If you’re a software company with a flimsy value proposition, then AI is a threat because it allows others to build something better and deploy it quickly. But for companies that have business models that process lots of proprietary data in an automated fashion, AI allows them to do all kinds of exciting things with it. That’s where Datadog $DDOG sits with their observability platform that is clearly benefiting from the growth of artificial intelligence. Last time we checked in with Datadog, they were projecting full-year revenue of about $3.2 billion, an increase of “just” 18% to 19% year-over-year. Instead, they delivered 2025 revenue growth of 28%.
Now they’re telling us again to expect 19% for the coming year which would still be impressive given they’re starting from a larger base. Perhaps what’s most notable is that a meaningful amount of that growth surprise is said to be coming from artificial intelligence.
Datadog Has Room For Growth
In their latest call, Datadog said they “continue to believe revenue is a better indication of our business trends than billing and RPO,” and we couldn’t agree more. For a SaaS company, increasing revenues comes from selling more to existing clients and adding new clients. Despite having 32,700 customers, Datadog still describes new logo additions as “very strong” with just 7% of the total number of potential clients captured. Even more impressive are the increasing revenue commitments being made by their largest customers.
We’re not given average contract duration, but that’s typically one to three years for SaaS firms. So when Datadog says, “We signed 18 deals over $10 million in TCV this quarter, of which 2 were over $100 million and 1 was an 8-figure land with a leading AI model company,” that implies strong demand for the services they provide, which is a leading indicator for revenue growth.
It also implies there’s loads of upside from existing clients alone, especially the larger ones. Around 90% of Datadog’s revenues come from clients spending more than $100,000 a year which means they’re firmly entrenched in enterprise sales as opposed to people paying for a few seats with a credit card. And get this. While 48% of the Fortune 500 are customers, they’re still spending under $500,000 at the median which suggests considerable upside from existing large accounts.
While Datadog’s net retention rate is an elusive number to capture, the latest earnings call talks about how “revenue retention percentage was about 120%, similar to last quarter.” That’s considered average for a quality SaaS firm, and it’s likely a good chunk of that new spend might be (or will certainly soon be, according to management) from upselling existing customers lots of cool AI stuff.
Datadog for AI and AI for Datadog
As we said in our “AI vs. SaaS” video, there is no better time to be selling AI software to C-level types who want to tell shareholders they’re “doing AI.” So it’s not enough to be talking about agentic AI solutions you’re working on (cough, UiPath, cough). You need to be actually selling AI functionality to clients. Bonus points if you develop and deploy something most AI companies need. Datadog has managed to do both.
“AI for Datadog” is about the AI-powered features they’re building and deploying, which should result in more revenues from existing clients. They talk about the Bits AI Dev agent, which detects code-level issues and generates fixes in production context. Additionally, their Bits AI Security Agent can help release and monitor a fix, providing complex troubleshooting as needed. This bot can autonomously conduct investigations and deliver recommendations based on Security Information and Event Management (SIEM) signals, which are simply alerts that Datadog’s software spits out when it encounters anomalies in a customer’s data exhaust.
“Datadog for AI” is the unique product offering they sell to AI companies, in particular, “capabilities that deliver end-to-end observability and security across the AI stack.” Last quarter, they landed “one of the largest AI foundational model companies” which had a fragmented observability stack comprised of more than five open source, commercial, and in-house observability tools. All of those patched-together solutions were brought under the unified Datadog platform which means the customer is now more productive and can scale with ease.
Datadog says they have 650 “AI-native customers” which are basically AI startups that form a cohort, which is significantly outpacing the rest of their business. Around 19 of these customers are spending $1 million or more annually with Datadog and “14 of the top 20 AI-native companies are Datadog customers,” which implies most notable AI companies are using their solution as they scale. And guess who has more money than they know what to do with right now? That’s right, AI companies, which now account for 11% of total revenues and growing quickly.
Datadog has also developed their own small observability LLM internally called Toto, which was trained on 750 billion data points unique to the company. It costs $750K to train compared to billions for leading frontier AI companies and is being fed trillions of new events per hour.
A software-as-a–service (SaaS) company for which AI is an opportunity, not a threat, ought to command a price premium. Even with all the AI excitement in the air, Datadog has a valuation that’s rich but not excessive. And lately, it’s been on the decline which is great news for people who like buying quality SaaS names at discounted prices.
Datadog’s Valuation Reverts to the Mean
Last year, we commented on Datadog’s rich valuation, with an average trailing four-quarter simple valuation ratio (SVR) of roughly 16. That average has now fallen to a more modest 14, with their current SVR clocking in at just 11.5. This means that despite Datadog’s impressive execution and persistently strong revenue growth, shares don’t appear expensive relative to past valuations. In fact, they’ve only gotten cheaper. Why?
The AI sword has two edges. While Datadog is seeing healthy demand for their own AI tools, they also face their fair share of headwinds from competing AI companies. The trio of Great Danes in the room would be the hyperscalers, specifically Amazon Web Services (AWS), Google Cloud, and Microsoft Azure. These three companies have begun to implement their own observability and AI monitoring tools onto their platforms. For example, AWS offers Amazon CloudWatch which comes built-in to over 120 AWS services. With CloudWatch, you can collect and aggregate various data metrics and analyze them for anomalies – basically, exactly what Datadog’s flagship software does. Azure Monitor and Google Cloud Operations Suite provide similar services.
In order to compete with the big dogs, Datadog needs to show that its third-party software can add value above and beyond these built-in tools. Our intern Tidder polled the developer community and found that the majority prefer Datadog over CloudWatch, thanks to its ease of use and enhanced features such as custom “health checks.” Datadog also has the added benefit of supporting “multi-cloud” operations, meaning it can be used on AWS, Azure, and Google Cloud simultaneously. That bodes well since most companies are moving towards utilizing multiple clouds.
There’s also the “AI eats software” narrative that’s been running rampant lately – basically, that AI agents will slowly start to replace humans. This will cause enterprises to drop seats from their software subscriptions, or abandon them entirely in favor of advanced AI models like Anthropic’s Claude. While Datadog is largely usage based, not seat based, it’s not totally shielded from disruption fears. For example, Datadog is using AI to highlight issues and suggest fixes in code. Why can’t any AI agent – not just the ones being built by Datadog – fix these issues proactively?
We can speculate until the Datacows come home, but until we see weakness in Datadog’s customer count and revenue growth, what management says seems credible. AI is more of an advantage than a threat.
Conclusion
With all the fear mongering around AI replacing SaaS platforms, Datadog stands out as a company which occupies a position of strength. Their highly-automated platform pulls data from over 1,000 integrations and compiles it into a single source of truth. Trillions of data points are added each hour and then fed to AI algorithms to make sense of. Now they’re moving towards agents that take action when needed. The addition of security in their stack creates stickiness and lends itself to vendor consolidation initiatives.
If you’re an AI-native company of any importance, it’s more likely than not you use Datadog for everything from observability to security. As always, the proof is in the revenue growth, and Datadog has that in spades. Let’s hope they can blow this year’s numbers out of the water and keep the acceleration going.