Presenters
Source
The Data Layer: The Unshakeable Constant in the AI Revolution 🌐
The tech landscape is evolving at breakneck speed, and at the forefront of this transformation is Artificial Intelligence. But amidst the whirlwind of LLMs, SLMs, and AI agents, one fundamental element remains constant: the data layer. In a recent insightful conversation, CJ Desai, President and CEO of MongoDB, sat down with Harry Stebbings of 20VC to unpack the future of data, AI, and enterprise adoption.
Will Databases Become Commoditized? 🤔
The question on many minds is whether databases will become just another commoditized commodity in the coming years. While the conversation around commoditization often touches on compute, storage, and even large language models, CJ Desai firmly believes the data layer will resist this trend. He argues that the data layer has been the only constant since the inception of the tech industry, from mainframes to cloud transformations. It’s not only here to stay but also represents a continuously growing and non-commoditized market.
Evolving Perspectives on Data 💡
When asked about beliefs regarding data that have shifted in the last 12 months, CJ shared a personal journey. Starting his career at Oracle, he gained an understanding of traditional databases. However, his surprise came when he learned about the widespread adoption of MongoDB, not just by frontier model companies but also by large banks for mission-critical workloads like payments, and by innovative companies like 11 Labs. This challenged his initial thought that MongoDB might be confined to a “NoSQL” category, revealing its true potential for critical applications.
He now sees three constants in the AI world:
- LLMs/SLMs: These models are here to stay, regardless of their specific type or origin.
- The Data Layer: This foundational element is indispensable.
- Agent Layer: Whatever the precise definition, this component is also crucial.
Everything else, he suggests, is subject to change and re-evaluation.
Navigating the “Hollowing Out” of the Middle 🚀
An investor’s concern is the potential “hollowing out of the middle” – companies that aren’t hyper-growth AI giants but are still growing steadily. CJ’s advice for these companies is stark: transform or face challenges. Simply infusing AI or rebranding as an “AI company” won’t suffice. The path to sustained growth lies in either leveraging AI to create brand new, relevant products or fundamentally reinventing the company with AI at its core. Companies need to demonstrate significant acceleration in growth, akin to 11 Labs’ impressive trajectory, to thrive.
The Agentic Future: Selling to Machines, Not Just People 🤖
The conversation then pivoted to a fascinating prospect: a world where we build for and sell to agents, not just people. This implies a fundamental shift in UI and API requirements. CJ highlighted that even leading AI companies are finding it challenging to sell to enterprises, as they initially believed their AI products would “sell themselves.” The reality is that in the B2B context, people still buy. While agentic buying is a future possibility, the current reality, even in 2026, emphasizes human decision-making.
The criteria for an agent to “buy” will be drastically different, moving beyond traditional methods like Excel spreadsheets and Gartner reports. The complex layers of control and governance within enterprises, even for something as fundamental as database selection, mean that an agent’s recommendation will face significant hurdles before becoming an autonomous purchase.
Enterprise AI Adoption: The Regulatory Hurdle 🚧
For large enterprises, particularly in regulated industries like finance and healthcare, adoption and integration of AI face higher barriers. Data sovereignty and compliance requirements are paramount, meaning that simply presenting an AI product isn’t enough. Companies must demonstrate a deep understanding of how these solutions align with strict regulatory frameworks.
While there’s immense pressure for companies to adopt AI (“where’s my freaking AI?”), the willingness to lower adoption barriers depends on the stakes. For internal tools or productivity enhancements, adoption is faster. However, for client-facing applications with significant regulatory oversight, such as loan approvals, the barriers remain considerably higher.
The Rise of Forward-Deployed Engineers 👨💻
The skill gap in AI adoption is palpable. Many companies, even those on the bleeding edge, require assistance. This has led to the rise of “forward-deployed engineers” – teams that help enterprises build and integrate AI solutions. This model, pioneered by companies like Palantir and embraced by others, is crucial for delivering ROI and achieving desired outcomes quickly. The question remains whether selling to enterprises in the future will necessitate this FTE (full-time equivalent) model, akin to the customization services of the past.
The Human Touch in Code: AI’s Role in Development ✍️
When it comes to core infrastructure software like databases, relying solely on AI for code generation is currently not feasible due to the complexity and numerous edge cases. CJ stated that over 90% of MongoDB’s code is still human-written, though this is expected to change.
The impact of AI coding assistants is already being felt, significantly boosting engineer productivity. This increased efficiency offers two key benefits: either accelerating innovation velocity with the same team or maintaining innovation velocity while potentially reducing the need for a proportional increase in engineers as the company grows.
Unlimited Budget, Unlimited Innovation 💰
If granted an unlimited budget and no scrutiny, CJ’s focus would be on two core areas:
- Product Innovation: Expanding MongoDB’s platform into adjacent areas, leveraging AI to disrupt markets and build world-class products. This includes strategically entering competitors’ spaces where MongoDB holds an “unfair advantage.”
- Go-to-Market Evolution: Preparing for a future where agents buy software and investing in geographies that haven’t yet been prioritized due to public market pressures.
The Future of Pricing: Beyond Seats 🏷️
The traditional seat-based pricing model is becoming obsolete. CJ advocates strongly for consumption-based or outcome-based pricing. The pandemic era highlighted the pitfalls of overbuying based on seats, leading to higher churn. In the dynamic landscape of cloud and AI, where future needs are uncertain, consumption-based models offer greater flexibility and alignment with the value delivered.
Navigating the Uncharted Territory of Compute Capacity 🚀
A significant challenge emerging is the scarcity of compute capacity. Recent events, like major AI companies securing future data center deals, underscore this demand. CJ shared an anecdote of a Fortune 50 customer, a major hyperscaler user, being denied additional capacity in their desired regions, forcing them to move workloads back on-prem.
This situation highlights a critical takeaway: compute is the resource of the next decade. Both software vendors and customers must become more intelligent about forecasting capital expenditure and architectural strategies. The rise of frontier model companies as potential “next hyperscalers” also raises questions for software vendors about their architectural choices and potential vendor lock-in for customers.
Quickfire Round: Mumbai Indians and Boardroom Brilliance 🏏
In a fun quickfire round, CJ expressed a desire to sponsor the Mumbai Indians IPL team due to his roots and the massive developer community in India. For the MongoDB board, he seeks a current, technologist who can keep pace with the rapid advancements in AI and provide invaluable guidance.
The Biggest Pivot: Embracing AI at Infrastructure Speed ⚡
CJ’s most significant pivot in the last 12 months has been realizing that AI is table stakes for the engineering team, even at the infrastructure software level. He’s learned from customers that speed is paramount, and even infrastructure companies must move fast to enable the creation of AI-enabled agents and applications. The accountability and privilege of providing infrastructure that powers critical functions, from payments to nations, demands agility in this new era.
The Next 5-10 Years: Agents, AGI, and the Enduring Data Layer ✨
Looking ahead, CJ is incredibly optimistic about the AI revolution. He foresees an explosion of agents far beyond our current imagination, with the potential for Artificial General Intelligence (AGI) to become a reality. Throughout this transformative period, he reiterates his core belief: the data layer will remain the constant, presenting a massive opportunity for companies to leverage it to truly deliver on the promise of AI.