But even with the high price, there's no other recon aircraft that can carry tons of spy satellite grade sensors 20 kilometers up, and stay there for 32 hours at a time. What it comes down to is that Global Hawk is a lot cheaper than a spy satellite, and a lot more useful because of its endurance.
Although the cost of the 13 ton Global Hawk continue to rise, the U.S. Air Force can't get enough of them, fast enough. Total program cost, including a lot of research and development, is now pegged at $7.8 billion. That pays for 54 aircraft, including spares, ground stations and so on. That's $145 million per aircraft. Seven have already been built, and another 17 are under construction. The air force wants to turn out seven a year. The original cost projections (with R&D) were under $50 million an aircraft. Several things happened to drive up the cost. First, the aircraft was a great success, and more powerful (and heavier) sensors were loaded on. This led to a slightly larger "B" version of the aircraft, and a big jump in price. Another issue was reliability. This is the first UAV to take automation so far, and there were problems that were expensive to fix. This was made worse by the fact that the Global Hawks were in great demand (it's wartime), but that the aircraft was still in development on September 11, 2001. Finally, there's the usual low balling carried out both by the air force and the contractors that build things. Congress always falls for this, often with a wink, and then loudly protests when the final cost spiral out of site.