Transforming VC with Gemini API: How Vela Partners Uses Grounding with Google Search for Deeper, Faster Insights

In the high-stakes world of venture capital, identifying the most promising startups early on requires sifting through mountains of data and making swift, informed decisions. Vela Partners, a leading quant venture capital firm, is revolutionizing this process by leveraging the Gemini API, turning the art of investment into a data-driven science with unprecedented speed and precision.

Vela Partners isn't your typical VC. With a mission to be "100x faster and better," they've embraced an AI-first strategy since 2017, investing exclusively in AI startups. To achieve this, they needed an AI platform that could keep pace. The Gemini API, particularly Gemini 2.0 Flash with Grounding with Google Search, provided the right solution.

Vela Partners uses the Gemini API to power two critical workflows:

• Enriching startup profiles with Grounding: Early-stage startups, especially those born from open-source projects, often have minimal online presence. Vela uses Gemini 2.0 Flash grounded with Google Search to automatically fill in the gaps. When analyzing a GitHub repo, Gemini can surface key details about contributors – location, affiliations, experience – enriching the data used by Vela's internal models to assess potential. "Gemini and Grounding with Google Search has been exactly the solution that we were looking for a long while," says Zakari Salifu, AI Engineer at Vela Partners. This capability surpasses resource-intensive multi-agent frameworks and advanced analysis techniques (SSFF and Founder-idea-fit) previously explored by the firm.
• Streamlining due diligence with real-time analysis: Once a potential investment is identified, deep due diligence is crucial. The Gemini API, integrated via the Vercel AI SDK, pulls real-time data from news, social media, and more, providing a comprehensive view of the company and its market.

Before the Gemini API, gathering comprehensive information on nascent startups was a major bottleneck. "We invest in startups and people at their earliest stages," explains Yigit Ihlamur, General Partner and CTO. "At this stage, information is scattered across the web."

The Gemini API's Grounding with Google Search feature proved transformative. By grounding Gemini 2.0 Flash's responses in Google Search results, Vela significantly reduces the risk of hallucinations and ensures their analyses are based on the latest, most reliable public information. This was a game-changer compared to their previous in-house solutions.

For Vela, the Gemini API offered the ideal balance. While they had developed sophisticated proprietary methods, they sought a more efficient solution. Afriyie Samuel, Frontend Intelligence Engineer, notes, "We needed a lighter, more rapid, and cost-efficient version of deep company analysis. Gemini 2.0 Flash with Grounding has been exactly the solution we were looking for."

Compared to other models, Gemini 2.0 Flash delivered superior performance for entity enrichment, even outperforming Vela’s own internal techniques.

Vela Partners is already seeing significant improvements in the speed and accuracy of their analyses. Looking forward, they plan to integrate the Gemini API further, exploring research features and file ingestion for due diligence.

Their advice to other developers? "Consider what Google is great at. Google is the undisputed leader in search technologies. Start using Gemini 2.0 Flash with grounded search and compare the quality of results with other LLMs out-of-the-box. You’ll experience the magic."

Vela Partners' journey shows how the Gemini API is not just a tool, but a strategic advantage, empowering them to lead the way in the AI-driven future of venture capital.

Ready to build? Explore the Gemini API documentation.