It changes the value chain in ways that destroy their competitive position. Short term, it's pretty likely that they'll adapt. Long term, they're dead.

A good example is the IBM PC. When it came out, everyone was saying "Of course, IBM will now dominate the new personal computer market, because they have the sales & marketing apparatus to reach into every business that will want to buy one. Nobody ever got fired for buying IBM, after all, and now that they own the technology to make a personal computer, their offering is clearly superior."

But that's not what happened. Instead, they did dominate the PC market - for approximately 5 years. But the PC had reduced prices so that it was now targeting a market that was cost-sensitive, and it had created a secondary market of applications that let it reach into many areas that had previously required custom software direct from the manufacturer. IBM did not own the critical matchmaking components of this, the instruction set, operating system, and BIOS. Intel and Microsoft did, and then Compaq reverse-engineered the BIOS. As a result, clones flooded in, IBM's sales & marketing prowess counted for nothing, and they found their market commoditized.

Uber's critical value proposition is serving as a market-maker in a two-sided market. That's the part that's really difficult for a startup to clone. You can make the Uber ride-sharing software trivially, and many people have [1]. But even if you do, riders won't use your service because you don't have the same number of drivers available that Uber does, and drivers won't join because they won't make as much in fares.

When self-driving cars come out, that two-sided marketplace becomes a one-sided marketplace. We've yet to see how Google will market the technology, but the most strategically advantageous approach for them is to contract out manufacture of the cars, own the hardware, put their own software on it (and not license it out), and then sell a service to riders, undercutting Ubers' prices. Under this model, Uber's competitive advantage counts for nothing - their supply chain costs more than the competition, in a price-sensitive one-sided market.

Google could then use a number of different tactics to lock Uber out. The most likely one is regulatory; in the interest of public safety, they could argue that all self-driving cars need to pass a very stringent safety test, consisting of real-world driving for X00,000 miles. Google's got a 10-year head start on Uber for developing this software, and once a critical mass of cars on the road are Google self-drivers, they have accurate position information on everybody else, a key factor in making this safer.

[1] http://www.businessinsider.com/homeless-coders-trees-for-car...