Key Moments
How Two French Engineers In New York Built The Company That Monitors The Entire Cloud
Y CombinatorWant to know something specific about what's covered?
We've already dissected every moment. Ask and we will deliver (with timestamps).
Key Moments
Datadog CEO Olivier Pomel claims AI has enabled developers to rebuild entire systems in days, a task that previously took teams six months, significantly accelerating software development.
Key Insights
Datadog was rejected twice by Y Combinator, with co-founder Olivier Pomel still possessing the rejection email from Paul Graham.
Datadog initially supported cloud infrastructure because smaller, more modern companies were cloud-native, which proved to be a significant tailwind as the cloud market exploded.
Datadog's core bet was to bring DevSecOps together onto one platform, a concept that was not called 'observability' at the time and was initially framed as 'infrastructure monitoring' to gain market understanding.
Despite being based in New York, Datadog faced a difficult time fundraising, with VCs in both New York and the Bay Area initially unfamiliar with or dismissive of the infrastructure monitoring market outside of Silicon Valley.
Datadog's culture is not written down in values or principles; instead, it flows from the top through the actions and decisions of leadership, exemplified by their hiring, promotion, and firing practices.
The company now has around 20-25 products, expanding its portfolio based on customer usage patterns and extensions built around their core platform, alongside more strategic, forward-looking projects.
Pomel believes that moving faster, particularly in hiring and firing, is crucial for startup success; it's better to hire quickly, even if it means needing to fire later, than to be paralyzed by searching for the 'perfect' hire.
AI accelerates development at an unprecedented pace
Datadog's co-founder, Olivier Pomel, revealed a significant shift driven by AI: "in two quarters, we're not writing any code anymore." This statement came with a striking observation that highly skilled developers, not beginners, could rebuild entire systems in a matter of days. These accomplishments previously would have required a team of six engineers working for six months. This dramatic acceleration underscores how AI is fundamentally transforming the development lifecycle, enabling much faster iteration and problem-solving. The implication is that companies need to adapt to this new reality, embracing automation and revised workflows to harness this efficiency. The challenge now is not just about writing code, but about strategically automating and adapting to AI-driven capabilities, where the focus shifts from frequent code writing to more strategic automation, albeit with the understanding that some code will still need to be written and rewritten.
Navigating early rejections and market skepticism
Datadog's journey began with significant hurdles, including two rejections from Y Combinator. Pomel humorously recounts having a rejection email from Paul Graham, which served as motivation to prove them wrong. The company also faced a tough fundraising environment, with VCs in New York and the Bay Area initially struggling to understand the infrastructure monitoring market, especially as Datadog was not based in the Bay Area. This skepticism, however, fueled a desire to build differently and better than established Bay Area companies.
The foundational bet on cloud and DevSecOps
Datadog's core premise, established in 2010, was to unite DevSecOps onto a single platform. This was a significant departure from the existing 'monitoring' tools, which were siloed, job-specific, and reactive, largely used only by operations teams. Developers were kept in the dark about production issues. The company's initial infrastructure support focused on cloud technologies because smaller, more modern companies were adopting it. This strategic alignment with the burgeoning cloud market proved to be a massive tailwind, even though the founders initially underestimated the cloud's explosive growth. The product was initially labeled 'infrastructure monitoring' to make its value proposition clear to customers and their managers, rather than a more abstract 'data platform for DevSecOps collaboration'.
Culture as lived experience, not written words
Datadog deliberately avoids codifying its culture in written values or principles. Instead, Pomel emphasizes that culture flows from leadership and is demonstrated through actions. This includes how people are hired, promoted, and, crucially, fired. The leadership team exemplifies the desired behaviors, expecting employees to understand and embody the company's ethos without needing a posted list of rules. This approach is pragmatic; while they re-evaluate annually if formal documentation is needed to train thousands of employees, the current model of leadership emulation has proven effective.
