Bloom's newly-announced fuel cell converts natural gas (and hydrogen, if you can get it) into electricity at approximately 48% efficiency, is presently $700,000 for a 100KW box, and is looking to scale its cells down to smaller size at something like the equivalent cost. Does this make Bloom a game-changer?
In California, the majority of utility electricity is generated from natural gas. While the turbines that burn the natural gas are of varying efficiencies depending upon their age and location, new power plants with good planning and use of cogeneration are about the same efficiency as the Bloom fuel cell (which affords much more limited opportunity for cogeneration). Therefore the question is the possible purposes of using the Bloom fuel cell.
Assuming we're talking about individual businesses and homes buying and installing their own fuel cells, the essential result is to cut out the utility's role in electricity generation and distribution. The fundamental reason you'd want to do that, aside from fear of the grid failing and a consequent desire to be independent of the grid, is cost. While it is certainly true that presently natural gas is cheaper than utility electricity on a per-unit-energy-converted-to-electricity basis, this may not be true in the future given market fluctuations, the effect of increased demand, the policies to encourage development and distribution, the costs utilities will be allowed to charge for distribution, etc. Given the capital investment in these fuel cells, it is an interesting weighing of known and unknown risks. It is clear that at current and short-term future natural gas rates, they can indeed pay for themselves after perhaps a decade or so (factoring in the total costs involved, including permitting, engineering, space costs, installation, grid-tying, inter-grid implications, etc.) and then save money. (Actually, it would be an interesting question whether users can employ fuel cells for all net-metering and tariff-in purposes, and thereby save and possibly make money by feeding power into the grid that utilities may have to accept and pay for.)
Regarding emissions, while it is true that fuel cells do not produce combustion pollutants, carbon dioxide is still produced. Of course, fundamentally natural gas is still a fossil fuel and is therefore ultimately limited (although this country appears to have a store of natural gas to last for decades at its present rate of use).
My concern regarding fuel cells is that I would really hate for the public to latch onto them as somehow "clean" and therefore viable alternatives to renewable energy sources: solar, wind, and geothermal (and eventually wave and some biofuels). These are truly pollution-free generators of electricity, and there are quite good analyses to demonstrate that they are equally good if not better economic bets, as they are subject to only the cost and market fluctuations of the sun and tides -- which is to say, their sources are free. And, if this is just about generating cheaper local electricity, medium to large-sized businesses can already do that with micro-turbines which, when coupled with cogeneration (example: a hotel with a regular need for hot water), can raise their combined efficiencies to above 50% efficient and thus save money while burning natural gas very cleanly.
Finally, if they can be made inexpensively enough, they may pose a reasonable approach for series hybrid electricity generators on larger vehicles such as buses and trucks (their presumed size would be too bulky for cars and motorcycles, and a size that would fit cars and motorcycles would likely be too low-powered to truly act as electric vehicle range extenders). Compressed natural gas station infrastructure already exists, and buses and trucks that run regularly may benefit from the high efficiency even at the greater capital cost as compared to conventional internal combustion engines (even those designed for speed-specific series operation).