeared to move faster. But Behive left Alvarenga with something a trend report could not provide: years of watching how people actually formed and maintained online groups. When the original product disappeared, that pattern recognition remained.
The clue that became Hubla arrived in an ordinary mess. A fitness creator was selling access to communities on WhatsApp. Every day she checked who had paid, added new customers, removed lapsed members, and tried to keep the whole thing synchronized. Alvarenga says the routine consumed nearly six hours a day. It was not a glamorous creator problem. It was clerical labor with a payment attached.
“She saved almost six hours a day of work.”Arthur Alvarenga, on an early creator using the automation
That was the wedge. ChatPay, founded in 2020, connected payment status to membership in WhatsApp and Telegram groups. Pay and you entered. Stop paying and the system handled the awkward removal. The software turned a spreadsheet-and-notification ritual into a rule. Its appeal was easy to explain because the work it replaced was easy to hate.
The channel was the strategy
Alvarenga did not lose interest in communities after Behive. He changed his view of where a community product should sit. The failed neighborhood network had asked users to adopt a destination. ChatPay worked inside destinations they already opened all day. That distinction was particularly important in Brazil, where WhatsApp was infrastructure long before software companies began describing chat as a platform.
The new company entered Y Combinator's Summer 2020 batch. The accelerator's current directory describes Alvarenga with two terse specialties: software engineering and product design. Both were visible in the early product. ChatPay was a technical integration, but its real design decision was subtractive. It removed administrative steps without trying to replace the conversation around them.
The long route to the narrow wedge
The timing helped. During the pandemic, teachers, coaches, analysts, and other specialists moved their work online. Many discovered that an audience could become a business through courses, subscriptions, mentorship, or paid access. Social platforms supplied reach, but not ownership. A creator could gather followers on Instagram or YouTube and still lack a checkout, a customer list, or a reliable way to deliver what had been sold.
ChatPay met one part of that need. The founding team soon widened the frame. In 2021, the company became Hubla and raised R$60 million in a Series A led by Kaszek, with FJ Labs, Big Bets, and investor Kevin Efrusy participating. At the time, Hubla reported more than 60,000 subscribers across 1,000 active communities and sixfold revenue growth over the prior 12 months.
A checkout grows up
Hubla gradually stopped looking like a utility for paid chat groups. The platform added customizable checkout, payment processing, course and membership areas, subscriptions, content delivery, analytics, and automation. The product followed a creator's customer journey rather than a conventional software category. Once Hubla handled the sale, it could also handle access. Once it handled access, it could measure engagement. Once it had the data, it could help with the next sale.
The company's business model sharpened that logic. In 2026, Alvarenga described Hubla's standard charge as 8.9 percent of a transaction plus R$2.49 per sale. A subscription-software vendor can collect revenue from an unused account. Hubla's fee requires a customer transaction. “We only make money when the creator processes a sale,” he said. Alignment is an appealing word. Here it is also a constraint: if conversion stalls, Hubla feels it.
“The creator looks at us and says: this is my business partner.”Arthur Alvarenga, describing Hubla's desired role
Calling software a business partner raises the bar. A tool can do one job. A partner is expected to notice what happens next. Hubla now pitches itself as the layer between a creator's expertise and the tedious machinery of a digital business: accepting PIX, cards, and boleto; presenting offers; granting access; tracking customers; and helping sellers understand what is working.
There is a defensible strategy in that accumulation. A creator with a checkout from one vendor, a course area from another, groups managed by hand, and customer data scattered between them has flexibility but also friction. Hubla's bet is that an integrated system can turn the scattered work into a loop. The creator makes a sale, the customer receives access, the platform observes behavior, and the next action becomes easier to automate.
The quiet reset
A large round does not remove the need to rebuild. The first paid-groups market attracted competition and offered a ceiling. Hubla had to become more than the original clever integration. According to the company's 2026 account, it concentrated on product depth, operating discipline, and a broader view of the creator as a company. It did not announce another funding round after 2021.
The results offer a cleaner scorecard than the funding news did. Hubla says revenue grew tenfold during 2024 and 2025, and the business reached breakeven at the end of 2025. The company's 2026 goal is to process R$2 billion in sales. Those are company-reported figures, but they show what Alvarenga wants the organization measured against: money moving through customer businesses, not attention around the startup itself.
This phase also puts artificial intelligence into the operating loop. Hubla Agents is designed for jobs such as recovering abandoned carts, supporting customers, and nudging conversion. At a Hubla Day gathering in 2026, Alvarenga presented the product to 150 digital entrepreneurs during a program centered on the impact of AI on their businesses.
For Hubla, an agent matters when it rescues a sale, answers a customer, or saves a creator's time. The measurement is operational: did the work move, and did the business improve?
It is consistent with the original ChatPay instinct. The interesting application is not a chatbot that can perform intelligence on command. It is a system that notices the abandoned checkout at the right moment, follows up in the channel the customer uses, and records the result. The technology changed. The appetite for dull, repeatable work did not.
What the detour bought
Alvarenga calls Hubla his third company. That detail matters because its apparent speed rests on a long prehistory. Behive taught the team about communities. The New Brand Design, which Alvarenga led from 2018 according to public career listings, added another operating chapter. ChatPay then arrived with founders whose backgrounds covered software, mobile systems, product, and payments. The company that joined Y Combinator in 2020 was young. Its pattern recognition was not.
People who worked with Alvarenga describe the company as a training ground as well as a workplace. When former Hubla CTO David Reis left to found a credit-software startup, he recalled Alvarenga advising him earlier that an executive role at a Series A company would be strong preparation for entrepreneurship. Reis later wrote that the advice had been right. The exchange suggests a practical view of ambition: help someone learn how the machine works, even if that eventually helps them build another one.
Alvarenga's public writing has earned him a LinkedIn Top Voice badge for sharing the challenges of entrepreneurship. His sharper lesson, however, sits in the sequence of products. A founder can be wrong about the destination and right about the terrain. Five years spent on the wrong neighborhood network created familiarity with group behavior. That familiarity made six hours of someone else's repetitive work legible as a company.
Hubla's future depends on whether it can keep turning that sort of observation into reliable infrastructure. The creator economy changes costumes quickly. One year the object is a paid Telegram channel; another year it is a cohort course, a membership, or an AI-assisted sales funnel. The administrative questions are steadier. Who paid? What did they buy? Did they receive it? Will they stay? What should happen next?
Alvarenga built his current business by respecting those questions. Hubla began where the audience was already talking and where a creator was already earning, then removed the labor between the two. Its larger ambition is to make an online expert operate more like a durable company. The path from Behive to a R$2 billion processing target is not a tale of spotting the future in one flash. It is a story about staying close enough to failed work to notice the useful piece it left behind.