Snow is a vital resource for a host of natural and human systems. Global warming is projected to drive widespread decreases in snow accumulation by the end of the century, potentially affecting water, food, and energy supplies, seasonal heat extremes, and wildfire risk. However, over the next few decades, when the planning and implementation of current adaptation responses are most relevant, the snow response is more uncertain, largely because of uncertainty in regional and local precipitation trends. We use a large (40-member) single-model ensemble climate model experiment to examine the influence of precipitation variability on the direction and magnitude of near-term Northern Hemisphere snow trends. We find that near-term uncertainty in the sign of regional precipitation change does not cascade into uncertainty in the sign of regional snow accumulation change. Rather, temperature increases drive statistically robust consistency in the sign of future near-term snow accumulation trends, with all regions exhibiting reductions in the fraction of precipitation falling as snow, along with mean decreases in late-season snow accumulation. However, internal variability does create uncertainty in the magnitude of hemispheric and regional snow changes, including uncertainty as large as 33 % of the baseline mean. In addition, within the 40-member ensemble, many mid-latitude grid points exhibit at least one realization with a statistically significant positive trend in net snow accumulation, and at least one realization with a statistically significant negative trend. These results suggest that the direction of near-term snow accumulation change is robust at the regional scale, but that internal variability can influence the magnitude and direction of snow accumulation changes at the local scale, even in areas that exhibit a high signal-to-noise ratio.