A Space-X booster failure isn't a disaster. Fortunately, Space-X fuels remotely. The USSR had booster explosions during fueling in 1960 and in 1980, each with tens of casualties. If Space-X can find out what's wrong and fix it, it's not a huge setback. If the craft had been manned, the pad abort system would have saved the crew; the payload module was unharmed for the first few seconds, enough time for the capsule to launch. The Falcon launch failure rate is about 4%, which is on a par with everybody else.[1]. But that's not counting this as a launch failure.

Tesla's "autopilot" was too much, too fast. It's not very good and was promoted as being more capable than it is. Tesla really needs to enforce hands-on-wheel strictly.

The Tesla Model 3 at $35K with big production volume is a big risk. It will probably come out late, at $50K+, and in lower initial volume. But it should make it out the plant door at some point. That's the biggest risk item in all this. Remember that when Musk announced the proposed Model 3 production schedule, the two top people in production quit. They clearly knew the goal wasn't realistic.

Tesla buying Solar City looks more like a bailout. The big Solar City risk is that they're building their own solar panels. They could be undercut by competition from China, as Applied Materials was. Solar City has the advantage that their plant sells to themselves, so they can resist competition to some extent. Solar City's big vulnerability is that they're heavily dependent on Government subsidies. Those are gradually going away as solar becomes able to stand on its own.

The big risks are financial, not technical. Hence the pressure at Tesla to cut costs, and probably use creative accounting, to get a profitable quarter.

[1] http://i.imgur.com/ei3h1B7.png