Staying grounded through direct customer and employee feedback
Despite Datadog's scale of 8,000 employees, Pomel and much of the management team remain deeply involved in product decisions and stay connected to raw customer feedback. This includes reading support tickets, sales transcripts, and employee survey comments. By sampling this ground-level information, leadership gains a realistic understanding of the company's operational 'fabric,' avoiding the polished, often misleading, summaries that can filter up through management layers. Pomel handles feedback by asking clarifying questions, which prompts managers to investigate issues within their own span of control.
Portfolio expansion and strategic bets in a dynamic market
Datadog now offers around 20-25 products. Portfolio expansion is driven by observing how customers use and extend the platform, uncovering their underlying problems. Additionally, the company makes 'strategic' bets on emerging trends, even when no immediate customer demand is apparent. The rapid pace of change, particularly with AI, necessitates being more proactive and comfortable with a higher rate of being wrong. This reactive strategy involves taking more 'shots' and iterating quickly, as predicting product needs a year or more in advance is becoming impossible.
Navigating the public markets and compensation challenges
Going public in 2019 presented challenges, including a significant stock drop on the day of COVID-19 lockdowns. While the company has experienced market volatility, Pomel notes that leadership behavior hasn't drastically changed. The primary impact is on compensation, as fluctuations in stock price affect Restricted Stock Units (RSUs). The main risk is not company survival but retaining key performers whose compensation is tied to stock value. This requires careful compensation management, considering that employees join at different times with varying RSU grants.
The urgency of rapid hiring and firing
The hardest lesson Pomel continually learns is the need to move faster, particularly in hiring and firing. He advocates for being willing to hire quickly the first person who looks promising, understanding that if it doesn't work out, they can be fired. This iterative approach, involving more cycles and chances, is preferred over being paralyzed by the search for a perfect hire, which founders often make initially. Similarly, firing someone for the first time is difficult, but employees often perceive it as overdue rather than harsh, highlighting the importance of timely decisions.
Mentioned in This Episode
●Companies
●Organizations
●Concepts
●People Referenced
Company Size and Valuation Milestones
Data extracted from this episode
| Event | Size/Valuation | Timestamp |
|---|---|---|
| Left previous company | 800 people | 289 seconds |
| Went public | Valued at $10 billion initially | 966 seconds |
| Post-COVID lockdown crash | Valuation dropped to approx. $4 billion | 996 seconds |
| Current company size | 8,000 employees | 619 seconds |
Common Questions
Olivier and his co-founder Alexy met at Centrale, an engineering school in Paris. They later reconnected and worked together at IBM Research in New York.
Topics
Mentioned in this video
A startup accelerator that Olivier and his co-founder applied to but were initially rejected from. This rejection served as a significant motivator.
The company founded by Olivier and Alexy, which provides monitoring and analytics for cloud infrastructure, built on the principles of DevSecOps.
Mentioned as a more 'sexier' place for research compared to IBM Research in the current landscape.
Mentioned as a more 'sexier' place for research compared to IBM Research in the current landscape.
Mentioned as the 'Amazon thing' that many companies initially dismissed as a 'toy' but which eventually grew into a major cloud provider.
A rapidly growing segment of Datadog's customer base, requiring advanced observability solutions to manage their complex infrastructure and rapid development cycles.
Software as a Service, a business model used by the educational software startup where Olivier and Alexy worked, which was successful and led them to Datadog.
A combined approach for security and operations in software development, which was central to Datadog's platform and vision.
Restricted Stock Units, a form of employee compensation that became a concern for Datadog when the stock price dropped, potentially impacting key personnel retention.
More from Y Combinator
View all 607 summaries
45 minOpencode CEO: Blocked, 20X Growth in 6 Months, Building the Coding Agent for the World
23 minThe Model-Agnostic AI Platform Betting That No Single Lab Will Win
29 minWhy Ambitious Startup Ideas Are Actually Easier To Sell
75 minThe Key Thing Human Brains Have That AI Is Trying To Learn
Ask anything from this episode.
Save it, chat with it, and connect it to Claude or ChatGPT. Get cited answers from the actual content — and build your own knowledge base of every podcast and video you care about.
Get Started Free