Welcome to the Chaos
May 2, 2024

HashiCorp Under IBM’s Wing

HashiCorp Under IBM’s Wing

Ned and Chris discuss IBM's acquisition of HashiCorp and the implications for the tech and DevOps communities.

IBM’s $6.4 Billion Acquisition 

In this episode of Chaos Lever, Ned and Chris discuss IBM’s significant acquisition of HashiCorp, a deal valued at $6.4 billion. They look into the implications of this merger for the tech community, oversee HashiCorp's journey from Vagrant to the role of Terraform in infrastructure as code, and speculate on the future integration within IBM's ecosystem. This episode also covers broader tech trends, mergers, and the impact of corporate culture on acquired entities!



Links: 

Transcript

00:00:00
Ned: I spent 45 minutes today messing around with my audio and camera setup.


00:00:06
Chris: How’d that work out for you?


00:00:07
Ned: Oh God, I hate everything [laugh]. Time is relative and not real. That’s what I’ve determined.


00:00:16
Chris: Time is relative to how it can disappoint us all.


00:00:19
Ned: Oh, that is so true. Hello, alleged human, and welcome to the Chaos Lever podcast. My name is Ned, and I’m definitely not a robot. I am a real human person who is not Hoggle, or Sarah, or Ludo, or that weird dog thing. No one’s that thing. No one knows what the hell it is either. Is it supposed to be a fox? With me is Chris. Is that a fox, Chris [laugh]? I don’t understand.


00:00:52
Chris: You watch this movie every goddamn day, don’t you?


00:00:55
Ned: [laugh]. Well, if I could just figure out the mystery of that weird fox thing that rides a dog, we could all move on.


00:01:03
Chris: But what does the fox say?


00:01:05
Ned: He says a lot in that movie, to be honest.


00:01:07
Chris: Actually, that’s true.


00:01:08
Ned: “Tallyho,” [laugh] is probably my favorite [laugh]. Oh, God. Did you try spinning three glass globes in your hand? Like, try to figure out how that works.


00:01:19
Chris: I cannot confirm nor deny.


00:01:21
Ned: You did it with pool balls. I know you did. Because everyone did. I got really good at two. I could never really get three. And one on top? Just forget about it.


00:01:30
Chris: Yeah, that’s clearly science fiction.


00:01:32
Ned: It was definitely CGI. That’s what I’m going to say, and no one can convince me otherwise.


00:01:37
Chris: I wouldn’t even bother trying.


00:01:38
Ned: [sigh]. It’s usually the best course of action when it comes to me.


00:01:41
Chris: Mostly because it’s exhausting talking to you.


00:01:45
Ned: [laugh]. Well, why don’t I just talk at you for the next half an hour? How’s that sound?


00:01:48
Chris: Much better, in fact. I’m going to go have a bit of a lie-down and just let you have the stage.


00:01:52
Ned: [laugh]. Oh, well, for folks that follow any kind of technology news. It’s probably completely unavoidable that they’ve come across the story that IBM, Big Blue, is acquiring HashiCorp.


00:02:04
Chris: [ba ba bah-nah… baaaww].


00:02:08
Ned: That’s the one. You’ve been watching too much Climate Town. Anyway, on April 24, it was announced that IBM has entered into an agreement to acquire HashiCorp for $6.4 billion. 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 some kind of activist investor. This thing is actually happening.


00:02:41
Chris: TL;DR?


00:02:42
Ned: Sigh [laugh]. Much longer—[laugh] so you can just stop now if you don’t care about this, but I care, and I have some feelings and some thoughts, and Chris, they’re not all negative. So, don’t flee in terror. It’s not going to be a 30-minute screed on the evils of IBM. [laugh]. I mean, there is going to be a little bit of that, let’s be honest. But you know, I’m hoping to bring some nuance to the discussion as well as indulge in a little catharsis. I went to Reddit to see what they had to say about it—


00:03:13
Chris: There’s your first mistake.


