Do you have even the foggiest of notions of polygons, resolution, and 3D rendering?
Seemingly, not. If I render a spinning box, that's six sides, 12 triangles. The output resolution is only how many squares/pixels I use to represent that box. The box itself is an infinite "resolution" because all the elements are vectors (relative directions in space). Resolution is completely disconnected from the object being rendered. For example, I can zoom in and out on it at a fixed resolution, like 640x480, and it will obviously use fewer or more pixels within that area to render. I can render it at 320x240, 640x480, 1024x768, 1600x1200, 1920x1080, or eleventy-billion by eleventy-billion. The difference is that more pixels takes more time and processing power.
Pixar did not only record a fixed resolution video output in digital format, they saved the models, objects, scenery, and animation data to completely re-render on a modern machine. When they made the 4:3 version of "A Bug's Life," they moved characters around in the scene to fit in the frame and re-rendered. When they re-released "Knick-Knack" with "Finding Nemo," a short that was made before their association with family-friendly Disney, they changed the model for the Barbie-doll knock-off to get rid of the pointy breasts.
What you are saying is like playing a video game and recording it and then saying that you are stuck with the result of the first play through. No you aren't. You can play again and do something different. You could continue developing it and make a special or GOTY edition. Devs don't just throw away the source code and resource data when the first version is compiled. You could upgrade your graphics hardware or port it to a more powerful machine. Similarly, Pixar still has the source material the film was made from and not simply the static final output.