I’m going to assume we are discussing the US and Canada, as these are the most auto-dependent places in the world. I’ll also divide my response into ideal and realistic scenarios.

The ideal scenario is not that ICE engines are replaced by EVs, but that ICE engines are replaced by walking, cycling, transit, and electric micromobility. This would require:

1. Pigouvian taxes - taxing the release of carbon into the atmosphere, taxing the registration of vehicles (more), taxing the use of highways (for road use and wear and tear) or the use of highway exit ramps (imposing the externality of your vehicle on the urban environment). These taxes can be imposed initially at a very low cost, and then increased over time to gradually make the social costs imposed by these activities equal to the actual cost that users bear. Note: while the taxes themselves are functionally regressive, these proposals tend to be paired with a citizens dividend to offset the increased cost of goods and services and to ease the transition to less carbon intensive ways of living. The only people who will be negatively impacted would be those who already have excess wealth and use it in especially carbon-intensive ways.
2. Relaxation of zoning and building regulations - aka, let people build things. Peoples need for full sized automobiles is driven largely by the fact that their homes are far from their work which is far from the grocery store which is far from their social spaces. This is not a solution which can be solved with infrastructure, as you geometrically cannot fit all the transportation infrastructure between these places in an efficient manner - you need to put these things closer together so that a person can, say, walk from home to a transit station, transit to work, transit to the grocery store, walk home with their groceries, then walk to their social activity. Part of this is ending single family residential zoning, and instead allowing mixed use in all areas, part is changing building requirements - like parking requirements and overly stringent aesthetic conditions, and part is reforming building codes to be more flexible and understandable (note - not in a slapdash DOGE capacity, but reviewed by multidisciplinary teams of experts, with an eye towards making things understandable enough that a fairly average DIYer could confidently do their own construction within the limits of the building code)
3. Intelligent investments in infrastructure and transit. These should prioritize low cost, quick to implement changes in high impact areas - like replacing parking with bike lanes or closing off streets to cars and instead allowing only pedestrians, cyclists, and transit. The most important changes are to rehabilitate old downtowns which were originally pedestrian-friendly, since this will be the highest impact change. Changes should then radiate outwards from urban cores to facilitate movement around the city. In first-ring suburbs, initial big wins are things like implementing BRT lines with frequent schedules along arterials, protected bike lanes on larger neighborhood streets, public protected bike parking and pleasant pedestrian shelters at transit stops, and raised pedestrian crossings, speed bumps, and other traffic calming measures anywhere where cars are driving too fast. Of course, this should also be paired with a mandate to not accept any more sprawling suburban style development into the city’s land portfolio, since these developments are a drain on city resources and would simply need to be rehabilitated later. Also note that these changes to infrastructure and transit do not prioritize things like inter-city high speed rail, since as we have seen with these projects in the past, these rail lines end up underutilized as long as their destinations are not walkable. An inter-city BRT line can achieve 90% of the benefits of high-speed rail using existing busses and some paint on the highway. As a rule, grand, ribbon-cutting-worthy transit projects tend to end up as expensive boondoggles which take decades to complete and which are underutilized. Instead, the vast majority of infrastructure improvements should be driven by walking around in neighborhoods and asking “how can we make this more safe and pleasant for everyone?”

In this case, most gas stations would continue to function more or less as they currently do. Fewer people would make the switch to EVs, and would instead simply drive their cars less as they become less dependent on them. But due to lower demand for gas, some gas stations would slowly become financially non-viable, and would go out of business. This wouldn’t mean all of them would go out of business at once - instead it would mean that at an intersection with a gas station at each corner, 3 would go out of business and the best one would remain. In denser urban areas, many would likely divide the parcel they are on and continue functioning as a convenience store, while the pump and parking areas turned into some other, better use. Near highways, the larger truck stop style gas stations would likely remain largely the same.

The more realistic scenario is that EV tech evolves and everyone replaces their ICE cars with EV cars. In this case, gas stations will try to predict how the market will move and will try to pivot in whatever direction they expect it will take.

One anticipated direction is that gas stations will turn into charging stations. Since charging, even in the best case, takes a while, these charging stations will provide a more pleasant customer experience, integrating restaurants, shopping, and entertainment to keep customers busy while their cars charge. You can already see the stations anticipating these trends with the rise of “luxury” gas stations like Buccees, Wawa, and Maverik.

Another direction it could go is that instead of a charging station, EVs will develop swappable batteries. This process might require human attendants, and provide jobs for a number of years until the process could be automated.

But in either case, demand for charging stations would be severly reduced in urban areas, as it will be cheaper and more convenient for people to charge at home and at work for their daily commutes. Again, we would see 3 gas stations at an intersection go under while the 4 takes all the remaining business. But under these conditions, the 3 that go under would likely sit as vacant husks, blighting the urban landscape, rather than being redeveloped into something that actually serves people.