The French Alps are characterized by complex topography, which influences precipitation regimes and associated risks in various ways. Extreme precipitation, especially when it happens on very short timescales, can lead to flash floods and landslides with important social and economic impacts. Local stakeholders, working on prevention and protection from such events, report a perceived recent increase in sub-daily extreme precipitation intensities in the French Alps. However, to our knowledge, this has not yet been confirmed statistically. In this paper, we apply a statistical approach based on extreme value theory on long time series of in situ observations. By comparing daily and hourly observations and analysing both months and seasons, we are able to present a climatology as well as trends of precipitation extremes in unprecedented detail. Here we define 'extreme' as 20-year return levels of precipitation intensity, focusing thus on very rare events. In a first part, we use a stationary framework to provide an updated climatology of daily and hourly extremes. We make use of an existing weather pattern classification to show the dynamic influences on precipitation maxima, highlighting the strong differences between the northern and the southern French Alps as well as between seasons. In a second part, we compute trends of these extremes, for which we also find partially large differences between regions and seasons. For annual extremes of daily precipitation, we find strong positive trends in the south and negative trends in the north, mostly driven by changes in autumn. Annual trends of hourly extremes are mostly non-significant and show more of an east-west division, with positive trends in the high Alps and negative ones in the pre-Alps.