As long as you are staying within SharePoint's hard limits, Nintex Form's does not have a hard limit in the number of controls or number of validations you can add; however based on the amount of information that is being committed, the form may not submit (and then open again) instantly. In SharePoint there are several hard limits like not being able to query more than 5000 items at a time. There are even hard limits on the number columns that you can have on the same list (no more than 276 Single Line of Text columns). Here are some factors that can effect performance for Nintex Forms:
Breaking Apart The Form/Redirects: You can configure a form to redirect to another form, URL, page, or list when a button is selected. This can allow you to segment portions of your form and minimize the number of controls on one form. The use of content types will allow you to have separate Nintex Forms on the same list. There are many methods to improving a form's performance and improving user experience. Below are a couple factors that could improve overall form performance:
Tabs: Though you may have the same number of controls, tabs allow you to break up the form for a better user experience. Create Tab like functionality on your Nintex Forms (Note: When designing a Nintex Form with Tabs, hidden controls will still trigger validation and prevent form submission if they are required.)
Tasks: On the initial form, have the required fields but not all fields. Then a workflow will start and eventually have a task for the initiator or other appropriate users. This task can be a request data task with a custom form to get the remaining data.
Number of Controls: Only use the required number of controls. You may have 40 columns in your list, but only really need 10 of them filled out in the form (as the rest will be updated manually).
Lookups and Calculated Values: Reduce the amount of lookup fields and calculated fields as much as possible on the form. Any queries made to display content on the form will take time. Especially unfiltered queries not optimized with filters on indexes.