On its face, it is indeed true. Cloud and bare-metal are different beasts. However, the implied part of that response is incorrect.

Here is the same response, with the implied part spelled out:

To the contrary. When you’re hungry, apples and oranges should be compared.

The cloud and bare-metal both fulfill a basic ask: “Run servers”. The fact that they do it differently from one another, with a wide variety of operational and monetary costs should be an even stronger motivation to compare them.

Having many options with different tradeoffs is better than having few options with the same tradeoffs.

Did you know cloud SSDs can be ~24x more expensive than the bare-metal equivalent?

Yes, the cloud has better redundancy1 and is managed, but remember being hungry? Now that you know the price difference, you can decide whether to pay extra for that juicy orange. It’s not like the apple has nothing going for it except for price, either. Bare-metal is usually much faster.

Be daring — compare the incomparable. Constantly reevaluate your costs. You may surprised.

The cloud has many advantages, but its cost is often vaguely buried at the back of CTOs’ minds in the form of efficient overhead. A premium finely refined through economies of scale, unbeatable by in-house bare metal.

While that is true for some of the cloud for some of its customers, it most definitely does not apply to all companies, for all of the cloud. The marketing causing some to dismiss the possibility of beating the cloud in-house is nothing less than astonishing.

I hope that a tool like BareMetalSavings, as young as it is, will help offset that and help people consider all the options for what they are. Sure, the cloud may still end up making the most sense, even with its considerable premium. That’s fine. Any decision is fine, as long as it is taken with all the relevant information in mind.