Here's how I see events unfolding. Of course, this is ENTIRELY from my imagination. I have absolutely zero inside knowledge. But it seems to fit how things operate there. Actually I could see this sort of thing happening at almost any startup when it involves an upgrade for a non-critical product. Well, that and a very sharp CEO that likes to maintain full control over details, and only likes to give good news. (I have been CTO at a startup with a CEO very much like that. Perhaps I am projecting here).
0. July 2013, at the Teslive conference, an attendee asks Musks about Roadster support. Musk answers, completely off the cuff, "in 2014, we'll circle back and do something nice for Roadster owners".
1. No other Tesla employee has any idea what he's thinking. If they ask, Musk says they'll worry about it later. Musk has several more important things to focus on.
2. Around late 2014, employees and tweeters remind Musk of his promise. "&%@!", he says. He thinks for 30 seconds, and remembers that employees have been periodically reminding him they have to provide replacement Roadster batteries, and soon because they are nearly out. His 3-legged stool for convincing the masses that EVs are great consists of:
a) Supercharging. Supercharging is hard to provide for the Roadster, not something that has to be done anyways, and won't provide more vehicle sales or additional brand halo. Next.
b) Performance. Making the Roadster really quick could help with the brand halo, but it requires many things (battery, PEM, motor, firmware, etc) to be upgraded. It will be incredibly expensive, yet still slower than an AWD Model S, so it won't grab headlines. Besides, he wants top performance to be in the Model S to attract more high-profit sales.
c) Range. He can't just throw a ton of batteries at the Model S because until the Gigafactory opens, they are very expensive and in limited quantities. But he can provide a brand halo without using a ton of batteries - and in a way where he doesn't care about costs - by making the Roadster's range incredible. And it can use very little valuable engineering time if they just replace cells in the pack. He spends 5 more seconds calculating from figures JB gave him that they could use cutting-edge batteries to make the Roadster go nearly 400 miles - a suitable number that will definitely get press.
He immediately uses that as his answer - a 400-mile battery is coming for the Roadster.
3. The next day, several Tesla employees freak out. Are we really going to charge that much - that will be more than most Roadsters are worth! It may weigh just enough we will have to do crash testing. How will we handle the cheap BROs when we offer a pack that expensive? And who knows, maybe there are, say, voltage-related technical issues using that type of cell in the old Roadster pack. "&%@!", Musk says again. But he is a quick problem solver. Put one battery engineer half-time on putting together a reasonably-priced pack that gets us as far as we can. And that aero/drag specialist that has been using a Roadster mule to get numbers for Model 3 planning, have her work half time to offer the best tire/wheel/brake options she can find for the Roadster. Let's see if that gets us close to 400 miles. Let me know tomorrow at 9am.
4. Calculations are done late in to the night, a powerpoint is thrown together, and it looks like without too much effort they can get to 340 miles by December, with a battery price below that of an 85kWh Model S. "&%@!", Musk says again. He starts saying "360 miles" and "August" in public. The managers of the two half time employees each say "it looks like you should try to do a little better". The battery engineer picks a slightly more expensive cell; while pushing the new pack price above Model S pricing, it will still make the new pack cheaper than the old pack. The drag expert starts adding aero mods. Essentially nobody else in the company has any inkling that any of this is happening.
5. The battery engineer finds a problem (PEM problem? ACP patent issue?) applying the battery change to 1.5 Roadsters. He notifies his manager, who mentions it to Musk in the next meeting. "&%@!", Musk says. He tells the manager to do the 2.x first because it is easier and affects more cars, then tackle the 1.5 problem - and then promptly forgets about it.
6. In early August, Musk asks the managers why these projects are not done. The managers point out the other S and X work that the employees have been pulled off to do. "&%@!", Musk says. When can I have something? The aero/drag stuff isn't ready to show, but the single battery that has been put together works as expected in a 2.x vehicle.
7. Just as Model X config happened at the last possible minute in August, Musk decides to announce the battery. He asks somebody from the comm team that has never heard of it to write something about it in less than an hour. He gives them the name of the battery manager, who is sleeping and doesn't answer his phone because it is the middle of the night. The poor communications employee writes what he can. Realizing customers will want more info, he alerts the online team that the store should have an entry for this. A hapless employee there cobbles together something; like the comm employee, she has no way to communicate with the employee that worked on the battery or his manager. And Musk is the only other person in the company that has ever been told it doesn't yet work with 1.5.
8. Word of the battery upgrade gets out, cries of joy spread, people place orders. Finally the battery employee realizes a public announcement has been made, and that it's missing some caveats. Word of the battery upgrade finally spreads to more than a handful of employees. Problems are realized - how fast can we build the batteries? How do we decide where to send them? How do we train the service center employees? And, oh no, they won't work with 1.5s.
9. The battery manager is elected to remind Musk that the batteries don't work with 1.5. Musk is a fast problem solver, and suggests a solution. This solution is being worked on now. But Musk only delivers good news. Saying "sorry, we goofed, 1.5s will have to wait and we don't know how long yet" is not good news. Customers complain, but Musk has always preferred silence until he can deliver good news. So we wait...