Markets for ancillary services often have minimum sizes for participation, so as to limit transaction costs. That's why there is a growing segment of "aggregators" who act as intermediaries between many small, distributed providers of ancillary services and the TSO markets. For example, EnerNOC plays this role in capacity markets in the PJM region, where they aggregate large amounts of demand response and bid the aggregate into capacity markets and other ancillary services. PJM then deals with one counter-party - EnerNOC - who in turn must manage the many smaller providers of services they engage with. It's thus on the aggregator, and not the TSO, to manage these many kilowatt-size market participants, and to the TSO, the aggregator looks like one big megawatt-sized provider.
In addition, I'm not sure that having many small providers is worse for system stability than fewer large providers. In fact, I would assume that many more distributed providers of services would be preferred from a redundancy and reliability perspective, since you have more finer increments of supply on your supply curve and you have less chance of one major provider of supply going offline when you need it. So perhaps the opposite is true: lots of smaller providers of the same aggregate quantity of service is better from a system management perspective, assuming that transaction costs can be minimized and managed (which is the role of the aggregator). Thoughts?