Rocketship.vc finds the company before the introduction

A founder in Brazil can build a business that a California investor never hears about. Nothing has gone wrong with the business. The missing ingredient is a person: somebody who knows somebody who can forward an email. Venture capital has spent decades making this arrangement seem natural. Rocketship.vc begins with the suspicion that a good introduction and a good investment are different things.

• The method: use startup data to discover companies, then contact their founders.
• The offer: investment capital, operating experience and help with data.
• The reach: a portfolio spread across 14 countries.

The firm’s wager is that an expanding company leaves evidence beyond the investor’s address book. Find that evidence systematically and the map of investable businesses gets bigger. For a founder outside familiar fundraising circles, that is an attractive proposition. For an investor, it creates a harder question: how much of your judgment depends on who happened to enter the room?

First, change the search

Rocketship is a venture firm, with startup equity as its investment and limited partners supplying its funds. Its research machinery serves that business. Founders receive financing and support; they are not buying a dashboard. The portfolio includes Wasabi, NoBroker, Moglix and Agnikul. Cloud storage, housing, industrial purchasing and launch vehicles make an eccentric shop window. The common thread is the selection process.

In a 2023 interview, managing partner Madhu Shalini Iyer described an outbound process: about 150 companies shortlisted monthly, followed by conversations and further investigation. She placed the emphasis on Series A, where there is more evidence to examine, with occasional Series B investments. Those were dated operating details, but they explain the logic. The machine makes the first list. People still have to establish what belongs on it.

The origin was an experiment with an escape hatch. In his 2020 fundraising essay, Ramakrishnan recalled asking whether information from sources such as Crunchbase and social media could support an investment decision. If the answer was no, the team could try marketing, hiring or recommendations instead. Backtesting encouraged them to proceed into venture investing in 2015. That is a useful detail for anyone tempted to copy the result: the commitment followed a test of the method. They had another use for the technology if the central hypothesis disappointed them.

The people who built the filter

Anand Rajaraman and Venky Harinarayan had already worked on finding useful things in enormous collections of information. Their shopping search company Junglee went to Amazon. Their search and categorization business Kosmix went to Walmart. Rajaraman also co-authored Mining of Massive Datasets. The intellectual inheritance is unusually literal: search, classification, then a decision about what deserves attention.

Co-founder and managing partner Sailesh Ramakrishnan adds a different piece of the story. Before Kosmix and WalmartLabs, he worked at NASA Ames. Later, he co-founded LocBox, acquired by Square. Iyer brought experience as Gojek’s chief data officer. These are backgrounds in systems that have to operate, rather than simply look plausible in a presentation. For portfolio founders, the promised help includes people who have built and scaled companies themselves.

Forty. One hundred. One hundred twenty-five.

The financial sequence is straightforward when the labels stay attached. Rocketship’s first fund closed at $40 million in late 2015. In August 2020, the firm announced a $100 million second fund, backed by investors including Vulcan Capital and Adams Street Partners. July 2022 brought a reported $125 million first close for Fund III.

These figures describe pools of investment capital. They do not describe the management company’s sales, and a first close is an intermediate fundraising milestone. Keeping those distinctions straight matters: venture firms deploy other investors’ capital into businesses whose eventual outcomes remain uncertain. A bigger fund says something about the resources available. It does not settle the investment argument.

The map got bigger. The uncertainty stayed.

By its 2020 announcement, Rocketship had invested in 44 companies, with 46 percent of investments outside the United States. Its current portfolio page describes investments in 14 countries. Those are meaningful measures of geographic reach. They also help explain why remote discovery was useful before the pandemic made video calls an everyday substitute for travel.

The limit appears earliest when a company has little observable history. It is difficult to recognize a pattern that has barely begun. Another problem follows: recorded activity can be incomplete, inflated or misleading. As an analytical matter, expanding the search cannot eliminate the need to question the measurements. Rocketship’s own 2020 discussion of bias acknowledged that people still shape the models and their inputs.

A data firm questions the data moat

Then comes a pleasing complication. In July 2026, Rocketship published an essay arguing that proprietary data and accumulated user memory may offer less protection in the AI-agent market than founders expect. Retrieval, public substitutes and increasingly capable inference can erode the advantage. The proposed alternative emphasizes integration and user control. This is a market argument, not an announcement that Rocketship has discarded its investment system.

Its May 2026 SaaS essay supplies a practical companion: watch revenue and margins together. Growth purchased through discounts is a different proposition from growth accompanied by improving economics. The interesting consistency is a willingness to ask what the evidence actually proves, even when the fashionable answer happens to flatter a data specialist.

“The signal we’d watch for: revenue growth matched by margin growth.”

Rocketship.vc · May 2026

Make the evidence travel

The copyable lesson is modest and useful. Start by widening the field of candidates. Decide which observations deserve attention. Then investigate the people and businesses behind them. Founders can apply the same discipline in reverse: make customer demand, repeat use and the economics behind growth legible to someone who has never met them.

That will be less useful for a company whose progress is difficult to observe, or whose market lacks comparable signals. It cannot manufacture trust or repair weak governance. But it can change who gets noticed. Rocketship’s most interesting intervention happens at that first moment, before the meeting and before the pitch, when somebody decides which company is worth calling.