00:03:14
Ned: I backed away slowly. No sudden moves. So, I thought I would set the stage with my history with HashiCorp, and why I do have strong feelings about this, and why you might not and that’s fine. I’ve been working with HashiCorp’s products since 2014, roughly. That’s when I discovered their product called Vagrant. Chris, have you used Vagrant in the past?


00:03:40
Chris: No, but I’ve been accused of being one.


00:03:42
Ned: Haven’t we all? A listless layabout, actually was, I think, what somebody called me.


00:03:47
Chris: That’s probably the name of an Ubuntu release.


00:03:51
Ned: [laugh]. Probably. Listless Lollygagger would be my guess. Anyhow, this was before Docker and containers were really ascendant, and I was responsible for managing the lab environment at work. A thankless job, and I don’t know how I ended up with that. But I really wanted a way for my coworkers 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, given that the salespeople had already sold it. Yeah. Could you really use AD certificate services to automatically enroll devices running RHEL 5 leveraging SCEP on a trusted network? What about testing out Microsoft DirectAccess using 6to4 Teredo tunneling? Inquiring minds want to know. Those are two real things that happened.


00:04:47
Chris: Yes, of course I know that they are real things.


00:04:50
Ned: You were involved in some of this.


00:04:52
Chris: Details.


00:04:54
Ned: Ah, fair enough. Anyhow, provisioning machines, running the right operating system, and the right software, and configuring it was all a huge pain, and when I found out about Vagrant, I was very pumped. It essentially allowed you to take vanilla OS images that were pre-configured with the correct software, and then build out a whole testing environment in, like, say an hour instead of a couple of days. I even ended up paying for the Vagrant license so I could push the images to our VMware environment instead of running them locally because at a certain point, I did run out of RAM. Nowadays, ten years later—good God, ten years la—yes. You’d just use cloud for most of this testing. But remember, this was 2014, and cloud was still very much in its infancy. Ask me about Azure Service Manager interface sometime if you want to see me visibly cringe. HashiCorp had created a thing, and that thing solved a major problem for me. That alone made them memorable as a company. So, when I stumbled across Terraform a couple of years later, I was immediately interested, in part because I was intrigued by the idea of a cloud-agnostic Infrastructure as Code solution, I’d just spent two months working exclusively in JSON for CloudFormation—that was fun—and also, in part because it came from HashiCorp, a company that already had 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’s basically the de facto tool for infrastructure provisioning across the entire DevOps landscape. Terraform also made me a ton of money, but not in the way you might think.


00:06:37
Chris: Embezzlement?


00:06:38
Ned: I mean, later, maybe. Not right now. I’m kind of full from lunch. I was so enamored with Terraform that I used it for my Pluralsight author audition. You have to do an audition video that’s, like, five to ten minutes of you teaching how to do something. Not only did I pass the audition, but they asked me to develop a whole getting started course around Terraform. Now, despite not actually having that much experience with Terraform at the time—I’d been using it for, like, six to nine months—I agreed to develop the course because Fortune favors the bold and ignorant. And I’m really glad I did. My Terraform getting started course has consistently been in the top 30 courses on Pluralsight since I published it in 2017. I get paid on hours viewed, so um… it’s made me a lot of money.


00:07:29
Chris: And it’s never been updated since.


00:07:32
Ned: Oh, you know that’s not true. I did actually learn a lot more about Terraform over the next seven years, and the course has been updated four times now, I think? So, yeah. I mean, it’s not that I haven’t done any work on it, but it has definitely paid me back. I also created a study guide for the Terraform Associate Certification, I created Terraform-focused content for Manning Publishing, and I also deliver live trainings for Terraform for HashiCorp Academy. These are all things that make me money.


00:08:01
Chris: Which is good.


00:08:02
Ned: Generally.


00:08:03
Chris: Good. Yes, yes, yes.


00:08:04
Ned: The most popular content on my YouTube channel is, of course, the 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.


