Where Data Meets Intelligence: Key Insights from Data Council 2025

I recently spent three immersive days at Data Council 2025 in Oakland, Calif. Billed as a “no bullsh*t” forum for practical insights, the conference brought together top minds in data and AI from Databricks, Google DeepMind, Meta, OpenAI, Perplexity, Snowflake, a16z, Insight Partners and many other leading companies and investors. 

The energy around the modern data stack and AI ecosystem was palpable from thought-provoking sessions to cutting-edge product demos, it's clear that innovation, capital, and customer interest are all converging rapidly. 

Here are 5 key takeaways that stood out to me: 

1.     The Rise of Agentic AI 

Multi-step agent systems face a tough hurdle: error compounding. A modest 5% error rate per step can balloon into a 40% failure rate over 10 steps. A promising solution? Embedding supervisory models or “eval agents” that guide decisions at each step, reducing compound error and improving reliability. 

2.     The Solo Bootstrapped Founder Era 

In a Gemini 2.5 demo, a presenter highlighted Carta data showing a recent shift in startup dynamics from 2023 and 2024: 

·       VC-backed single-founder startups dropped from 19% to 17% 

·       Non-VC-backed single-founder startups rose from 32% to 38%  

Notably, several boot-strapped founders I spoke with at the conference see capital raising more as a strategic lever to differentiate, consolidate, or outpace competition rather than a requirement to get to a Minimally Viable Product (MVP). It’s never been easier to launch but standing out and potentially exiting still requires capital and scale.   

3.     Model-First vs. Product-First? 

Should startups focus on building differentiated models or winning UX? Some argue the model is the product like how “Intel Inside” defined the computing market during the PC era. There’s also the view that foundational model companies are well positioned to capture more of the AI value stack, moving upstream via R&D or acquisitions. This echoes concerns from the cloud era (“What if AWS builds it?” now becomes “What if OpenAI builds it?”). Regardless, nimble product-first companies are finding success by tuning into what users truly want, not just building smarter models.    

4.     The Cloud Cost Challenge 

Cloud and AI adoption is surging, but so are costs, inefficiencies, and billing complexity. Some vendors pitched serverless architectures, while others showcased observability and cost optimization tools. Reducing resources spent on optimization deployments and managing spend are becoming just as important as productizing AI. 

5.     Vibe Coding & Future of Dev Work 

Google Cloud was a major sponsor of the conference and had a teach-in session on Gemini. I had my first crack at “vibe coding” a month ago using Google Firebase Studio and found the experience intuitive and rewarding. The OpenAI - Windsurf deal (just rumored at the time) got a lot of buzz. And also supported a point made in my Model-First vs. Product-First takeaway - "What if OpenAI builds it [or buys it]?” Interestingly, conversations around AI coding tools focused heavily on developer augmentation, not replacement - likely tailored to a crowd full of software engineers and data scientists.

Still, the real-world signs of workforce impact are showing: 

·       Salesforce reported ~30% productivity gains from AI tools and is slowing engineering hiring

·       Duolingo cited AI efficiency gains as a reason for cutting ~10% of contractor roles 

As I left the conference, I was struck by the sheer momentum and energy surrounding data and AI. This landscape is evolving at a breakneck pace. For founders and operators alike, the challenge now isn't just to be in the game, but to build with purpose and focus on what delivers meaningful, differentiated value. In today’s environment, M&A and capital raising have become more critical than ever as strategic levers to outpace the competition and capture market leadership in this space.

At Full Send Partners, we are actively engaged in the enterprise software, data and AI sectors. We offer strategic and informed guidance to our clients navigating this rapidly evolving landscape. With extensive experience in identifying and capitalizing on market opportunities, we are committed to helping businesses succeed in this essential and growing industry.

For a deeper conversation, please reach out to me at ggarg@fullsendib.com.

ABOUT FULL SEND PARTNERS

Full Send Partners is a Los Angeles-based middle-market investment bank that provides corporate, entrepreneur, and private equity clients with customized M&A, capital raising, buy-side advisory, and strategic alternatives. Combining Wall Street expertise with a boutique approach, Full Send Partners is dedicated to delivering exceptional advisory services and outcomes. To learn more, visit fullsendib.com.