On April 24th it was announced that IBM has entered into an agreement to acquire HashiCorp for $6.4B. The all cash transaction will see HashiCorp’s outstanding shares purchased at $35 a share. 43% of outstanding common stockholders have already agreed to the acquisition, making it highly unlikely to be scuttled by activist investors. This thing is actually happening.
Okay, so I have some feelings and some thoughts. They’re not all negative, so don’t flee in terror. This isn’t going to be a 30 page screed on the evils of IBM. I’m hoping to bring some nuance to the discussion as well as indulge in a little catharsis.
My Hashi History
I’ve been working with Hashi products since 2014. That’s when I discovered Vagrant. This was before Docker and containers were really ascendant, and I was responsible for managing the lab environment at work. I really wanted a way for my coworkers and myself to be able to quickly spin up temporary environments for testing and learning. As consultants, we were constantly being tasked with learning how a particular product worked, or testing whether a given scenario was actually feasible. Could you really use AD certificate services to automatically enroll devices running RHEL 5 leveraging SCEP from a trusted network? Or what about testing out Microsoft's DirectAccess using 6to4 Teredo tunneling? Inquiring minds want to know.
Provisioning machines running the right OS and software was a huge pain, and when I found Vagrant, I was pretty pumped. You could pull boxes that were preconfigured with the correct software and build out a whole testing environment in an hour instead of a couple days. I even ended up paying for a Vagrant license, so I could push images up to our VMware environment instead of running them locally.
Nowadays, you’d just use the cloud for most of this testing. But remember, this was 2014 and cloud was still in its infancy. Ask me about the Azure Service Manager interface sometime if you want to see me visibly cringe.
HashiCorp had created a thing that solved a problem for me. That alone made them memorable. So when I stumbled across Terraform a couple years later, I was immediately interested. In part b/c I was intrigued by the idea of a cloud-agnostic IaC solution and in part b/c it came from HashiCorp. A company that had already earned my trust.
Needless to say, my trust was not misplaced. Terraform turned out to be a huge deal in the IaC space, so much so that it is the de facto tool for infrastructure provisioning across the DevOps landscape. Terraform also made me a ton of money, but not in the way you might think.
I was so enamored with Terraform that I used it for my Pluralsight author audition. Not only did I pass the audition, but they asked me to develop a whole getting started course around Terraform. Despite not actually having that much experience using Terraform- I’d probably been using it for about 9 months at that point- I agreed to develop the course.
And I’m glad I did. My "Terraform - Getting Started" course has consistently been in the top 30 courses on Pluralsight since it published in 2017. I get paid based on hours viewed, so um, it has made me a lot of money.
I also created a study guide for the Terraform Associate certification, created Terraform focused content for Manning Publishing, and I do live trainings on Terraform for HashiCorp Academy.
The most popular content on my YouTube channel is Terraform stuff, and I’ve attracted sponsors that are part of the larger Terraform automation community. Without HashiCorp and Terraform, there might not be a Ned in the Cloud LLC.
It’s not just the money though. I’ve been part of the HashiCorp Ambassador program since its inception in 2020. I’ve gotten to know a ton of people over at Hashi and maintained those relationships even after some of them have left the company. I’m not a Hashi employee, but I’m probably as close as one can reasonably get without receiving a direct salary.
The reason I’m bringing all this up is to set some context around my feelings regarding recent HashiCorp drama. We’ve already done a whole episode around the switch from open-source to source-available, and I’m not going to re-litigate my feelings about that. But the announcement that IBM was acquiring HashiCorp? That stirred up a lot of swirling emotions. I’m deeply invested in HashiCorp from both a personal and professional perspective. That’s the context I feel like I need to set up before we move forward.
IBM? Not So Much
Given all these warm and fuzzies about HashiCorp, you might reasonably ask me, “Ned, how do you feel about IBM?” Whelp, slightly less warm and fuzzy. My general interactions with IBM have been minimal and frustrating at best.
My first tech job was helpdesk for retail. We had an AS/400 in the data center. Big mama-jama of a beast plopped right in the center of the DC floor. I interacted with it through a terminal emulator. It felt arcane and sluggish.
When I went into consulting, I worked on a data center migration project that IBM won based on their SoftLayer acquisition. The IBM sales people made a lot of big promises about what SoftLayer could do, and most of it was bullshit. The lead on the project from our team spent weeks of their life yelling at SoftLayer and IBM, finding weird workarounds, and just generally being stressed out. It wasn’t me, but I was adjacent to it.
For a while I was an analyst doing benchmarking and product eval. I had to put OpenShift and IBM Spectrum Protect Plus head to head with Portworx. It did not go well for IBM Spectrum Protect Plus, despite the fact that they paid for the research. I don’t believe that report ever saw the light of day.
My general impression of IBM has always been of a bloated behemoth that maintains 40 year old hardware, has 20 year old ideas, and smothers everything it comes in contact with. I’m not saying that’s accurate or fair. It’s just my impression. Sorry (not sorry) to IBM.
I probably would have maintained that opinion, if not for Red Hat.
IBM’s Red Hat Acquisition
It’s been almost five years since IBM acquired Red Hat for $34B in 2019. I know that doesn’t seem right, but yes, five years.
