This does not only apply to consumer software. I was just thinking that the past decade of business software has been about connecting off-the-shelf saas and cl

This does not only apply to consumer software. I was just thinking that the past decade of business software has been about connecting off-the-shelf saas and cloud systems together to run businesses. This was good for businesses wanting to quickly get onboard new technologies but it has real downsides because all these systems are scattered across the world and managed independently, making them more distributed, less transactional (more corruptible) and less tailored to the specific industries they're used in. Overly distributed systems are fundamentally hard and fundamentally inefficient ( https://groups.csail.mit.edu/tds/papers/Lynch/podc89.pdf https://groups.csail.mit.edu/tds/papers/Lynch/podc89.pdf ). Data that spends more time in memory and cpu cache instead of being thrown over global networks can be processed a lot more efficiently. Also, if your infrastructure is made of a lot of components that anyone can buy, it's much likely for competition to emerge. You lose an important moat. Plus, oftentimes, the outside organizations holding your data can leverage it to make you pay high prices. The next decade might be about recreating the logic of these cobbled together generic saas systems into streamlined and fully tailored software systems for each industry. There is now a lot of mature open source code for companies to use as a foundation. As the article suggests, this could also allow moving towards fully owning data and stack instead of having too much in other companies' platforms. There is probably more incentive for this reshoring of data and software to happen for businesses so it might happen there before it does for consumers.