In non-urban KwaZulu-Natal, South Africa, very few households escape the impacts of HIV and AIDS, either the direct impacts as a result of illness and death, or the indirect impacts through providing care and support to family, friends and neighbours. HIV and AIDS becomes part of the context or situation within which households arrange their lives, generate livelihoods and arrange and provide care. The differential impacts of HIV and AIDS on male and female members of different ages within households is poorly documented and understood. How people arrange care, especially for household members who are chronically ill, while generating livelihoods at the same time, is even less clear in the context of HIV and AIDS. This research assessed household living and care arrangements and livelihood generation in non-urban Mbonambi in the KwaZulu-Natal province of South Africa, in the context of HIV and AIDS. The study used a combined approach of quantitative and qualitative methodologies. Demographic, socio-economic and health data were collected at the level of the household by means of a survey and results were verified and clarified by means of focus group discussions. For the survey, two research locations were selected, one close to town, with a high population density and fairly good infrastructure and the other further from town and with poorer infrastructure. In the latter location, lack of access to electricity and clean water close to the home adds to the burden of domestic work. In addition, this location also has fewer individuals who are working, with many of those who are working, employed in low paying elementary occupations or working as unskilled labourers. Households at this location also have lower household incomes, are more dependent on state grants and own fewer assets that can be converted to cash if need be. Female-headed households proved to be bigger than male-headed ones, having significantly more demographic and effective dependents residing at their homesteads. Female heads are significantly older than their male counterparts, the majority of them widows relying on state old-age pensions as the main source of household income. Female-headed households have significantly lower average incomes and fewer assets than male-headed households. All the households in the survey sample were categorised based on whether and how they were afflicted and/or affected by HIV/AIDS and/or TB, where TB was used as a proxy indicator for HIV infection. Households were allocated to four clusters. Households in Cluster 1 did not experience any impacts attributed to AIDS and included just more that half of al the households. Afflicted households in Cluster 2 had at least one ill member diagnosed with HIV or TB and requiring some care, but did not experience any deaths and were not taking care of orphans. Affected households in Cluster 3 had no ill members, but took care of orphans and/or experienced deaths, while households in Cluster 4 were both afflicted and affected by HIV and AIDS. Progression from Cluster 1 to Cluster 4 showed a significant difference in household size, with households in Cluster 4 having on average two more members than households in Cluster 1. Households in Clusters 3 and 4 had significantly more demographic dependents than those in Clusters 1 and 2, while the households hosting orphans in Clusters 2 and 4 had significantly more effective dependents than the households in the other clusters. Although not significant, households in Clusters 2, 3 and 4 had lower household incomes and fewer assets. Of all the households it is clearly visible that households in Cluster 4 that host ill persons and orphans, and experienced deaths, are in all regards worse off than the households in the other clusters, and are extremely vulnerable to livelihood insecurity. Considering that these households have more dependents they will be more severely affected by the lower household income that has to be shared by more persons. Having fewer assets also mean that they do not have anything they can sell when they need money to cover household expenses or to pay for transport or a funeral. Case study households were selected from each cluster for further study of their living arrangements and livelihoods. This was done by means of interviews and observations, and each household was visited at least two times over a period of six months. This revealed that the majority of households experienced changes in their living arrangements, regardless of whether and how they were affected by HIV and AIDS. It was especially young people and children who were mobile and individuals were leaving or joining households for a variety of reasons. Young women with or without their children were leaving to look for work, get married or provide care. Mobile children moved between the homesteads of unmarried mothers and biological fathers. The case study households included several households where unmarried mothers were living with their children at the homesteads of their frequently unmarried or widowed mothers. Although changes in living arrangements can be caused by many factors other than morbidity and mortality, the majority of cases described experienced changes as a direct result of TB and/or AIDS-related illness and death. The time frame of inter-household movements varies from a few months to several years. The variation in cases presented illustrates that when movements between homesteads take place, the impact of HIV/AIDS-related morbidity and mortality on the livelihood and resources extends beyond the single household. It is clearly visible that the majority