And in those five years, IBM has mostly let Red Hat do their thing. Open source projects like Ansible are still thriving. Red Hat is still intensely active in the open source space, contributing to hundreds of projects. OpenShift has gained traction in lots of big enterprises. And while I think we’ve dedicated an entire show to the CentOS debacle and the placement of Red Hat source code behind a sign-in wall, the truth is that IBM hasn't smothered Red Hat, quite the opposite.
Buying Red Hat was a huge bet for IBM. They had been losing money for something like 22 straight quarters. They knew they had to do something different or be doomed to slowly wither on the vine. Red Hat was their big swing. And it's mostly panned out.
In 2019, they took a massive write down in revenue based on some hinky financial math they’d been engaging in. Revenue was $57.7B in 2019 and in 2023, it was $61.8B. These aren’t gangbuster numbers like we see coming out of Amazon or Microsoft, but they are respectable and going in the right direction.
Hashi’s Current State
There’s a really good article from Fintan Ryan on Medium. He pulled the HashiCorp historical trends and asked the question, why is Hashi for sale anyway? The answer is one of growth.
89% of their revenue comes from about 19% of their customers. Those customers are largely US based. And growth has been slowing since they launched their IPO. When Hashi IPO’d they were valued at $14B. Based on revenue and growth, they cannot even sustain their current $4.4B valuation. In the eyes of the public market, Hashi has simply failed to maintain growth rates and therefore investors are looking for an exit.
Best Case
Being acquired by IBM is not the worst outcome, and it could actually be good. It largely depends on where Hashi lands within the IBM structure. When Red Hat was acquired, they basically became their own division. IBM had spent a lot of money getting them, and chose to leave them largely untouched. Some of the Red Hat leadership even found their way into the IBM leadership structure.
But Hashi is not Red Hat and the purchase price (while significant) is not $34B. They’re not going to get carved out like Red Hat did. There’s basically two places they could land: Red Hat or IBM Cloud.
The best case scenario is that they end up within the Red Hat grouping and the Red Hatters treat them nicely. Many of Hashi’s products are complimentary to Red Hat, think Ansible and Terraform. Also, integrating Vault and Consul into OpenShift.
Landing in the IBM Cloud is basically a slow painful death. IBM Cloud has never invested the proper amount of CapEx to compete in any realistic way with even the niche clouds out there. Digital Ocean, Linode, OVH all have better stories to tell when it comes to cloud.
The only companies that seem to use IBM Cloud are in finance, and there’s only so many finance companies in the world. This is not a growth opportunity for Hashi. They’re much better off hanging with the Red Hat crew from a sales, marketing, and product standpoint
Stupid Rumors
There’s a lot of speculation going around in regards to the BSL licensing change and when and why it was done.
First of all, I want to say that ALL of this is conjecture. The people who know the truth are absolutely not telling, and the simplest explanation is probably the correct one.
One line of reasoning is that IBM requested that HashiCorp switch to BSL licensing as a condition of sale. The theory being that IBM wanted the licensing change, but didn’t want to take the flak for it.
Red Hat has a history of taking open-core products and making them open-source, so there’s also a theory that once the acquisition is complete, all the products will be put under the Red Hat umbrella. The licensing will be reverted and the closed-portions of each product will be open-sourced.
There’s also a theory that IBM is playing some kind of 4D chess by making Hashi change licensing, then reverting it back post-acquisition to seem like heroes of the OSS community.
I’d like to point out that IBM was one of the cofounders of the OpenBao project which is a fork of HashiCorp Vault. I don’t think they would have done that if they were behind the BSL change or playing any kind of 4D chess. Unless… no. Just stop.
I’m sure there’s some other stupid, wild theories out there, but I am already exhausted. The most likely answer? Hashi knew that had to sell. They thought BSL might make them more attractive to potential suitors. And maybe it would make them more revenue. It failed on the latter, but may have succeeded on the former.
Worst Case
If we take it as given that Hashi couldn’t go it alone. Then the possible worst cases were pretty bad. There’s a nonzero number of terrible private equity companies that would have absolutely eviscerated the Hashi product portfolio, jacked up prices by 200%, loaded the company up with debt, and then sold it off to someone worse.
It could have also been acquired by one of the big cloud providers. Hashi’s whole thing is being cloud-agnostic, so that would have been a little like Intel buying VMware.
Speaking of VMware, Hashi could have been bought by Broadcom and folded into VMware. There’s some kind of argument to be made for that, but I’m not going to try.
Oracle could have also moseyed along, since they heavily rely on Terraform and possibly Vault for OCI. Today’s Oracle actually could have been sorta okay, but not really.
Final Thoughts
IBM's purchase of HashiCorp is neither the best or worst case. I'm hesitantly hopeful that IBM will do the right thing regarding placement of Hashi within their organizational structure. I also expect a product rationalization to happen fairly swiftly, with some of the less popular projects being reverted back to open-source and donated to the community. Sorry Nomad and Waypoint.
Then there's OpenTofu, which I've barely touched on. If we take it as a given that Terraform is HashiCorp's most important contribution, then regardless of what happens with Hashi as a company, their legacy will live on in OpenTofu.
HashiCorp is a company made up of people. Some of whom I like very much. My final thought is that I wish them all the best and hope that this transition is beneficial to them personally and for the company at large. Mitchell and Armon built something pretty cool and I'm certain the Tao of HashiCorp will live on, even if the company does not.
