Last week, Snowflake, one of the largest data platform vendors, made people sit up with an interesting move to acquire Observe Inc., a small observability vendor founded in 2017. In July 2025, Observe raised a Series C, bringing its total funding to around $393 million, with Snowflake Ventures among the investors in several of these rounds. In the announcement, the key messages appear to be that the combined company will enable Agentic AI for faster troubleshooting for Snowflake customers and create an opportunity for Snowflake to enter the Observability space. And that is where the message starts getting muddy. I can see clearly how this acquisition can benefit Snowflake, but the PR messaging seems to be a bit all-inclusive, and the script is getting lost in the process. I will try to break this down for you here.
Observability is a crowded market
There are dime-a-dozen vendors in the observability space, and each faces its own challenges with market capture, customer retention, growth, and innovation. The established ones have a rich platform but have become super expensive. Some of the younger or smaller ones are narrowly focused on their offerings and seem to thrive in niche markets. As every vendor in the space claims they can deliver full-stack observability today, the real differentiators will come down to factors such as ease of use, alert noise reduction, lower MTTR, support for your choice of clouds/containers, and out-of-the-box integrations with various systems across your enterprise. Other factors, such as support for OpenTelemetry, faster search indexes, and longer, yet more cost-effective, log retention, are also critical when evaluating these vendors. Entering this market is easy, but finding traction and growing your customer base can be challenging if you don’t have real differentiators. With the subtle growth of OpenTelemetry adoption and the stupendous rise of cloud-native / SaaS applications on standard clouds, switching costs between observability vendors have started coming down. However, customer familiarity with tools and the general resistance to change mindset are still hindering IT Ops audiences from moving to different observability platforms as easily.
Observe’s customer growth and retention
Observe has been around since 2017, has over $393 million in funding, and yet has only 100+ customers. This shows the real struggle in this market when it comes to customer acquisition. They do have some big name customers but the market situation is real. I have seen a company like LogicMonitor stuck at 2200 customers for several years. While it may indicate lack of customer growth, it also shows that observability vendors tend to retain customers for a long time unless there is a serious security breach (like what happened to Solarwinds) or some incident like that which can cause customer distrust and eventually customer drain. So, the NRR numbers that Observe has aren’t really surprising. It is expected in this space.
Observe’s smart play
Observe did something really smart from the get-go. They developed their platform on/for Snowflake with native integrations. They also ensured support for open standards such as OTel and Apache Iceberg. They have a separate product called “Observe for Snowflake“, which is primarily a data/infrastructure observability platform meant for Snowflake workloads. This will allow the user to optimize performance and understand resource utilization (credits, hours, costs) on your Snowflake Cloud instances. You can monitor the health and performance of Snowflake Data Cloud, including query execution, Snowpark User-Defined Functions, Snowpark Container Services-deployed applications, ML/AI models, data pipelines, and underlying infrastructure. You can also debug and troubleshoot SPCS applications with a comprehensive set of telemetry signals.
Snowflake has developed a reputation in the industry for becoming very expensive over time. One of the biggest complaints that customers have (not just Snowflake customers) is being hit with a surprisingly high usage bill at the end of the month. This is due to the lack of visibility tools on these platforms that can help customers track and understand why and how they incur additional charges. Having such visibility can bring responsibility and ownership in user behavior. It can also protect the platform’s brand value. This is exactly why Snowflake should acquire Observe. This would make a ton of sense.
Observe said that they will operate as an independent unit for the foreseeable future. That again makes a lot of sense given their recent innovations with their AI SRE offering as well. However, now with the additional baggage of the “expensive” Snowflake brand, the question arises whether their pricing will change. Will Snowflake have to be part of every deal they make going forward? Selling their “Observe for Snowflake” offering to the extensive Snowflake customer base will be a strong sales play. If I were Snowflake, I would even offer this capability (at least a basic version of it) free to all of my customers to help them gain such visibility. They can always choose to buy premium features based on their needs.
Should Snowflake enter the observability space?
Entering the crowded observability market (primarily full-stack monitoring) doesn’t make much sense for a player like Snowflake for a few reasons. First, the primary use cases within observability around debugging and troubleshooting involve extremely fast ingests of real fresh data, indexing them for highly performant queries and performing highly effective pattern matching. Moving such an architecture into Snowflake makes no sense. Pushing telemetry data into Snowflake for analytics is an acceptable use case, but that is not what real-time observability is about.
Second, if the argument is about cost savings, particularly for long-term data retention, Snowflake hasn’t earned that reputation yet. Moreover, most of the observability platforms leverage S3 for such storage – cheap and effective. You cannot convince an enterprise out there that storing that in Snowflake will be cheaper than S3.
Full-stack observability is not really Snowflake’s core competency. Getting the petabytes of data eventually into Snowflake (either Iceberg or other) is a profitable business for sure, but the way to make it happen is very hard. However, the data observability angle is absolute gold and I firmly believe that Snowflake should pursue that instead.
Will we see more data platform vendors acquiring other observability vendors?
I really hope not. Based on the arguments above, I hope that if a data vendor acquires an observability vendor, it is primarily for data observability and cost monitoring. If they want all that juicy telemetry data in their data warehouse/lake/pool/puddle, they should just partner with observability vendors rather than acquiring them.