After ten months of regulatory review, on Feb 27 2025 IBM closed its acquisition of HashiCorp for $6.4B (or $35/share in cash.) The RedMonk team has collectively been busy on the road, planning conferences, and writing books, but we still wanted to come together and share our thoughts on the merger. We’ll also share some initial reactions about the announcement of the intent to buy DataStax.
Below is a team Q&A with our thoughts about the deals.
Q0: Before we dive into our deal analysis, let’s give some background on HashiCorp.
Rachel: At a high level HashiCorp plays in the infrastructure automation and infrastructure management space, providing tools for provisioning, securing, connecting, and running cloud-native and hybrid environments.
Its products like Terraform, Vault, Consul, and Nomad, focus on Infrastructure as Code (IaC), secrets management, service networking, and workload orchestration.
Terraform and Vault are the products that garner the most attention and usage. From a revenue perspective, Vault is the most monetizable product in the HashiCorp portfolio; from a community perspective, Terraform is the heart of HashiCorp.
HashiCorp went public in December 2021 and struggled to find their financial footing in the market thereafter.
HashiCorp originally operated as an open core company before changing their license to the BSL in 2023.
If you want more details about IBM and HashiCorp at the time of the deal announcement, you should read this amazingly good piece of analysis by former Monk Fintan Ryan.
Q1: Why does this acquisition make sense for IBM? How does it benefit HashiCorp?
James: It hasn’t been easy for IBM to find its footing in recent years, but historically it has strong enterprise presence and when it comes to middleware, IBM pretty much invented the concept. And as of late, IBM is having a bit of a moment in the sun; it’s turning in good financial results and its share price is pretty good. Enterprises are willing to have conversations with IBM at the moment, pretty clearly; they’re interested in its story around AI. The Red Hat acquisition is working pretty well.
Adding HashiCorp as a string to the bow in its middleware portfolio makes sense.
From HashiCorp’s side, their financial results post-IPO were down; that’s one of the reasons it was on the block as a possible acquisition target. It was clear they were going to need a shot in the arm, and it is up to IBM to provide that boost and drive sales again. The alignment is good around IBM’s core strategic planks of AI and hybrid cloud. Infrastructure as Code is going to be absolutely foundational to hybrid cloud. We’ve talked about the hybrid cloud market for a long time, but because of geopolitics right now it is likely to be even more attractive in 2025.
Q2: How do you see the relationship between HashiCorp and Red Hat? In particular, how do you anticipate them finessing the portfolio overlap between Ansible and Terraform?
Kelly: Both Red Hat’s Ansible Automation Platform and Hashicorp’s Terraform play in the infrastructure automation space. Notably, both products have their roots in open source, with Red Hat AAP building on upstream OSS projects and Terraform previously available under an open source license prior its relicense in 2023 (see more from Steve O’Grady on this). While both Ansible and Terraform can be (and are) used for a wide range of automation objectives, Terraform uses a specific language and has strengths in areas like infrastructure provisioning and other IaC processes, whereas Ansible leverages YAML and shines in configuration management, software deployment, and orchestration.
Most organizations leverage multiple automation tools, with the Ansible/Terraform combination common enough that it has earned the nickname “Terrible” (or, alternatively, “Terrable” or “Ansiform”), and led to “better together” stories from Red Hat and Hashicorp. With IBM now owning both, many are curious to see how the “better together” story plays out vis-a-vis more intentional integration (or at least usage guidance) between the two.
Rachel: Even if we leave product overlap capabilities out of it, it feels like it will be a challenge for their sales and marketing teams to effectively communicate a nuanced shift to something like BSL licensing. Red Hat is a bastion of open source software, and BSL is a squishy in between.
Steve: Indeed. The BSL is a source available, non-compete license. The intent was to improve HashiCorp’s ability to monetize its previously open source assets by attempting to make the company the exclusive commercial supplier. In practice, however, the returns on this strategy – for HashiCorp specifically and the market more generally – are best described as mixed. Certainly the relicensing produced no abrupt change in direction from a financial returns standpoint.
The question facing IBM post-acquisition is whether to continue with this strategy or to revisit it given the difference in context between HashiCorp assets independent of IBM and now as part of its portfolio. To date, neither the acquiring or acquired company has given any indication of a change of heart in its license, and days after the acquisition has closed the projects’ various licenses are still BSL in public repositories. A fact which would seem to indicate that the status quo will persist.
