The paper investigates the speed of scalar-induced gravitational waves (SIGWs) using data from pulsar timing arrays (PTAs) such as CPTA, EPTA, PPTA, and NANOGrav. The authors employ a Bayesian analysis to explore the speed of SIGWs, assuming that the stochastic signal observed by PTAs originates from these waves. They find that to be consistent with PTA observations, the speed of SIGWs should be \( c_g \gtrsim 0.61 \) at a 95% credible interval. This constraint aligns with the prediction of general relativity that \( c_g = 1 \) at a 90% credible interval. The study underscores the potential of PTAs as powerful tools for probing the speed of SIGWs, particularly in the nanohertz frequency range, which is accessible through PTAs but not to the same extent by ground-based detectors like LIGO and Virgo. The findings contribute to our broader understanding of gravitational theories and the nature of gravity.The paper investigates the speed of scalar-induced gravitational waves (SIGWs) using data from pulsar timing arrays (PTAs) such as CPTA, EPTA, PPTA, and NANOGrav. The authors employ a Bayesian analysis to explore the speed of SIGWs, assuming that the stochastic signal observed by PTAs originates from these waves. They find that to be consistent with PTA observations, the speed of SIGWs should be \( c_g \gtrsim 0.61 \) at a 95% credible interval. This constraint aligns with the prediction of general relativity that \( c_g = 1 \) at a 90% credible interval. The study underscores the potential of PTAs as powerful tools for probing the speed of SIGWs, particularly in the nanohertz frequency range, which is accessible through PTAs but not to the same extent by ground-based detectors like LIGO and Virgo. The findings contribute to our broader understanding of gravitational theories and the nature of gravity.