Hashicorp gave away basically everything valuable they had, and charges extortionate prices for the few features they've paywalled. Only way they're going to survive is by becoming a feature of some bigger cloud offering.
We've been leaning into Terraform-based Crossplane providers recently and the out of the box experience is so much better than using terraform cloud once you get crossplane set up, I can't imagine ever going back to vanilla terraform pipelines, and certainly not paying for terraform cloud.
The high price of Hashicorp doomed them in the end. I don’t know what their financials looked like internally but the quotes we got from sales reps got them laughed out of the room.
I had the same experience some years ago. When we told them they were out of their minds and they wouldn't budge on price, the sales rep came back after a few months pleading to get us back to the table because we would have been a gigantic sale and a foothold into the other orgs of the large company I worked for at the time. We told them we already had a different solution in flight because of their original failure to negotiate.
Thirded. Some of their products looked useful so we talked to a sales person, and it was exorbitant. Not "oof, that's a little more than I'd budgeted for", but "LOL does that include equity?" The upside is that it pushed us to check out the alternatives.
AWS Secrets Manager isn't as nice as Vault in a lot of ways, but at a starting price of $0.40/mo for 1 secret + as many clients as you'd want vs $1250/mo for a standard vault + 1 client, it's fine.
I worked in the HashiCorp sales org for a couple of years -- Vault's weird identity-based pricing model was by far the biggest issue with getting prospects to buy Enterprise. It's great if you're a large multinational with highly-consolidated infrastructure and application identity, but if you're operating on a smaller scale than that or are not highly-centralized it's totally infeasible. The sales segment that focused on the smallest deals was rarely able to sell Vault Enterprise at all -- both because the pricing was so high and because the pricing model itself didn't scale in a way prospects felt they could understand and budget predictably for as they grew. We heard "we love the features, but there's no way we can afford that price" all the time. Various people beat the drum on this internally for literally years right from the initial switch to the current pricing model, but other than minor "client counting" tweaks, nothing ever changed.
I see they have a SaaS product called Vault Secrets now and I bet that's exactly why that product exists.
Okay but how much does that really say? All HN/startup advice says that when you sell to enterprises you have to charge enterprise prices. Meanwhile techies are notoriously for preferring to spend a month's worth of salary building an FTP server to replace Dropbox with.
Hashicorp's prices are not what i'd consider to be 'enterprise prices'. When we looked at buying "enterprise vault" a few years ago it was six figures a year for a relatively small install. It was basically "we don't want your business" pricing.
I mean, yes, it is a bit expensive. On the other hand, it also shows that it is nearly impossible to have an opensource product and be profitable at the same time, especially with dev/infra tooling like this.
> All HN/startup advice says that when you sell to enterprises you have to charge enterprise prices.
HN/founder culture is a bubble though.
And the original statement was Joel's and said you have to either charge something that the lowest level manager can approve on their own OR enterprise pricing. Basically you can charge $30 or $3000 but not $300.
I remember seeing quotes for Hashicorp Vault Enterprise, the price for Vault dwarfed the entire cloud bill of the company. And yet the cloud bill (through the marketplace) included other third-parties that were being paid an "enterprise price".
I can tell you that I worked at a large company that was moving pieces like identity and secrets management to SaaS or open source solutions years after rolling their own and failing to keep up. I'm well versed in built vs buy vs opensource.
For the record, Hashicorp vault was 10x the cost of our fully managed Identity system for a million+ unique logins a month.
Was there still people buying Vault at that price? From my perspective it looks like Hashicorp has been overplaying its hand there, but maybe it worked somehow!?
I've got a github actions pipeline that stands up a kind cluster, installs cluster-api on it, uses cluster-api to stand up the control plane cluster and then copies over the cluster-api config to it along with argocd, and after that it's self-managing. We did have to do a little bit of terraform to get iam roles set up, but that's it.
Everything after that is pure argo cd, cluster-api and crossplane.
I work at something that's basically a startup incubator and we can stand up and manage the infrastructure for an entire enterprise, on either aws or azure (including stuff like SSO artifact storage and CICD pipelines), from a central control plane with just a few lines of yaml, and they can be deploying code within 30 minutes. And it's continuously reconciled, so there's no drift, and if we push out an update to something, it rolls out to everything within minutes.
We've been leaning into Terraform-based Crossplane providers recently and the out of the box experience is so much better than using terraform cloud once you get crossplane set up, I can't imagine ever going back to vanilla terraform pipelines, and certainly not paying for terraform cloud.