Seriously, you do not need massive amount of money to build search engine. The V0 version of Google was literally a single cheap desktop running whole thing end to end. Average webpage size even today is just 2MB. You can just start with crawling 1B pages which you can comfortably store and index on 8TB HDD (you don't need SSD at this stage). You should make your algorithm work for single user on this 1B page crawl. Don't worry about freshness or speed, just focus on relevance. If things works beautifully, find a VC and show them the goods. If things works out, you might probably end up with $100K seed money. Then you go out and buy dozen good desktops, increase crawl size to may be 15B pages and support few dozen simultaneous users. Now you can go out to big guns, send them a link to your search engine and get next level of funding. At that point you hire real employees and now your job is scaling up to thousands of invite-only users and squeezing out the long tail of performance in terms of relevance. Constantly measure your customer satisfaction and complaints, use gmail-like invite your friends model to incrementally get high-value users who are willing to try out something new. At this point, if your search algo is really much better, you should be able to plot exponential DAU/MAU growth and get funding rounds in $100M range to really scale up.