In the context of the power system reform, the transmission and distribution price reform has changed the profit model of power grid companies, from the difference between the purchase and sale of electricity to the acquisition of the total permitted income from transmission and distribution, which is closely related to the effective investment scale. Higher requirements are put forward, and the investment scale under different boundary conditions needs to be analyzed and predicted. Therefore, this paper first analyzes the main influencing factors of grid enterprise investment, and then uses the support vector machine optimized by salp swarm algorithm to predict and analyze the investment scale of the grid enterprise under different boundary conditions, and finally gives the corresponding conclusions and recommendations. The research results show that the factors affecting the investment scale of a grid enterprise are GDP, electricity sales, resident population, peak load, secondary industry structure, net asset value ratio, asset-liability ratio, return on equity, and operation and maintenance costs; through the support vector machine model, it is predicted that the investment scale under different boundary conditions in 2025 will be between 122 billion yuan and 153 billion yuan. In the sensitivity analysis, the investment scale of the grid enterprise has the highest sensitivity to GDP and electricity sales, with sensitivity of about 0.6 and 0.5 respectively.