We re-estimate the World Technology Frontier (WTF) non-parametrically, using the Data Envelopment Analysis method, with a dataset covering both OECD country-level and US state-level data on GDP per worker and the stocks of physical capital, unskilled labor and skilled labor. The WTF 2000 is found to be spanned by a few US states such as Colorado, Connecticut, Delaware, Nevada, Utah, and Washington, while the USA as a whole falls markedly behind these leader states. The auxilliary use of US state-level data adds extra precision to cross-country growth and levels accounting exercises. We also calculate the "appropriate technology vs. efficiency" decomposition, disentangling dynamic shifts of the WTF from movements along the WTF. Our results indicate that previous estimates of the WTF might have been downward biased and previous estimations of the technical efficiency might have been upward biased.