In a rapidly changing society, young adults may play an important role in teaching older adults about social, cultural and technological changes. Thus older people who lack regular contact with younger people are at risk of being excluded from contemporary social developments. But how age-segregated are older people? The level of age-segregation of older people can be studied by examining the age-composition of personal social networks. Using NESTOR-LSN survey data from The Netherlands, we are able to determine the number of younger adults that people aged 55–89 years identify as members of their social networks, and to examine the factors that are associated with segregation or integration. The findings show that there is a large deficit of young adults in the networks of older people, and that few older people have regular contact with younger non-kin. If age were not a factor in the selection of network members, one would expect the age distribution of adult network members to be the same as the age distribution of the entire adult population, but the ratio of actual to expected non-kin network members aged under 35 years for those aged 65–74 years is only 0.10. And only 15 per cent of the population aged 80 or more years has weekly contact with any non-kin aged less than 65 years. The number of children is strongly related to the total number of younger network members, because most younger network members are adult children. Further, participating in organisations (work and volunteer settings) that include people of diverse ages increases the likelihood of an older person having significant cross-age interactions with non-kin.