Read the statement again. You left out "generate enough money to...."
He's talking prospectively; he's speaking about what is (or is not) going to happen.
He's saying there isn't enough money for an enterprise to produce an innovative product profitably in the open source model. That seems obvious to me.
But I'd like to understand how he's wrong.
Where he's wrong is in assuming that software innovation is a direct result of any particular package's profitability. I thoroughly detest Richard Stallman for the aging, hypocritical, hippie socialist he is, but he always had one piece of it right. Free software doesn't have anything to do with price. The innovations in the applications I mentioned were a direct result of the availability of the source code and the desire to build a better mousetrap.
The open source business model has many shortcomings, not the least of which is the tendency for it to attract people like Stallman. You cannot, however, assert that without a direct market benefit that software innovation will not occur. The applications most people see are just the tip of the iceberg. The entire realm of parallelism and software supercomputing lives mostly in the open source world. Amazing innovations occur there on an almost daily basis. Some more useful than others.