Infinity Constellation Raised $24M to Mass-Produce AI Companies. That Is the Studio Thesis.

On June 8 2026, Infinity Constellation told Axios it had raised a $24M Series A to build up to eight AI companies a year inside the professional services market. The round was led by Freestyle, with Backed VC, Rafferty, Oxford Funds, BY Ventures and others, as reported in the [Axios exclusive](https://www.axios.com/pro/all-deals/2026/06/08/infinity-constellation-24-million-multiple-ai-companies) and the company [press release](https://www.prnewswire.com/news-releases/infinity-constellation-raises-24m-series-a-to-take-on-professional-services-302794633.html). This is not a bet on one product. It is a bet on a factory.

Read it next to the model [Avante Ventures](/why-avante) runs in Brazil and Latin America and the signal is hard to miss. The venture studio, once a niche structure argued about by a handful of believers, is now being capitalized like an asset class. The interesting question is no longer whether the studio model works. It is who runs it, and where.

What Infinity actually raised money to do

Infinity Constellation, led by CEO Brennan Pothetes and chaired by Francis Pedraza, the founder of [Invisible Technologies](https://www.invisible.co), is going after the $6 trillion global professional services market. Pothetes framed it plainly. Professional services is a $6T industry still built around people billing hours, and that era is ending. The thesis on the [company site](https://www.infinityconstellation.com) is sharper still. The future of the firm is built, not bought.

The structure is the story. Infinity does not buy legacy firms and bolt AI onto them. It builds new AI-native companies from a central platform, then lets data, pricing, and expertise flow between them. Per Axios, it prefers repeat founders, who take 25% of the newco plus a stake in the holding company itself, so every operator is incentivized to make the next company better. It already runs several live units across executive assistance, compliance, design, hiring, finance, and data.

This is a studio story, and the numbers explain why

Strip the branding and Infinity is a venture studio. A central team builds companies on repeat instead of writing checks into other people's companies and hoping. That distinction is not cosmetic. It changes the returns. Per the Global Startup Studio Network, studio ventures average roughly 50% IRR against roughly 19% for traditional venture-backed startups, about 2.5x over realistic time horizons. The same dataset shows studio ventures reaching Series A in about 25 months against 56 for traditional startups, and graduating at 72% versus 42%. A model that halves the time to a priced round and nearly doubles the survival rate is not getting lucky on deal selection. The full case is in our breakdown of [why venture studios win in LATAM](/library/why-venture-studios-win-latam).

Infinity says the quiet part in product language. One shared brain. Compounding by design. Each company makes the next one faster and smarter. That is the studio flywheel, and it is the exact reason a studio outperforms a fund that gives advice across a dozen boards. For the side-by-side on structure, see [studio versus accelerator versus VC](/library/studio-vs-accelerator-vs-vc).

AI is what makes a company-a-year believable

A studio launching eight companies a year would have sounded reckless in 2019. Axios put the reason it no longer does in one line. It illustrates how rapidly companies can be spun up in the age of AI. The cost of standing up the first working version of a company has collapsed, because [LLM inference prices have fallen by orders of magnitude](https://a16z.com/llmflation-llm-inference-cost/) and the tooling now does work that used to need a team.

That is why a serious studio can launch AI-native and reach revenue before it would have historically reached a priced round. Infinity claims one month to first revenue across its units. The same dynamic is what lets a disciplined builder ship without burning a Series A first, which we cover in [how to build AI-native without a Series A](/library/ai-native-without-series-a) and the [AI infrastructure cost curve in LATAM](/library/ai-infrastructure-cost-curve-latam).

Where Avante and Infinity part ways: the map

Infinity is attacking the global professional services market from the United States. Avante runs the same playbook where the gap between the size of the opportunity and the level of software is widest. Services account for roughly 70% of Brazilian GDP, and most of that economy is barely digitized. A studio that supplies operating depth on day one compounds hardest exactly there, where that depth is scarce and a generalist fund cannot manufacture it.

This is the structural edge behind [why Avante builds where it builds](/why-avante). Domain operators with ten or more years of Brazilian-market scar tissue, paired with a Silicon Valley playbook and first-ticket capital, assembled before the company exists. Infinity validates the model. Geography is where the next decade of studio returns gets decided, and the deeper read on that is in [the Brazilian services economy opportunity](/library/brazil-services-economy-opportunity).

The line the headline buries: data

The funding number gets the headline. The durable advantage is quieter. When repeat founders build inside one shared brain, the data each company generates flows into the next one. That is a moat that a single startup, however well funded, cannot assemble alone. It is also the part most reporting on studios skips.

Avante names the loop directly. Copilot to data to fund. Build an AI copilot to generate proprietary data, then use that data to raise and deploy capital. A holding company that shares a brain across ventures turns that loop into a portfolio asset rather than a single-company trick. The mechanics are in [the copilot to data to fund flywheel](/library/copilot-to-data-to-fund-flywheel), and why the data compounds into defensibility is in [data network effects in vertical AI](/library/data-network-effects-vertical-ai).

What an LP should take from a $24M round

Read the Infinity raise as a market signal, not a competitor scare. Serious investors are now funding the studio model at scale, in public, with a number attached. The debate about whether building companies on repeat beats betting on them one at a time is closing. What remains open is execution and geography.

Avante is the answer for Latin America. A venture studio building AI-native companies in Brazil and Latin America, launching 3-4 ventures per year, deploying $500K-1.5M per venture, and running every one through a six-stage system from Research to Partner to Build to Traction to Revenue to Compound. The same model the market just priced at $24M, pointed at the economy where it should compound hardest. For an LP, the takeaway is not to chase Infinity. It is to ask which studio is positioned where the model has the most room left to compound. Read [how Avante operates](/principles), or [start a conversation](/contact).