Researchers have developed a new text-to-speech system called X2Streaming-TTS that can generate speech from uncertain prefixes and maintain perceptual continuity over an unbounded stream. The system uses a framework that consumes asynchronously arriving text tokens and emits speech without accessing future input. It introduces two key concepts: causal commitment, which keeps ambiguous expressions provisional through uncertainty-aware buffering, and causal speech-state inheritance, which carries the complete Code2Wav state across segment boundaries. Experiments show that X2Streaming-TTS outperforms existing pseudo-streaming models on most subjective and objective metrics. --- Why it matters: This matters to AI researchers because it enables strict token-level synthesis with quality comparable to offline baselines, reducing latency in spoken dialogue systems. The system's ability to handle uncertain prefixes and maintain acoustic continuity is particularly relevant for applications where low-latency speech generation is critical. Source: https://arxiv.org/abs/2608.18661