Due to an odd quirk in how the game industry developed, went bankrupt, and then re-emerged and marketed itself during other cultural changes in the 1980s that were pushing women away from tech more broadly, for the next couple decades the vast majority of gamers were straight, white, cisgendered boys between the ages of around 12 and 25. Tons of market research and focus groups went into figuring out what kind of characters this specific audience wanted to play, and the result was a bunch of main characters that all looked like this.
There is nothing necessarily wrong with this in every case. It's not that gamers were all necessarily racist/sexist, or that the industry had a lack of imagination, or that there was some conspiracy. It's that we researched the hell out of it, and found that characters who looked like that hit the largest overlapping segment of our existing customers in terms of what they wanted to play, for the budgets we had. So we stuck with that. It was a risk calculation.
The trouble is, the game industry started to hit saturation in terms of that customer space back 10-15 years ago. In other words, the gamers who want to play characters who all look like that are a relatively stable number, and they were already all playing our games. In order for the industry to grow, and to get the increased sales to experiment with new types of game designs like everyone wanted, we needed to find new customers outside of that already saturated customer base.