Preserving muscle mass and strength is critical for long-term health and longevity. Age-related muscle lipid accumulation has been shown to be detrimental to muscle health. In healthy older individuals, we sought to determine whether muscle lipid content, determined from computed tomography, is associated with self-reported physical function, laboratory-measured performance, and the response to progressive resistance training (PRT), and how metformin may alter these responses (N = 46 placebo, 48 metformin). Using multiple linear regression models adjusted for confounders in a large cohort, we show that intermuscular adipose tissue (IMAT) was not associated with baseline function or response to PRT, contrary to previous reports. On the other hand, thigh muscle density (TMD), as an indicator of intra- and extramyocellular lipid (IMCL and EMCL), remained strongly and independently positively associated with physical function and performance following adjustment. Baseline TMD was inversely associated with gains in strength, independent of muscle mass. Percent change in TMD was positively associated with improved chair stand and increased type II fiber frequency but was not associated with muscle hypertrophy or overall strength gain following PRT. For the first time, we show that metformin use during PRT blunted density and strength gains by inhibiting fiber type switching primarily in those with low baseline TMD. These results indicate that participants with higher muscle lipid content derive the most performance benefit from PRT. Our results further indicate that muscle density may be as influential as muscle size for strength, physical function, and performance in healthy older adults. ClinicalTrials.gov , NCT02308228, Registered on 25 November 2014.