The Ernest Orlando Lawrence Berkeley National Laboratory (LBNL) is working with the California Energy Commission (CEC) to determine the potential role of commercial sector distributed generation (DG) with combined heat and power (CHP) capability deployment in greenhouse gas emissions (GHG) reductions. CHP applications at large industrial sites are well known, and a large share of their potential has already been harvested. In contrast, relatively little attention has been paid to the potential of medium-sized commercial buildings, i.e. ones with peak electric loads ranging from 100 kW to 5 MW. We examine how this sector might implement DG with CHP in cost minimizing microgrids that are able to adopt and operate various energy technologies, such as solar photovoltaics (PV), on-site thermal generation, heat exchangers, solar thermal collectors, absorption chillers, and storage systems. We apply a mixed-integer linear program (MILP) that minimizes a site's annual energy costs as its objective. Using 138 representative mid-sized commercial sites in California (CA), existing tariffs of three major electricity distribution utilities plus a natural gas company, and performance data of available technology in 2020, we find the GHG reduction potential for this CA commercial sector segment, which represents about 35% of total statewide commercial sector sales. Under the assumptions made, in a reference case, this segment is estimated to be capable of economically installing 1.4 GW of CHP, 35% of the California Air Resources Board (CARB) statewide 4 GW goal for total incremental CHP deployment by 2020. However, because CARB's assumed utilization is far higher than is found by the MILP, the adopted CHP only contributes 19% of the CO2 target. Several sensitivity runs were completed. One applies a simple feed-in tariff similar to net metering, and another includes a generous self-generation incentive program (SGIP) subsidy for fuel cells. The feed-in tariff proves ineffective at stimulating CHP deployment, while the SGIP buy down is more powerful. The attractiveness of CHP varies widely by climate zone and service territory, but in general, hotter inland areas and San Diego are the more attractive regions because high cooling loads achieve higher equipment utilization. Additionally, large office buildings are surprisingly good hosts for CHP, so large office buildings in San Diego and hotter urban centers emerge as promising target hosts. Overall the effect on CO2 emissions is limited, never exceeding 27 % of the CARB target. Nonetheless, results suggest that the CO2 emissions abatement potential of CHP in mid-sized CA buildings is significant, and much more promising than is typically assumed.