Qualcomm can use their cash to fund more CPU/GPU design teams to cover more design points.
I'm thinking there's a law of diminishing returns here. Intel has armies of designers, and they have been able to sustain dozens of designs. A handful dedicated to the phone biz would be easy for Intel, if Qualcomm started funnelling money. There's also a limit to the number of qualified designers in the industry, and many of them are already working for Intel or other large companies. But assuming they did grow their design teams, there's only so many building blocks available from the ARM ecosystem. Dual core, quad core, their own custom core, big graphics, small graphics, medium graphics, etc. They could make a giganto level SOC design and a 6" superphone, and Intel could do the same thing on a smaller process node for less cost and smaller die size, put it in a 4.5-5" phone, and end up with better appeal. Hard to win when at the whim of the foundry process....
Another option is to work more closely with TSMC and fund some process R&D to tune for their designs.
You don't think they're working hard with TSMC already? Another thing that's hard to beat is the process/architecture codesign that Intel benefits from by having an in-house process.
They can acquire companies that fill in missing pieces of their SOC solutions such as the recent Atheros acquisition.
Yes, many things are under consideration with the size of the war chest that Qualcomm has. I just can't think of any at the moment that would be real game-changers in terms of competing with Intel in the silicon business.