Main idea

A big Oracle database, properly sized servers, a machine room and the infrastructure teams to run them: for years, that equipment was enough to open a gap between a software vendor and its competitors. What differentiated was not the quality of the use, it was the price of entry — you already had to be big to equip yourself like a big player.

AWS, Google Cloud and Azure removed that barrier by changing one single thing: the mode of acquisition. The same building blocks remained, but they are rented by the hour instead of bought upfront. An advantage that rested on the ability to tie up capital then disappears, since what founded it was not the building block but the fact of being able to afford it.

The mechanism therefore reads the opposite way round from common intuition: it is not technical progress that commoditized infrastructure, it is the shift from buying to renting. Technology did not lose its usefulness, it lost its scarcity.

Why it matters

This gives a criterion for judging the durability of a technical advantage before it evaporates: ask what, in that advantage, rests on an investment the others cannot make. Anything that reduces to that will fall as soon as a supplier turns it into an on-demand service.

It also explains why the execution gap persists when the access gap disappears. Some still operate, secure and architect better than others — but that gap is built in practice, not in acquisition, and it no longer protects anyone by the mere fact of existing.

Nuances and limits

Commoditization is never complete: scale, price negotiation and operational mastery recreate gaps among very large consumers of resources.

And renting on demand shifts dependence rather than removing it: the supplier that removed the barrier is in a position to set the price of what it commoditized.

Open questions

• Which current technical advantages still rest on a cost of entry, and which would survive being made available on demand?