AI companies worldwide secured over $100 billion in venture capital funding in 2024, according to Crunchbase, marking an impressive 80% increase compared to 2023. This accounts for nearly one-third of all VC dollars invested last year, underscoring the monumental growth and interest in the AI sector.

However, this explosive growth has brought challenges. The AI landscape is increasingly crowded, with a mix of overlapping startups, companies touting AI in marketing but lacking practical application, and a few genuine game-changers. For investors, identifying the next category leaders amidst this noise is no small task. So, where do they begin?

The moat that matters: Proprietary data

According to a survey by TechCrunch involving 20 venture capitalists focused on enterprise AI, the consensus is clear: proprietary or hard-to-access data is the key differentiator. Startups with unique data sources and the ability to leverage them effectively have the greatest potential for long-term success.

• Paul Drews, Managing Partner at Salesforce Ventures, emphasized the rapid evolution of the AI landscape. He prioritizes startups combining differentiated data, technical innovation, and compelling user experiences.
• Jason Mendel, Venture Investor at Battery Ventures, echoed this sentiment, highlighting the importance of “deep data and workflow moats.” Unique data access allows companies to create superior products, while sticky workflows ensure they become indispensable to customers.
• Scott Beechuk, Partner at Norwest Venture Partners, pointed out that startups focusing on proprietary data, particularly in vertical solutions, are best positioned for lasting success.

Data as a feedback loop and differentiator

For many VCs, proprietary data is more than just a resource—it’s a feedback loop that strengthens AI systems. Andrew Ferguson, VP at Databricks Ventures, noted that rich, feedback-driven customer data not only enhances AI effectiveness but also sets startups apart.

Case in point: Valeria Kogan, CEO of Fermata, shared how her company uses a mix of customer data and in-house research to train its AI models for detecting crop diseases. By handling all data labeling internally, Fermata ensures a level of accuracy that competitors struggle to match.

Domain expertise and workflow integration

Investors are also drawn to startups with deep domain expertise and strong customer workflow integrations. Jonathan Lehr, Co-Founder at Work-Bench, explained that vertical AI opportunities—those tackling specific business workflows—are highly attractive. Startups enabling the acquisition and cleaning of previously inaccessible data hold immense value.

Beyond data, other factors driving investor interest include:

• Teams with strong AI talent
• Seamless integration with existing technologies
• A deep understanding of customer pain points and workflows

The road ahead for AI startups

As the AI industry evolves, the bar for success continues to rise. Startups must go beyond generic applications of AI, focusing instead on unique, actionable data and tightly integrated workflows. For investors, the challenge remains finding the diamonds in the rough—those startups poised to shape the future of AI.

With $100 billion already flowing into the sector, 2024 may well be remembered as the year of the AI gold rush. The question now is: which startups will stand the test of time?

Source: https://techcrunch.com/2025/01/10/vcs-say-ai-companies-need-proprietary-data-to-stand-out-from-the-pack/