00:08:20
Chris: So, finally we know who to blame.


00:08:23
Ned: [laugh]. It’s true. 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 HashiCorp, and I’ve maintained those relationships even after some of them have left the company. I’m not a HashiCorp employee, but I’m probably as close as one can reasonably get without receiving a direct salary from them. The reason I’m bringing all this up is to set some context around my feelings regarding the recent HashiCorp drama. We already did a whole episode about the switch from open-source to source-available licensing, and I’m not going to relitigate my feelings about that. Go listen to that episode if you care. They also issued a cease-and-desist order to OpenTofu. I’m not going to comment on that other than to say it was a bad idea. But what about the news that IBM was acquiring HashiCorp? That stirred up a lot of swirling emotions. I am deeply invested in HashiCorp from both a personal and professional perspective, and that’s the context I feel like we need before we move forward. Good? Questions?


00:09:31
Chris: No, I haven’t been listening. So far you’re doing great, I’m sure.


00:09:33
Ned: Okay. Thumbs up? You can go back to sleep [laugh]. IBM. Not so much. I got a lot of warm fuzzies about HashiCorp, and you might reasonably ask, “Ned, how do you feel about IBM?” Well… slightly less warm and fuzzy would probably be my response to that.


00:09:52
Chris: It’s all been downhill since PC DOS.


00:09:54
Ned: Mmm. Isn’t that the truth? My general interactions with IBM had been minimal at best, and not overwhelmingly great. The very first job I had in technology was working help desk for retail. And we had an AS/400 in the data center that did all kinds of logistics, and also crunched the numbers for the store sales. Big [mamajam 00:10:18] of a beast sitting in the middle of the data center floor. I interacted with it through a terminal emulator that was a green screen when you brought it up, and the whole thing felt arcane, sluggish, and obtuse. Well, kind of describes IBM, doesn’t it? Anyway. Then I went into consulting—well, at some point [laugh] later I went into consulting—and I worked on a data center migration project that IBM won based on their SoftLayer acquisition. Chris, I think you may have been somewhat involved in this as well. The IBM salespeople made a lot of big promises about what SoftLayer could do, and most of it was bullshit.


00:11:00
Chris: Well, they use the word ‘could.’


00:11:02
Ned: In theory.


00:11:04
Chris: They never said ‘would.’


00:11:06
Ned: If we spent R&D dollars and had five years, it could do all of this.


00:11:11
Chris: If this was a completely different product.


00:11:13
Ned: [laugh]. That’s the one. The lead on the project, which thankfully was neither of us, spent weeks of their life, basically yelling at SoftLayer and IBM on conference calls, then finding weird workarounds, and just generally being stressed out about it. It wasn’t me, it wasn’t you, but we were adjacent to it, and we had to listen to it. After that, once I quit being a consultant, I was an analyst for a little while doing benchmarking and product evaluation, and I had to put OpenShift and IBM Spectrum Protect Plus head to head against Portworx. It did not go well for IBM Spectrum Protect Plus. Despite the fact that they paid for the research to begin with. I don’t think that report ever actually saw the light of day.


00:12:01
Chris: Why?


00:12:03
Ned: In addition, working with the software, just getting a working copy of the software was a nightmare. I had to fill out some forms, create an account, log in, ask for a key, that went into some kind of work order queue, get a key back three days later, and then get the executable, which I didn’t have access to, and I had to put another order in the queue.


00:12:29
Chris: Sounds completely straightforward.


00:12:31
Ned: [sigh]. To get the Portworx software, I went on their website and I click the download link. That accurately describes the rest of my interaction with both products. So, 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. Not saying that’s accurate, not even saying it’s fair, it’s just my impression. Sorry, IBM. And I would have probably maintained that opinion forever, if not for Red Hat. And here’s where you’re allowed to have opinions, Chris.


00:13:08
Chris: I never got a Red Hat, and I’m still mad about it.


