crawling the web is only easy if all you ever do is following hrefs... we're in the world of single-page turing-complete web apps and you're underestimating the amount of pure compute, ingress and egress bandwidth necessary to keep the index up to date. there's a reason google has designed and built their data centers, their server chassis and even power supply.
Mere difficulty of a problem does not make those who do solve it a natural monopoly. Of course these player may still be a practical (in effect) monopoly due to their being way better than the competition.
Making a kickass image editing application like Adobe Photoshop is difficult too (as anyone who has had to suffer GIMP would attest), but that doesn't mean Photoshop forms a natural monopoly.
