Crunchbase CEO on Startup Investment and Data Strategy in the AI Era

In this article, based on an interview with Crunchbase CEO Jager McConnell, we provide a professional yet accessible explanation of a wide range of topics, from the current state of startup investment to the role of data companies in the AI era, as well as predictive models and organizational transformation.

Below, each section is numbered, and subsections are provided where necessary for explanation.

1. Current State of Startup Investment

1-1. Changes in the Funding Environment

According to Jager McConnell, startup investment as of 2025 is "close to 2019 levels, but has decreased by more than half from the 2021 peak." McConnell points out, "Compared to the past it's not that great but compared to the recent past it's doing pretty good. We’re at like 2019 levels of funding, but more than half off of our peak in 2021." In other words, while funding from VCs (venture capital) still exists, the supply itself has decreased, and funds are being issued only after more rigorous screening.

1-2. The Gap Between Valuations and Funding Needs

Many entrepreneurs feel that seed-stage valuations remain high compared to the past. As evidenced by the statement, "The average YC company today is raising money at a higher valuation than ever before," investor interest is concentrated on a few "treasure deals" in popular sectors, particularly AI, which command high valuations, while it is difficult for many other deals to secure funding. McConnell states, "Valuations of seed-stage companies are as high as ever, but it's harder to get funding," symbolically illustrating this supply-demand gap.

2. The SaaS vs. Internal Solution Debate

2-1. The Advantage of Internal Development in the AI Era

McConnell predicts that due to the rapid evolution of AI technology, the movement to custom-develop business support tools (such as CRM and ERP) that were previously purchased externally will accelerate within companies. In the interview, he says the following:

In other words, by making full use of AI tools, the possibility of quickly building functional internal applications without having a large number of engineers is increasing. He argues that this trend of "in-house production" is particularly notable in companies with complex and detailed business requirements, making it difficult to differentiate from external SaaS.

2-2. Cost and Sustainability

On the other hand, there is also the point that "If you have more than one full-time developer, it's often cheaper to buy than build." There are many cases where it is cheaper to utilize existing SaaS rather than waiting for the evolution of AI technology. Therefore, companies need to carefully consider whether they should "completely shift to in-house production right now" and make a comprehensive judgment based on costs, future scalability, and maintenance and operation systems.

3. Business Defensibility in the AI Era

3-1. Fluctuating Technology and VC Challenges

McConnell points out that the current technology environment is in the midst of such rapid disruption that it "changes on a weekly, monthly, and yearly basis," making it "extremely difficult for VCs to predict changes and place bets (invest)." McConnell sharply criticizes this, saying, "VCs don't really grok what it means to invest in a tool now when the tech will be obsolete in three years." In other words, as the technology lifecycle is compressed, traditional investment methods of "reading the market's future" are becoming obsolete.

3-2. What Assets Truly Have Defensibility

Under these circumstances, to make a business sustainable, it is not enough to "just be the first to enter" or "have a unique process"; a "moat" that completely prevents third parties from easily entering is required. McConnell asserts that "Proprietary, unique, changing time-series data is the only thing that truly has defensibility." In other words, since certain static data and processes are quickly copied by AI and LLMs (Large Language Models), the idea is that only data that continuously changes and increases in value will be a weapon for survival.

4. The Role of Proprietary Data

4-1. The Difference Between Historical Data and Real-Time Data

In traditional data businesses, the core business model was collecting and selling historical time-series data. However, it has been pointed out that "Past data can be absorbed by LLMs and becomes commoditized." Moving forward, real-time changing usage data and data on what information users view or edit will become the differentiating factors. Mr. McConnell states the following:

In other words, data known as Usage Data, which tracks who accessed what information and when, is becoming more valuable than traditional static facts (financial metrics and company profiles).

4-2. Crunchbase's Data Strategy and Challenges

Since its inception, Crunchbase has grown by leveraging user-generated content (UGC) data, but according to Mr. McConnell, "User-generated content now accounts for only 5% of the total, and the main battlefield has shifted to other data pipelines." Furthermore, because LLMs "retain facts forever once absorbed," simply accumulating and providing historical data makes it easy for AI to catch up.
Therefore, the company is shifting toward dynamic, value-added services such as "Usage Data" and "predictive models," and is being forced to rethink its traditional business model of "charging for deep data access." Mr. McConnell emphasizes the need for transformation, stating, "If we only charge for deep access today, we may have to give away the data tomorrow and sell only insights."

5. M&A Trends for Data Companies

5-1. Current State and Challenges of the M&A Market

Data companies have been in a situation over the past few years where "both sellers and buyers find it difficult to move." Mr. McConnell says, "Sellers cannot sell at the prices they expect, and buyers cannot find suitable candidate companies." Specifically, there is a reality where "many data companies cannot even trade at 4-5 times EBITDA, and conversely, the market is shrinking due to industry consolidation."
On the other hand, with the spread of AI, large companies are accelerating their efforts to "incorporate data into their products as quickly as possible," and the M&A market is becoming active again. Mr. McConnell explains, "AI is improving the profitability of data companies, and acquisition needs are rising for some firms. However, because deals are not being concluded at appropriate valuations, mismatches are occurring for both sellers and buyers."

5-2. Characteristics of Companies Likely to be Acquired

Regarding what kind of data companies are attracting attention as acquisition candidates, Mr. McConnell lists the following conditions:

1. Possessing essential data

   • Example: Credit card transaction data essential for credit judgment by financial institutions, etc.

2. Equipped with technical capabilities or APIs that buyer companies can immediately integrate into their products

   • Example: A pipeline that can provide real-time fluctuating Usage Data.

