Last week HashiCorp went public, raising more than $1.2 billion and ending the week with a $15.3 billion valuation. Not bad for a company that gives away software for free (the vendor made more than $211 million last year from open source).
HashiCorp’s open source tools such as Terraform, Vagrant, and Consul enable enterprises to automate and manage their cloud infrastructure. Quite often, they are used more often than the homegrown infrastructure automation services the clouds offer.
This is one reason it’s worth wondering which of the big technology companies is going to regret it most when HashiCorp gets bought and not by them.
There are plenty of reasons to bet against HashiCorp getting bought, but also tens of millions of developer-driven reasons to suggest that it would be worth whatever it might cost. Like GitHub (acquired by Microsoft in 2018), HashiCorp could offer an on-ramp to a company’s cloud services.
Yes, I know that HashiCorp isn’t for sale. I also know that the company takes an emphatic multi-cloud stance. As to the latter concern, GitHub also supported multi-cloud (or rather it was agnostic about where its code repositories would run). As to the former, the day HashiCorp went public, it became available for a price.
Not that the HashiCorp team is looking to sell. If anything, they’ve spent a long time pushing away offers (like this early $50 million offer — a “flabbergasting amount of money,” as HashiCorp Cofounder Mitchell Hashimoto tells the story).
The trouble with this approach is that HashiCorp is becoming incredibly valuable to a number of different companies. From Amazon Web Services (AWS) to VMware, many companies are seeing that there’s a lot to love about owning the company that builds some of the industry’s most beloved developer tools. Yes, including AWS.
Although AWS rarely acquires anything in this sort of price range, and although AWS has its own mature tools (as former AWS serverless specialist Marek Kuczyński calls out), developers tend to default to HashiCorp’s Terraform for provisioning their cloud infrastructure.
Perhaps Kuczyński is right. Perhaps “Google and Microsoft could have [more] use for it as their [infrastructure as code] is still very immature. Terraform is a more pleasant alternative to their basic capabilities.” This isn’t wrong.
And while there would be a need to walk a fine line of independence, both Google Cloud (Kubernetes, for example) and Microsoft (GitHub, see below) have done this for years. For both companies, as they seek to close the market share gap with AWS, owning prime developer assets could help.
HashiCorp’s tools are popular in significant part because they are not owned by one of the clouds. They are independent. But a cloud could maintain that independence while still slowly building better on-ramps to their cloud services through tighter integrations or additional features.
Before you dismiss the notion, look at how Microsoft has stewarded GitHub and how that stewardship appears to be changing several years into the acquisition.
GitHub: Paving the way to Azure adoption
For years, some in the industry have been waiting for the other shoe to drop on Microsoft’s interest in GitHub. Three years in, that may be starting to happen.
To be clear, Microsoft has been a great steward for GitHub, largely leaving it independent even as the company increased budgets to drive innovation. Nat Friedman, widely trusted within the industry as an open source advocate (though a pragmatic one), helped ensure GitHub’s ongoing independence. Other executives, such as Erica Brescia (GitHub’s COO), also joined and helped to paint a picture of GitHub thriving in its operational independence within Microsoft.
Until it no longer was.
In early November 2021, Friedman announced that he was stepping down to go back to his start-up roots. He was replaced by Thomas Dohmke, GitHub’s chief product officer, who had previously spent nearly four years working in Microsoft product management.
Brescia had no tenure with Microsoft, and that may have hurt her possible selection as Friedman’s successor. In early December, Brescia announced that she would be leaving to become a VC.
Dohmke is probably great, but he will now report to Julia Liuson instead of Scott Guthrie (to whom Friedman reported). Liuson has been with Microsoft for nearly three decades and was just promoted right after overseeing one of Microsoft’s only real open source missteps in many years: the Hot Reload debacle.
As reported by Paul Krill, Microsoft recently removed the Hot Reload feature from .NET 6, making it available only through the proprietary Visual Studio. This marked the seeming continuation of a stream of decisions to prioritise Microsoft’s proprietary, paid products over open source alternatives, according to a leaked internal email.
Rather than speak to some nefarious grand design at Microsoft however, this probably just shows the internal battles the company is fighting. According to one commentator, “On one hand, they want to be seen as a new version of Microsoft who loves open source, but on the other hand, they want to actively block advances in OSS projects like the .NET SDK which could undermine their own commercial offerings.”
Others are not so sanguine. As Geoffrey Huntley argues, “GitHub is now a marketing proxy for old Microsoft tech. GitHub Actions is Azure DevOps [and] GitHub Codespaces is Azure Visual Studio.”
This might be overly harsh. Still, it’s not unreasonable for Microsoft to expect to derive considerable financial returns from its open source investments, including the $7.5 billion the company paid for GitHub. This isn’t because Microsoft is bad. It’s because it’s rational. Or selfish, as I’ve written. All open source is selfish (read: self-interested), and Microsoft’s open source efforts are no different.
This brings us back to GitHub. And HashiCorp. And on-ramps.
Microsoft has been more deeply integrating GitHub with Azure for years. It’s becoming ever easier to move code from GitHub into the cloud, especially Microsoft’s cloud. In like manner, though HashiCorp has a resolute, multi-cloud independence, so did GitHub.
Three years in, some of that independence may be wilting in light of the need to grow Microsoft’s business. By going public, HashiCorp put itself up for sale, even if that is not its goal. HashiCorp is perhaps the dominant infrastructure automation company, so there’s a lot at stake in keeping it independent — or in owning it and (eventually) paving on-ramps to one’s own cloud services.