The pathophysiology of diabetic nephropathy likely has little if anything to do with tubular glucose. Diabetic nephropathy results from microvascular damage from a high glucose environment in the blood, not the urinary filtrate. Therefore, simply having an elevation in glucose concentration in the renal tubular lumen will not predispose to diabetic nephropathy in my opinion.
An interesting experiment of nature supports this hypothesis. There is a relatively rare condition known as renal glycosuria which leads to elevated glucose concentration in the urine. Patients with this familial disorder have no predisposition to renal disease. It is generally thought to be completely benign. It would appear then that pharmacologic glycosuria caused by SGLT2 inhibition would mimic this benign disorder and not predisposed to diabetic nephropathy.