3. Having built brand power or a user base in a specific market

   • Example: A platform like Crunchbase that is used daily by many investors.

Companies that meet these conditions are "easy to match because sellers can demand high valuations and buyers can gain added value immediately after acquisition." However, Mr. McConnell warns that if even one requirement is missing, the mismatch of "companies that want to sell cannot sell" and "companies that want to buy cannot find targets" will continue.

6. Predictive Models and Corporate Acquisitions

6-1. Importance and Accuracy of Predictions

Crunchbase is focusing on developing features to predict "which companies are likely to be acquired." Mr. McConnell says, "We have over 95% precision in our acquisition predictions," and they are building machine learning models using vast feature vectors. Specifically, they are increasing accuracy by making predictions that combine multifaceted indicators such as past funding information, Usage Data, and the frequency of company profile edits.

This "acquisition prediction" is expected to be a new form of "data-driven M&A brokerage" that does not rely on traditional human-based investment bankers.

6-2. Impact on User Experience and Business Models

Mr. McConnell reveals that Crunchbase plans to implement a UI on its applications that allows companies to check their own acquisition probability score in real-time (e.g., via a bar graph). Furthermore, by providing user companies with information via API such as 'which companies are potential buyers' and 'which companies those buyers are looking to acquire,' they aim to realize a self-service M&A solution that bypasses investment banks.

This means that even small and medium-sized enterprises, which previously paid millions to tens of millions of yen to investment banks, are now gaining an environment where they can discover acquisition candidates and initiate negotiations at a low cost and with high efficiency.

7. Transformation into an AI-First Organization

7-1. Changes in Internal Mindset

With the arrival of the AI era, organizations and individuals are being forced to fundamentally rethink their traditional ways of working. Mr. McConnell states the following:

In short, from management to frontline staff, there is a requirement to move beyond 'using AI as a mere tool' to 'switching the foundation of daily work thinking to an AI-first premise.'

7-2. The Importance of Employee Development and Vendor Selection

Mr. McConnell also argues that 'the most important skill moving forward is not hiring and managing talented people, but the ability to select appropriate vendors and AI tools and integrate them into the company's business processes.' Specifically, he recommends the following steps:

1. Break down daily work and identify the tools currently used for each task.

2. Research vendors, including AI tools, and consider more efficient options.

3. Share results across each department and establish best practices cross-functionally.

Mr. McConnell himself says he instructs all executive officers to 'break down their daily activities by percentage, list the AI tools currently in use for each, and constantly research tools that can improve efficiency.'

Thus, the era is shifting toward one where talent capable of mastering AI tools takes center stage, making the review of traditional hiring criteria and training programs an urgent task.

8. San Francisco and Startup Culture

8-1. Geographic Concentration and Trends Among the Youth

Mr. McConnell points out that 'as of 2025, ambitious young people in tech are still gathering in San Francisco.' He says the following:

While it seemed that people had dispersed to regional cities due to the spread of remote work over the past few years, it is said that there is a return to San Francisco, where investment capital is once again concentrating. McConnell notes, "This is just a temporary cycle, and it may shift to another trend in a few years," but he also evaluates it by saying, "It is currently the most exciting place for young people."

8-2. The Value of VC Funding and the Trend of Bootstrapping

On the other hand, McConnell also says, "VC funding is not necessarily the top priority." In particular, bootstrapping (a method of expanding a business with self-funding or initial revenue) is being re-evaluated, and he points out that "how quickly to become profitable and stabilize cash flow" is becoming more important than fundraising. He states the following:

From an investor's perspective as well, the idea that "companies with business models that can achieve maximum growth with minimal capital are truly valuable, rather than raising more money than necessary" is beginning to spread.

9. Future Business Education and Careers

9-1. The Decline in the Value of Academic Degrees and Entrepreneurial Orientation

In the latter half of the interview, McConnell states, "The value of school education is declining rapidly, and in the future, many young people will choose to start businesses without going to university." He presents the following points:

• There is a gap between the skills learned at university and the skills required in actual business settings.

• Using AI tutors (such as ChatGPT), there are more areas where one can learn faster and deeper than with textbooks.

• Creative fields and art, in particular, are the last areas that are difficult for AI to replace, and it can be said that now is the time to enter.

These claims may sound extreme, but they suggest that a value system of "practice over education, and AI utilization over practice" is emerging.

9-2. The Evolution of AI and Learning Methods

Conventionally, learning has been acquired through educational institutions such as universities and vocational schools, but McConnell foresees that "if personalized learning by AI becomes widespread, one can take individually optimized curricula without going to school." Specifically, he proposes the following flow:

1. Preparation and review using AI tutors

   • Example: If you want to learn about the "Peloponnesian War," have an AI summarize the key points and resolve any questions through dialogue.

2. Discussions in online communities

   • Example: Exchange opinions with peers and industry professionals in specialized communities on Discord or Slack.

3. Project-based practical learning

   • Example: Participate in open-source projects and learn while actually writing code.

As a result, the reduction of educational costs and the improvement of learning efficiency will be achieved simultaneously, and it is highly likely that the conventional "degree-first" ideology will be reconsidered.

In this article, based on an interview with Crunchbase CEO Jager McConnell, we explore a wide range of topics, from the current state of startup investment and the role of data companies in the AI era, to the use of predictive models, the necessity of organizational transformation, and even the startup culture of San Francisco and the future of education. The key takeaways are as follows.

AI technology and data businesses will continue to evolve, bringing significant changes to investment environments, organizational management, and even individual career development. We hope this article helps readers adapt to these changes and build next-generation businesses.