Water deficit is a major yield-limiting factor for many crops, and improving the root system has been proposed as a promising breeding strategy, although not in groundnut (Arachis hypogaea L.). The present work was carried out mainly to assess how root traits are influenced under water stress in groundnut, whether transgenics can alter root traits, and whether putative changes lead to water extraction differences. Several transgenic events, transformed with DREB1A driven by the rd29 promoter, along with wild-type JL24, were tested in a lysimeter system that mimics field conditions under both water stress (WS) and well-watered (WW) conditions. The WS treatment increased the maximum rooting depth, although the increase was limited to about 20% in JL24, compared to 50% in RD11. The root dry weight followed a similar trend. Consequently, the root dry weight and length density of transgenics was higher in layers below 100-cm depth (Exp. 1) and below 30 cm (Exp. 2). The root diameter was unchanged under WS treatment, except a slight increase in the 60–90-cm layer. The root diameter increased below 60 cm in both treatments. In the WW treatment, total water extraction of RD33 was higher than in JL24 and other transgenic events, and somewhat lower in RD11 than in JL24. In the WS treatment, water extraction of RD2, RD11 and RD33 was higher than in JL24. These water extraction differences were mostly apparent in the initial 21 days after treatment imposition and were well related to root length density in the 30–60-cm layer (R2 = 0.68), but not to average root length density. In conclusion, water stress promotes rooting growth more strongly in transgenic events than in the wild type, especially in deep soil layers, and this leads to increased water extraction. This opens an avenue for tapping these characteristics toward the improvement of drought adaptation in deep soil conditions, and toward a better understanding of genes involved in rooting in groundnut.