Multiple-input multiple-output (MIMO) is a technique that has been extensively studied and successfully implemented in various communication systems over the past few decades. It has been a key technology in both 4G and 5G mobile communication systems and will be a key enabler in 6G to support the extreme requirements of data rate, energy efficiency (EE), etc. However, meeting these requirements necessitates evolution in the architecture of massive MIMO systems, and distributed MIMO (D-MIMO) has garnered significant attention as a potential solution. This article articulates one of the key findings on adaptive large MIMO systems from the Hexa-X-II European 6G Flagship project and provides a vision of what large MIMO systems may entail in 6G. To achieve this, we examine an adaptive deployment strategy in a joint co-located and distributed MIMO setup, including the utilization of different carrier frequencies and coordinated transmission under an ideal channel state information assumption. We evaluate EE and data rate across various user densities. Additionally, we provide a framework to select the MIMO architecture based on user density and spatial distribution. The key challenges and future research directions on robust large MIMO systems towards realistic 6G networks are also discussed.