Tough question. Like Tom said, a lot of the high expectations of the 1990s were based on dreams of several companies operating large constellations of LEO comsats. However, the market never really materialized. The Iridium constellation was launched at a cost of about $3 billion if memory serves. However, Iridium went bankrupt and they were on the verge of deorbiting the satellites when the system was sold to someone for the bargain price of about $60 million. At that price, they could lower their prices enough to actually make a profit (the US government is one of their biggest customers). However, no one else is going to be able to launch and maintain a competing constellation for that price, so there won't be hundreds of new satellites launched. The other companies with LEO comsat constellations are also on the verge of bankrupsy so they'll have a hard time building, expanding, and maintaining their constellations. Teledesic suspended work on their systems last year, so where is the market for LEO launch services?

Really small satellites (nanosats and microsats) are getting a lot of interest in the research (and space control) communities. Most of them are launched as secondary payloads rather than by themselves. For weight and balance purposes, it isn't uncommon for a booster to carry ballast. A secondary payload can replace some or all of the ballast which increases the utility of the booster. AMSATs have been launched that way for a long time. I was reading one of the May issues of Aviation Week yesterday and it had an article about an upcoming Russian launch that will carry 8 payloads including 2 or 3 nanosats.

It's hard to find good cost figures on launchers due to the highly competitive nature of the business. I've read that a Pegasus flight costs about $10 million and can put around 1200 pounds into LEO and about 800 pounds into a sun-synchronous orbit. That works out to about $8000 a pound to orbit - pretty damned pricey. The Taurus booster probably costs more but it can also lift more, so the price per pound may be similiar to the Pegasus.

One problem with developing a new medium lift class booster is that you have to demonstrate a high enough reliability to get reasonable insurance costs (unless you're talking about the US government, which self-insures all of its flights). It does no good to save a few million on launch costs if that gets eaten by higher insurance costs. Companies planning on launching a revenue generating satellite (say an imagery satellite) will buy insurance, but that typically only covers replacement costs. Insurance wouldn't pay for the lost revenue following a failed launch (say a year or two of downtime due to the lost satellite), so launcher reliability is very important.

You have to be able to amortize your R&D costs over an expected number of flights. You have to compete with other countries like Russia that don't have to worry much about R&D costs (their Soyuz booster dates back to the 1960s, has a very good reliability rate with about 2000 flights in its record, can lift upwards of 16,000 pounds to LEO, and is even man-rated) and low labor costs. If allowed by international agreements, the Russians could probably sell Soyuz launch services for about $1000 a pound to LEO and still make a profit. It's damned hard to compete against that.

I've read that Burt Rutan thinks he can launch small satellites from his Proteus or Viking aircraft. By small, I'm talking about 80 pounds or so. I've read that such a launch might cost about $500,000, which would be over $6000 a pound. I suspect a lot of that is due to guidance system costs and my memory may be failing me in regards to the price. If he could cheaply recover the guidance system, then perhaps the price could be even lower. The question then comes back to your original point - is there a market for such services?