00:13:12
Ned: Aren’t we all? I always envisioned it to being a red Fedora because of Red Hat, but I believe the original Red Hat was a baseball cap.


00:13:21
Chris: Well, that’s what they give you. They never actually gave out Fedoras unless you are an employee, I think.


00:13:26
Ned: Yeah. So, that’s weird, but not as weird as IBM acquiring Red Hat, which they did in 2019 for $34 billion. I know that doesn’t seem right, but it’s been five years [sigh]. Time, man [laugh]. No, thanks. In those five years, IBM has mostly let Red Hat just 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 and a lot of big enterprises. And while I think we’ve dedicated an entire show to the whole Santos debacle—I’m pretty sure we did—


00:14:07
Chris: We did.


00:14:08
Ned: And the placement of Red Hat source code behind a sign-in wall, the truth is that IBM has not smothered Red Hat. Quite the opposite. They’ve kind of embraced them. That’s because buying Red Hat was a huge bet for IBM. They had been losing money for something like 22 or 23 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 seriously suspect financial math they’d been using. The revenue was $57.7 billion in 2019. And in 2023, it was $61.8. Now, that’s not gangbuster big growth numbers like something we see out of, I don’t know, AWS, Microsoft, but they are respectable growth numbers, and they’re going in the right direction, which is up. Any thoughts on Red Hat, Chris?


00:15:13
Chris: No. I mean, we talked about a little bit of this. And actually, now that you say it this way, it might be time for a review episode of exactly what has happened. But generally, what people seem to think is that Red Hat infected IBM and took them over like—


00:15:30
Ned: Yes.


00:15:30
Chris: —a parasite.


00:15:32
Ned: [laugh].


00:15:32
Chris: Like one of the brain eating kinds. So, it’s like the IBM body is still lumbering around, but it’s Red Hat behind the controls.


00:15:40
Ned: That is a visual that I needed today.


00:15:42
Chris: You’re welcome.


00:15:44
Ned: Yes, I’ve heard it called Blue Hat now, is the name of IBM. So, IBM has not totally ruined Red Hat. That’s an encouraging sign. But $6.4 billion is not the same as $34 billion. It’s still material to IBM’s bottom line. What about some other acquisitions since 2019? How have those fared? I looked it up. IBM has acquired 19 companies since Red Hat, not including HashiCorp, since that hasn’t actually happened yet. Of those 19, only Apptio—and I’m guessing that’s how you say it—merited a disclosure of the price paid, meaning that was material to the company, and they bought that company for $4.6 billion. Chris, do you know who Apptio is or what they do?


00:16:32
Chris: Yeah, it’s FinOps stuff.


00:16:33
Ned: Is it?


00:16:34
Chris: Yeah.


00:16:34
Ned: My God, I had no idea [laugh]. I had to look this up. My guess is that they have been mostly folded into the IBM Cloud division and never thought about again.


00:16:45
Chris: You know, that famous and robust IBM Cloud.


00:16:49
Ned: I’ll get to that.


00:16:50
Chris: Okay.


00:16:51
Ned: So, HashiCorp is being acquired. How did this come to happen? What is their current state that allowed for this? There’s a really good article that I’ll link to from Fintan Ryan, and it’s on Medium. He’s an analyst and a reporter—or at least has been in the past—and he tracks this general area of technology. He pulled up the HashiCorp historical trends, and he asked the question, “Why is Hashi for sale?” And the answer basically comes down to anemic growth: 89% of their revenue comes from about 19% of their customers. That’s not uncommon—it’s the 80/20 rule, right—


00:17:30
Chris: Right.


