The Three AI Giants That Agreed on One Thing: Daloopa

Anthropic, OpenAI, and Perplexity have one thing in common: they are all customers of Daloopa.

The New York startup just raised a $47 million Series C — its third fundraise in two years — to be the connective tissue between AI agents and financial data. Brighton Park Capital led. The pitch: as hedge funds and banks push AI from pilot into production, someone has to make sure the numbers feeding those models are accurate, traceable, and not hallucinated. Daloopa is betting it can be that layer for both Wall Street and the AI labs competing to replace it.

The customer list is the story. Daloopa's press release, published May 28, lists MCP connectors — Model Context Protocol integrations — with OpenAI's ChatGPT, Anthropic's Claude, Perplexity, and a fourth firm called Rogo. The company covers more than 5,500 public companies globally, delivers ten times more data points per company than competing providers, and links each figure to its original SEC filing. Its own benchmark study claims AI agent accuracy improves up to 71 percentage points when grounded in structured data versus web-based retrieval.

Those are Daloopa's claims. What makes the names worth noting is that Anthropic, OpenAI, and Perplexity are not passive data buyers. They are building AI products that compete with each other across search, research, and agent workflows. For all three to appear on the same vendor's customer list — in a press release, not a regulatory filing — suggests either that Daloopa's data is good enough that competitive loyalty doesn't apply, or that the supply of trustworthy financial data infrastructure is thin enough that everyone shops at the same store. Whether this represents genuine ecosystem consensus or coincidental convergence is not settled by the available evidence; what is settled is that Daloopa has named them publicly, and none has objected.

The funding environment for AI data infrastructure has tightened alongside everything else. But Daloopa's ability to close a $47M Series C in a market where many infrastructure plays are struggling suggests the category has found product-market fit in a specific, high-stakes problem: AI agents will not outpace human analysts until they stop making up numbers. Hallucinated earnings per share in a valuation model are not a minor inconvenience — they are a liability that blows back on the firm that deployed the model.

Thomas Li, Daloopa's CEO, put it this way in the company's announcement: "It's no longer enough for models to simply generate answers; they must be accurate and fully traceable." That framing — accuracy as a product requirement, not an aspiration — is the sales pitch to a market that has had enough of AI outputs that sound confident and are wrong.

The early-mover logic is real. The companies that locked Daloopa into their AI stacks before this funding round have a structural advantage: they shaped the product roadmap through feedback, hold reference relationships with the team, and are embedded in the benchmark data that Daloopa uses to prove its value to new customers. Companies coming in after a $47M raise and a published customer list are competing for access to the same data layer on terms set by a company that now has more resources, more leverage, and a roster that reads like a who's-who of AI research. That is the second-order effect the funding round quietly announces: the window for locking in AI data infrastructure at founder-friendly terms is narrowing.

The broader context: May 2026 was a significant month for NYC AI infrastructure. Modal Labs closed the largest round at $355 million at a $4.65 billion valuation, with revenue reportedly growing fivefold in six months — from approximately $60 million annualized in September to $300 million in May. The jump happened in two stages: first investors came in at a $2.5 billion valuation, then a second tranche priced at $4.65 billion after more investors piled in. Sacra independently estimated Modal at $50 million annualized revenue three months before the company disclosed $300 million — a gap that may reflect different definitions of what counts as revenue, or may reflect something Modal has not explained. The Modal and Daloopa rounds together — compute and data, $355 million and $47 million — map the two most contested layers of the AI stack right now: who runs the models, and what facts those models are allowed to use.

Daloopa's prior disclosed total equity funding was $101.4 million before this round, meaning the new capital pushes cumulative funding past $148 million. No public valuation signal has been disclosed. For comparison: Modal's $300 million annualized revenue against its $4.65 billion post-money valuation implies a roughly 65x revenue multiple — a price that requires either extraordinary growth or extraordinary faith. Daloopa's path to justifying any comparable multiple runs through the same constraint: proving that its data layer is sticky enough that Anthropic, OpenAI, and Perplexity cannot easily replace it with something built in-house.

Whoever controls the data layer controls part of the answer the model produces. That's not a venture-scale business yet. But it is a $47 million bet that it will be.