Building an OS is feasible for a small team or even a single programmer.
Providing enough of a subset of the expected APIs and the device support and the databases and web browser and the compilers and file systems the rest of the stack that the customers and third-party partners expect? That's a bigger project and a bigger budget.
Simply being better isn't enough.
Being faster isn't enough.
Being cheaper isn't enough.
Even running Microsoft Windows on your (non-x86) hardware isn't enough.
You need some combination of "betters" and of application and document compatibility, and you need to get to critical mass of applications and tools and device and hardware support, or you need to get to "massively better" in one or more dimensions to get enough early adopters on-board, or you need enough money to buy the tools and ports you need.
And then you have to get to big volume and to enough of a profit margins to get your prices down to where you attract application developers and resellers, or pay for the developers.
As for competition, you're working against Microsoft Windows on x86 and Apple, and FOSS limits your margins. Or against embedded vendors that excel in one or more dimensions.
There are little-known and niche and embedded operating systems and vendors all over the place. Wind River (now at Intel) is one. HP has at least three operating systems (NSK, HP-UX, VMS) and has retired others including MPE and Tru64 Unix and Domain/OS, and not counting embedded software platforms such as EFI and all those HP printers. IBM has its own OS offerings.
Building an OS is comparatively easy. Anyone with enough skills or enough cash can certainly build or buy one, and various folks can buy enough partners. But to create a self-sustaining environment within your target market and to avoid creating a massive write-off, you need to build an ecosystem around your operating system. That's a much tougher and much bigger effort.
Note that both litl and chrome use Linux under the hood, so you don't have to do all of that to make a "new" OS, for values of "new" used in this article.
Providing enough of a subset of the expected APIs and the device support and the databases and web browser and the compilers and file systems the rest of the stack that the customers and third-party partners expect? That's a bigger project and a bigger budget.
Simply being better isn't enough.
Being faster isn't enough.
Being cheaper isn't enough.
Even running Microsoft Windows on your (non-x86) hardware isn't enough.
You need some combination of "betters" and of application and document compatibility, and you need to get to critical mass of applications and tools and device and hardware support, or you need to get to "massively better" in one or more dimensions to get enough early adopters on-board, or you need enough money to buy the tools and ports you need.
And then you have to get to big volume and to enough of a profit margins to get your prices down to where you attract application developers and resellers, or pay for the developers.
As for competition, you're working against Microsoft Windows on x86 and Apple, and FOSS limits your margins. Or against embedded vendors that excel in one or more dimensions.
There are little-known and niche and embedded operating systems and vendors all over the place. Wind River (now at Intel) is one. HP has at least three operating systems (NSK, HP-UX, VMS) and has retired others including MPE and Tru64 Unix and Domain/OS, and not counting embedded software platforms such as EFI and all those HP printers. IBM has its own OS offerings.
Building an OS is comparatively easy. Anyone with enough skills or enough cash can certainly build or buy one, and various folks can buy enough partners. But to create a self-sustaining environment within your target market and to avoid creating a massive write-off, you need to build an ecosystem around your operating system. That's a much tougher and much bigger effort.