00:17:31
Ned: However, those customers are largely US-based, so they don’t have a big international presence, growth has been slowing since they launched their IPO, which is bad. And when Hashi IPOed, they were valued at $14 billion. Based on revenue and growth, they can barely sustain their current $6.4 billion valuation that IBM is paying for them. In the eyes of the public market, HashiCorp has simply failed to maintain growth rates, and therefore investors are looking for an exit. On a personal note—and I think, Chris, you can get behind me on this one—I hate this. I don’t know why Hashi had to IPO in the first place. I mean, like, I do know, but I still hate it. Venture capital investors invest in startups with different rounds. They expect the startup to IPO or get acquired so that they can cash out, and even though nine out of ten startups fail, they’re counting on that one that succeeds wildly, and so they need a big multiple for that. So, they do everything they can to pump up the IPO before it goes public so that they can cash out. If the startups IPO, the investors expect a return on their purchase, so if the company doesn’t look like it can do it, then the investors who are now shareholders will force them into an acquisition, or private equity will stroll up and buy the company, load it up with debt, and sell, it in pieces if necessary. Modern business is just the worst, and capitalism was a mistake.


00:19:06
Chris: You forgot humanity is doomed. But otherwise, yes. Completely agree.


00:19:09
Ned: [laugh]. Okay. So basically, Hashi was getting pressure from investors to sell, and eventually that pressure was enough that they did that. Ah, being acquired by IBM is definitely not the worst case outcome, and it could actually end up being good. It took me a full day to come around to this opinion. It largely depends on where Hashi lands inside IBM’s organizational structure. When Red Hat was acquired, they basically became their own division. They just got, you know, “Okay, you’re Red Hat. You sit off on this side. You’re your own division with your own reporting structure, and we’re just going to take the money that comes out of you. You can leverage our sales and marketing a little bit, but that’s it.”


IBM spent a lot of money getting them, and just left them untouched. And in fact, as you sort of indicated, Chris, some of the Red Hat leadership found their way into IBM leadership. It’s worked out pretty well. But HashiCorp is not Red Hat, and the purchase price, while significant, is not $34 billion, so they’re not going to get carved out like Red Hat did, not getting their own division. So, there’s basically two places they can logically land in the organizational structure of IBM: in Red Hat, or IBM Cloud. Guess which one is bad.


So, the best case scenario is that they end up with the Red Hat group, and the Red Hatters treat them nicely, which they probably will. They’re friendly folks. Many of HashiCorp’s products are complimentary to things Red Hat has, think Ansible and Terraform: some people like to run them together. Also, integrating Vault and Console into OpenShift could be a benefit. And they’ve got some other cool stuff too. Landing in IBM Cloud is basically slow, painful death. I’m not going to sugarcoat that one. IBM Cloud has never invested the proper amount of CapEx to compete in any realistic way, even with the niche clouds out there. DigitalOcean, Linode, OVH, they all have much better stories when it comes to being a public cloud provider. As far as I can tell, the only companies that actually seem to use the IBM Cloud are in finance, and there’s only so many finance companies in the world. And even that’s a little shaky. This is not a growth opportunity for HashiCorp. They’re much better off hanging with the Red Hat crew, from a sales, marketing, and product standpoint.


00:21:40
Chris: Thoughts? Yeah, I mean, you’re going to—let me see—no, you’re not going to specifically go into it as much as I thought you were going to.


00:21:47
Ned: [laugh]. It depends on what you’re talking about.


00:21:50
Chris: No, in terms of… what happens next, the reality of this situation is that this HashiCorp… portfolio is very, very modern tools—DevOps, DevSecOps, whatever you want to call it—which does not fit with your, I think, accurate assumption or description of IBM as 40-year-old technology and 20-year-old ideas.


00:22:12
Ned: Yeah.


00:22:12
Chris: That’s honestly as simple as it has to—as you need to get with what’s the difference between being Red Hat-managed, and IBM managed? This is effectively bleeding-edge stuff. You know, there is an argument to be made that HashiCorp has overextended with the amount of products they have in the first place, which is also probably a problem.


00:22:30
Ned: Yeah.


00:22:31
Chris: Having said that, all of it is forward-looking, and it all ties tightly into Linux-type things, of which obviously, Red Hat is the leader.