of households depend on social transfers, either grants from government or private grants, as their only or main source of income, emphasizing the strategic importance of grants in coping with poverty. The financial situation of households may even improve when children receiving grants join a household and are ‘accompanied’ by their grants. But when such children move, the gain of income in one household will translate into a loss for another. Furthermore, some cases show that accessing grants for children is difficult when the status of the child changes and/or the foster parent does not have the required papers. The role of maternal parents or grandparents becomes clear when looking at intra-household cooperation to arrange health care or take care of vulnerable or orphaned children. All the households are visited regularly by paid Community Health Workers and/or volunteer Home Based Caregivers, all of them female. These people are well-trained and work closely with the local public health clinic to assist households with care activities, caregivers with emotional support and patients with nutritional advice and traditional treatments to maintain health and relief symptoms. This is very important, as none of the HIV-positive persons in this small sample were on antiretroviral (ARV) treatment at the time of the research. Although treatment is free, to access it means regular blood tests and frequent hospital visits, which translates into indirect costs. The cases clearly reveal that women are still the main providers of health- and childcare. When the demand on their time to provide care increases, they have less time to devote to income generating and community activities, which means less time to invest in social networks. This will cause already poor households with weak safety nets ‘to fall through’ the vulnerability threshold. All case households reveal the significance of social capital, the network of kin in particular, as a source of material and immaterial support. Relatives may take in a child to relieve the household’s burden, may send money, or may provide emotional and practical support. When no relatives are living nearby, the neighbours provide the latter kind of support. At the same time, the cases also show ‘missing’ partners and parents who have opted out and whose whereabouts are sometimes not even known. Although the majority of children in the case study households manage to stay in school, they are absent from school more often due to HIV/AIDS-related morbidity and mortality. As a result they fall behind and are at risk of eventually dropping out. Some children choose to stay at the homestead of their late parents, with or without adult supervision rather than moving in with grandparents or other relatives, in an attempt to retain their parents’ homestead and land. This may make children vulnerable to exploitation. Child migration as a strategy to cope with HIV/AIDS-related morbidity was employed by some of the households. Although migration in search of employment has long been common in Southern Africa, migration of ill persons and children seeking care is a much more recent phenomenon. Inter-household movements are likely to occur when a household affected by AIDS-related morbidity and mortality does not have the capacity to meet the additional demand for care. Moving of ill persons or vulnerable or orphaned children across household boundaries may make for more efficient use of human, material and financial resources. The cases show a continuous adaptation of living arrangements in response to illness and death. While the homestead and the kinship network still function as important anchors for people’s lives, at the same time HIV and AIDS induce flux and instability, changes dependency relations between homesteads, makes ‘holes’ in safety nets, and undermines relations between partners, in particular those that are not sanctioned by traditional marriage, turning their children into de facto orphans. The homestead also seems to be losing its unified and patriarchal character, though more analysis is needed to prove this, and the supportive role and authority of grandmothers and maternal relatives is increasing. Care is not only morally grounded, it can also add to moral authority. The government should look into ways to facilitate better access to ARV treatment, because this would not only improve and prolong the life of people living with HIV, but also contribute to a better quality of life for household members. Streamlining access to foster care grants will prevent households taking care of orphans or orphans living on their own from living in extreme poverty. Increasing the number of well-trained paid community health workers, liaising with formal health care and social workers, will enhance the much need support required by households living with the burden of HIV/AIDS-related morbidity and mortality. Although consisting of a very small sample of households studied over a relatively short period of time, this study shows significant HIV/AIDS-induced changes in living arrangements, the variation in the timeframe of these changes, and the impact of these changes on the livelihoods of households and their potential to arrange health- and childcare, thus revealing the mechanisms of micro-level social change induced by the AIDS epidemic. It demonstrates the importance of qualitative research to complement cross-sectional survey research. More qualitative and longitudinal research is needed to know whether in the wake of the epidemic the cultural and social landscape of rural KwaZulu-Natal is fundamentally changing.