In theory, innovation processes lie behind the evolution of national systems as they create interacting dynamics among organisations. Institutions and policies are considered means for influencing these interactive dynamics, such as shifting innovative focus from traditional to environmentally oriented production, more environmentally friendly types of energy use, or environmental protection measures, products or services. Institutions and policies are thus considered drivers of change in technologies, processes, markets, raw materials or organisational forms--innovation in a Schumpeterian sense. Shifts in energy sources, from fossil to non-fossil sources, among organisations in the Swedish innovation system therefore call for explanations in terms of changed institutions and policies and their resulting eco-innovations. This paper looks more closely into climate gas emission and the shift to non-fossil energy in Sweden; what types of organisations are behind the shift to non-fossil energy use, what are the relative effects on emissions, to what extent can these interactive dynamics be considered eco-innovations, and if so, can they be related to specific institutions and policies? Quantitative analysis of evolving innovation processes in national systems is not always possible due to a lack of reliable and multi-level time-series data sets. This is also true for eco-innovations ('green' innovations). In the Swedish case, there are detailed data sets at national, regional, organisational and employee levels, making possible the estimation of evolutionary models. Register data can be merged with time series on environmental energy consumption and emissions. Data allow for a detailed analysis of environmentally oriented innovation from at least 2003. Analyses in this paper are based on time-series of data on the recent shift from fossil to non-fossil energy sources in the Swedish innovation system, as well as data on emissions, and potentially innovation promoting parameters at organisational and employee levels. Methods are quantitative, and Cox regression is used. Previous investigations of the energy use of Swedish organisations reveal a clear shift from fossil to non-fossil energy use. This is described both in terms of cumulative energy use and effects on emissions of carbon dioxide. Data provide us with information for conclusions on why energy sources change and in interaction with what organisational parameters. For example, wood fuel and solid waste increase as sources of energy while fossil oil has decreased during the years 2003 to 2010. This result is in line with national industrial and environmental policies and presented as institutionally and policy related 'green innovation'. But a quantitative analysis contests such a conclusion and it is noticed that the shift to non-fossil sources of energy has not led to verifiable decreases in green-house gas emissions. Public ownership is the single most important contributor to green innovation in non-fossil energy use. Still, CO₂emissions are not fundamentally reduced by this low-tech shift, since they do not affect end-of-pipe reductions. What we observe is in fact wind-fall gains rather than eco-innovations behind the Swedish shift from fossil to non-fossil energy use.