00:22:42
Ned: [laugh]. They’re slightly enamored with it, yeah.


00:22:44
Chris: You know, sorry, Ubuntu, but for all of their frustrating choices, Red Hat Enterprise Linux is the leader.


00:22:52
Ned: Yeah.


00:22:53
Chris: So, that’s the best case scenario. If there is any wisdom in the IBM board, that will be the only scenario.


00:23:01
Ned: Agreed. I think, as you mentioned, the product portfolio at HashiCorp has grown a little bit larger than it needs to, and there’s going to be a rationalization of what’s in that portfolio. Exactly what gets cut, hard to say, but there’s definitely going to be some changes. Not everybody’s going to like them.


00:23:20
Chris: Yeah, and this is an opportunity, too, because I don’t necessarily mean that every single thing that they’re doing is bad, or is a bad idea, but it’s confusing and confusing means people don’t buy it.


00:23:32
Ned: Right.


00:23:32
Chris: So, now you have this opportunity, you grab all of these things—11 of them, I think; something like that—maybe seven of them stay fundamental and stay under the HashiCorp banner, the other four which still have promise, either get rolled into a different company product and the name disappears, or it stops being part of HashiCorp because people didn’t recognize it is that anyway.


00:23:55
Ned: Right.


00:23:55
Chris: But it doesn’t mean that the intellectual property is lost or all that work is just dumped down the toilet. This might give it an opportunity to sort of have a second life.


00:24:05
Ned: Yeah. And honestly, the majority of HashiCorp’s profit comes from running Vault, and then Terraform cloud stuff. That’s where they make the bulk of their money. So obviously, the big focus is going to be on those two products.


00:24:18
Chris: Well, and services consulting for $500 an hour.


00:24:22
Ned: Don’t sell yourself short there [laugh]. So, I want to talk about the licensing thing in a moment, but I just want to say that this is not the worst case scenario. If we take it as a given that Hashi couldn’t go it alone, then there’s a nonzero number of terrible private equity companies that would have absolutely eviscerated the HashiCorp product portfolio, jacked up prices by 200%, loaded the company up with debt, and then sold it off to someone worse.


00:24:50
Chris: True.


00:24:51
Ned: Yes. So, I’m glad that didn’t happen. It also could have been acquired by one of the big cloud providers and Hashi’s whole thing is being cloud-agnostic, so that would be a little bit like Intel buying VMware. Speaking of VMware, they could have been bought by Broadcom and folded into VMware. There’s probably an argument to be made for VMware having acquired HashiCorp a few years ago. I’m not going to try to make that argument, but it would have made sense at the time, maybe.


00:25:19
Chris: I could see it.


00:25:19
Ned: And of course, Oracle could have moseyed along, and since they already rely heavily on Terraform for OCI, and I think they actually use Vault as their secrets manager—though, check me on that—Oracle could have actually been maybe okay, but not really. Now, like I said, I wandered into Reddit and then backed away slowly. I also wandered in some of the LinkedIn comments, and those were more civil because it’s LinkedIn, but still, there’s a lot of speculation going on in regards to the BSL licensing change, when and why it was done, and what might happen in the future. First of all, I want to say that everything I’m about to say is pure conjecture and a lot of it was created by other people. The people who actually know the truth are absolutely not telling, and the simplest explanation is probably the correct one—Occam’s Razor and all that stuff. So, there’s one line of reasoning—and you can tell me your thoughts on this, Chris—one line of reasoning is that IBM requested HashiCorp to make the switch to BSL as a condition of sale. The theory being that IBM wanted that licensing change, but they didn’t want to take the flack for it.


00:26:32
Chris: Yeah, that’s absolutely what happened.


00:26:34
Ned: I don’t think so, but I have some reasons.


00:26:36
Chris: People have to remember the company IPO’ed $14 billion. That promptly fell into a well so deep that Lassie couldn’t help them. They IPO’ed at probably the worst time in economic history.


