I worked at a startup that (despite my protests) had a 5-figure AWS bill, when we could have fit the whole company on a $50/mo dedicated server. Maybe you're falling into the same trap?

It feels like people think an exit plan needs to include infrastructure that supports 1M reqs a sec

A startup using reserved instances seems like a bad thing to do and is the opposite of scale to demand.

Unless you’re actually using all that reserved compute, in which case you just didn’t generate enough value. It happens.

Taking money from someone who is handing it to you is not ruthless…

Reminds me of people blaming UberEats/DoorDash for their poor financial situation.

AWS sells a convenient service, if you cannot afford that service maybe just don't use it?

You've already burned through 100K+ (at least[1]) in start-up credits?

1. https://aws.amazon.com/startups/credits?lang=en-US

that 100k is often miscited as a figure any startup can apply for. the real number is 1k

100k has to be associated with an "Activate Provider" and a few other criteria

*1 year in credits

It's common to hit the time limit before the credit limit.

I once had a contract where they spent a decent chunk of money for me to get them running on AWS but set up in prep to do a zero downtime migration to Google, and then to Hetzner, in order to first milk AWS and Google credits... It worked well, but the savings from moving to Hetzner were so large they could've just set things up there from day one (they had massive egress relative to compute, and at the time the reduction in egress fees vs AWS alone was more than their total cost at Hetzner)

you should switch to digitalocean. they're awesome for startups. i've used them for a long time and i can tell you the uptime is really good.

How so? I'm building a startup now and am using aws, curious if you'd still recommend.

You should avoid most of their offerings because of vendor lock-in.

My two cents is to build something that could be deployed anywhere else also.

Maybe you are fine staying on AWS forever but maybe you could save a lot of money by migrating somewhere else later.

AWS fees could cost more than your startup makes, depending on how you build your service, who knows what the future brings.

So it's better to make your deployment flexible enough that you can ditch AWS and migrate to dedicated servers easily. It's something to think about from the start.

The answer depends on where you would like to spend your time on. If you believe your customers will be yours because yourygreat at operating poatgress datbase - do it, if your value in something else, go with AWS RDs postgress. Younwill save a ton of time in exchange of delegating database management to aws. Use this time to develop features your customera are looking for. If you would like to leave aws eventually it still be the same postgress that you can run anywhere you like

I love RDS, but it really does not save all that much time beyond initial setup that's trivially contracted out. A low retainer to someone who knows how to do this will be far cheaper than RDS. By all means use RDS when you can justify the cost - if it's a rounding error of your total cost base there's no point finding someone to do a proper setup. But it doesn't all that large or complex a setup before there are huge savings to make even at inflated contractor rates from having someone manage it for you. More so if you do it outside of AWS entirely, though.

Presumably, you know your cost base, and how large a percentage of your cost is hosting.

If that percentage is low, AWS is fine. It's expensive, but it's not worth your time optimizing a small proportion of your cost in an early-stage startup.

My current (not a startup) employer is on AWS and it makes sense for us as it's not a big deal relative to what customers are paying us, and it's an easier sell at the kind of deal size we deal with and our customers often require specific providers.

If it's a high percentage, but you're still at bootstrap level costs, it's fine to stick around for a while - just keep an eye on it and avoid AWS-specific services/lockin.

If it's high, and your costs in absolute terms are starting to bite, then look at your options. AWS is expensive, especially if egress bandwidth makes up a lot of your AWS bill. But you also need to consider your customer base. If it's a B2C business, then nobody cares. If it's a B2B SaaS or similar where your customers are businesses that might assign a lot of trust to AWS or have their own infra on AWS, you need to understand whether or not they care, because worst case any savings will be eaten up by lost business.

In the past, I've worked somewhere that was nicely profitable on cheaper hosting but where the margins meant using AWS would have been certain bankruptcy. And many other places where we made significant savings.

You need to understand your costs well, and whether optimizing hosting is even worth it, in other words, and even if it is worth it at some point, it might not be worth it now.

It really depends on your needs and your staff, but 3x $50/mo servers in 3 different data centers will give you excellent uptime and performance. You could host millions of page loads a month with that.

How is that ruthless? If their prices suck wouldn't you just switch to another provider?

Vendor lock in combined with sunk cost, combined with my favorite, keeping up with the joneses, aka "industry standard".

How?