This paper focuses on the competitively optimal power-control and signal-shaping for "ad-hoc" networks composed by Multiple-Antenna noncooperative transmit/receive terminals affected by spatially colored Multiple-Access Interference (MAI). For this purpose, the MAI-impaired network is modeled as a noncooperative strategic game, and sufficient conditions for the existence and uniqueness of the Nash Equilibrium are provided. Specifically, the following main results are achieved. First, we develop an iterative, fully decentralized, asynchronous and scalable power-control and signal-shaping algorithm that is competitively optimal and maximizes the information throughput sustained by links active over the network. Second, we test that the proposed decentralized access algorithm outperforms the (conventional) centralized orthogonal ones (as TDMA) in terms of aggregate network throughput. Third, we show that, when the throughput requested by the users are no sustainable by the network, the proposed algorithm converges to the allowable operating point at the minimum Euclidean distance from the requested one. Finally, we propose two fully decentralized Connection Admission Procedures (CAPs) that rely on the proposed decentralized access algorithm and optimize the tradeoff between aggregated networking throughput and users QoS requirements.