Assuming it will for the sake of argument, that leaves questions to be answered.
How will IBM, and more problematically Red Hat, salespeople manage a third license type? To date, customer conversations were generally simple: software was either proprietary or it was open source. Adding HashiCorp assets to the portfolio makes those conversations – particularly with legal staffs and Open Source Program Offices (OSPO) – potentially more challenging.
What does this mean for forks such as OpenBao and OpenTofu? Up until the acquisition closed, some potential participants in the forks of Vault and Terraform respectively could have been waiting to see whether IBM, a company with a long history within open source itself and of managing the world’s largest pure play commercial open source software vendor might return the projects to their roots. With that option now seemingly off the table, does it create new incentive for participation in the project forks, or will they continue unimpacted?
Q3: What does this merger mean for competitors and the IaC market as a whole?
James: The combination of Terraform and Ansible is pretty clearly a roll up of assets in the IaC market. I suspect this market concentration was part of the reason why regulatory review took so long. One of the questions is how IBM’s strategic imperative around AI will impact IaC. Can IBM use AI to add substantive value to all of these automation tools, as opposed to AI being something that at times might begin to look almost a replacement to them?
Rachel: There are definitely market questions about we are headed in terms of declarative infrastructure, even without factoring AI in. In general we are hearing less and less about earlier generations of configuration tools like Chef, Puppet, and SaltStack. Things in the CNCF ecosystem like Crossplane and Pulumi and more visible. Then we see new competitors like System Initiative aggressively targeting HashiCorp install base. Ansible and Terraform are a powerful combination, but there are a lot of future-facing tools that are trying to replace them. Then, as noted, adding in AI muddies the water even further.
Q4: Just as the IBM and HashiCorp deal finally closed, IBM also announced its intent to acquire DataStax. Who is DataStax and why are they an interesting acquisition target for IBM?
Steve: DataStax is the primary commercial vendor behind Cassandra, a columnar database originally written by Facebook to manage its Inbox and later made available as an Apache Software Foundation (ASF) open source project. The open source Cassandra project has long has a reputation of being a highly scalable database, but also being difficult to stand up and operate, hence the commercial opportunity for a DataStax.
As with many of its commercial database vendor counterparts, DataStax has expanded aggressively from its original columnar roots over the years, embracing among others streaming and vector workloads. The latter, in combination with Langflow, is presumably one of the primary justifications for its acquisition by IBM.
Q5: Say more about the Langflow and the collision of databases and app platforms. Why is that of particular interest to IBM right now?
Steve: One of the things that has seemed apparent for some time is that application and database layers are likely to become less distinct. Compute, database and other primitives will clearly exist independently indefinitely, but the rise of abstract platforms like Fly.io, Netlify, Render, Vercel, et al following in Heroku’s original footsteps means that database and application platforms are increasingly comingled.
AI, for its part, seems likely to accelerate this trend. As large data platforms like Databricks and Snowflake invest heavily in models, and LLM tooling becomes more and more adept at making data accessible to all personas via natural language queries rather than just those familiar with SQL, the lines between data and analytical applications blur.
DataStax’ Langflow targets just this type of opportunity, with a low code tool designed to simply link models, APIs and databases. With IBM focused on bringing AI to enterprises with its watsonx tooling and Granite models, DataStax and Langflow make for a logical acquisition target.
Closing thoughts
Rachel: As with the Red Hat-IBM merger, we will be watching the specifics of the HashiCorp-IBM merger closely. Much of the success will ride on execution: how and where will the products be integrated into the company? what decisions will be made around licenses? how will sales teams be trained to talk about nuances of the portfolios? how are things like cultural mismatches or leadership changes handled?
Mergers are fickle and statistically difficult to pull off, even when all parties enter with a shared vision for success. The products in the HashiCorp portfolio play integral roles in the technical ecosystem, however, so there’s a lot riding on this particular transaction for both IBM and the market more broadly.
Disclaimer: Hashicorp, IBM, Red Hat, and Salesforce (Heroku) are RedMonk clients. CNCF, Databricks, DataStax, Fly.io, Netlify, Pulumi, Render, Snowflake and Vercel are not.
No Comments