I'm pretty sure the winner would be the IRS's modernization program, which was to write new software to process tax returns. It began over 16 years ago and has consumed over $12 billion (just for the software development portion).
$4 billion was spent on the first attempt at writing this software before they gave up in 1997 with no usable product. That's a $4 billion "app" nobody ever used.
It's been a few years since I read about it so things may have changed, but last read was they spent another $8 billion in the 2000s trying again, this time outsourcing the development to consulting firms with thousands of developers. And around 2005/2006 had yet to produce a system the IRS could use, forcing them to continue processing tax returns with paper and antiquated systems.
The estimated payroll costs for Windows Vista are estimated to be 10B. Now consider the costs for every Windows version released since 1985, and include R&D/marketing expenses.
But that's an operating system, not an application. You've got to give credit to the IRS spending more on one application than Microsoft spent on their OS.
OSes are not particularly complicated. I could imagine that tax laws require a lot more code (and tests) than dragging some windows around and scheduling processes. (Neither are easy, of course.)
That "one application" was software+hardware tasked with managing millions of hand-written paper forms. The electronic version of tax filing was far cheaper, and launched successfully.
You're getting a lot of replies that aren't mobile apps, but just software products in general.
In terms of mobile apps my money (hah!) is on one of the many medical apps that are in development - we're talking about apps intended for use by medical professionals in actual medical institutions.
Not only are they extremely costly to develop (someone's got to pay the army of consultants, after all), they face incredibly costly certification processes before they can see any actual users.
You're right. We're an app dev company, and we're in the process of developing a medical app for a client. The FDA approval process makes it very time consuming.
Actually, thinking in terms of largest total return for a project (ROI is infinite if your I is 0).. I can't think of too many projects that have had larger returns.
Probably the iPhone? Maybe some Cars (camry? Model T?) Maybe some really big mining/oil operations? What else is in the same league?
Depends on your definition of "App". The article is concentrating on mobile application, which is probably the definition you are using.
Regardless, I don't have an opinion on the most expensive mobile app, but your question made me think of a recent bit of news where the US military spent over 1 Billion dollars on an ERP system and the project is being scrapped. Ouch.
I would guess Facebook because they would be changing it on a daily basis to keep it consistent with their web version, and adding dedicated functionality unique to the mobile platform.
Here's a competitor for that title, the NHS National Program for IT. The plan was to build a patient records system to run in every hospital in the UK. It was billed as the biggest IT project ever.
And its still counting. I have a friend working in IT in the NHS and they are still spending money hand over fist on transforming digital records. Basically, nobody wants to be accused of putting tax money before accuracy of patient records which essentially means blank cheque and unbounded scope.
EDIT: Clarification - mobile app, as stated in the article.