00:26:52
Ned: Yes. That did not help.


00:26:54
Chris: And they have been fighting that ever since. So, when it comes to the terms that they agreed to with this acquisition, IBM had them over a barrel. In fact, there’s a lot of people that are saying $6.4 billion is overpaying by a lot. The normal premium is 24%. This is like 38%.


00:27:12
Ned: Mm-hm. I don’t disagree that IBM had lots of leverage over HashiCorp, but I’ll also say that after the BSL thing happened, IBM went and created a Project OpenBao, which was a fork of Vault. It seems strange that they would go and help create a fork of a product that they were going to own in short order, and in fact, had pushed for that licensing change to begin with.


00:27:38
Chris: Well, it’s not OpenTofu though, is it?


00:27:41
Ned: I’ll talk about that soon enough. Anyway, so Red Hat has a history of taking open-core products and making them open-source. In fact, that’s kind of their whole thing. When they acquire a company that uses an open-core model, at some point in the future, they will release a cleaned-up version of that product that’s fully open-source. Once the acquisition is complete, all of the HashiCorp products might be put under the Red Hat umbrella, and then just naturally, the licensing that was BSL will be reverted to something completely open-source, and the closed portions of each product will be open-sourced. That’s another theory. Maybe? Yeah, and that’s forward-looking, and it’s all conjecture. No one who actually knows has commented. There’s also some theory that IBM is playing, like, 4D chess, and by making Hashi change the licensing, and then post-acquisition, reverting the licensing to make themselves look like heroes to the open-source community. I don’t give them that much credit [laugh].


00:28:45
Chris: No.


00:28:46
Ned: No, just stop, people. I’m sure there’s some other stupid wild theories out there. I’m already exhausted by all of them. The most likely answer is that HashiCorp knew they had to sell, and they thought that BSL might make them more attractive to potential suitors, and maybe that would also make them some more revenue in the meantime. It largely failed on the latter. They did not make more revenue, but it may have succeeded in the former by attracting IBM who may have already been interested. So, while I don’t believe that IBM asked them to make the change, I don’t think they minded it. Ultimately, whatever happens to HashiCorp, OpenTofu exists, and they forked it at a point that Terraform, it wasn’t feature complete, but it was good enough for most things. Even if HashiCorp ends up being completely destroyed, the tool that is used by the largest swath of the DevOps community has an alternative in OpenTofu, and I think that’s probably going to be a good thing. It’s very possible that Terraform and the other products will be reverted back to open-source licensing, in which case there will need to be some concessions made before Tofu gets merged back in, but that’s also a reasonable scenario, and it’s something that happened with Node.js, I believe, so there is precedent for this happening. Ultimately, after a couple days, I could sum it up by saying that I am cautiously optimistic. And I’m just leaving it at that.


00:30:31
Chris: Okay. I’ll allow it.


00:30:33
Ned: Hey, thanks for listening or something. I guess you found it worthwhile if you made it all the way to the end, so congratulations to you, friend. You accomplished something today. Now, you can go sit on the couch, read through copious notes on BSL, eat a chili dog, and just chill for the rest of the day. You earned it. You can find more about this show by visiting our LinkedIn page, just search ‘Chaos Lever,’ or go to our new website, pod.chaoslever.com where you’ll find show notes, blog posts, and general tomfoolery. We’ll be back next week to see what fresh hell is upon us. Ta-ta for now.


Ned: Did you know about the new website?


00:31:14
Chris: What’s the website?


00:31:15
Ned: Ehhh.


00:31:16
Chris: I’m not really a technology guy.


00:31:18
Ned: That’s true. You more of just, like, a, I don’t know, Fahrvergnügen kind of person?


00:31:24
Chris: I would have gone with Vienna Fingers, but allow whatever it is that you said.


00:31:29
Ned: Oh, you keep rolling, like Kid Rock.