On July 29, 2025, a major earthquake with a moment magnitude of MW = 8.8 occurred off the coast of Kamchatka. The most accurate and rapid estimation of the magnitude of such events is a crucial task in modern seismology, particularly in the context of earthquake and tsunami early warning systems. In this study, we propose and test an approach for rapid MW estimation from accelerometer records based on an empirical relationship between peak ground displacement (PGD) and moment magnitude. Unlike most similar studies, we rely exclusively on accelerometric data from 13 seismic stations on Sakhalin and the Kuril Islands, without the use of GNSS observations. The final MW estimate derived from all stations was 8.75 ± 0.2, which is consistent with values reported by international agencies. Simulating real-time data processing demonstrated that a first stable magnitude estimate of MW = 7.6 ± 1.3 could be obtained as early as 4.5 minutes after the earthquake origin time, with a value of MW = 8.65 ± 0.3 available after 8 minutes. The results indicate that, given an existing accelerometer network, this methodology can provide reliable and sufficiently rapid estimates of the moment magnitude of major earthquakes without the need for GNSS data. This opens possibilities for enhancing the timeliness of source parameter estimation and for improving tsunami early warning systems in regions with developed seismic infrastructure, without requiring major modifications to existing seismic data processing systems.