Groceries are one of the few large markets that require some proximity to customers due to costs and spoilage. Each grocery store is a type of mini-distribution center for grocery products.
Shipt and Instacart have succeeded to date because they use existing distribution channels and set up marketplaces for the "last-mile" of delivery. This is in contrast to Webvan in the early 2000's who tried to do grocery delivery by building their own distribution and failed spectacularly.
Amazon has become an expert in distribution and logistics. But it is clear that using their current model doesn't generally work with groceries (RIP Webvan, 1998-2001). Bananas need to be treated much differently than books.
So what does Amazon do? But Whole Foods!! A moderate sized grocery store with a significant national footprint and lots of higher income customers.
Now they instantly have a pre-built distribution channel that is already optimized for the grocery business (which again is much different than non-perishable consumer goods etc).
Things definitely just got interesting in this space!! I still believe that Instacart and Shipt can succeed, but they need to maintain a laser focus on making their shoppers and customers happy! And grow as fast as possible while Amazon digests Whole Foods!
[Note: I was the early CTO for Shipt responsible for building their grocery delivery platform and initial engineering team. Go Shipt!]