Still, there are certain consumption expenditures that are highly unlikely to ever move online (e.g. Restaurants, Gas and Groceries). Below is a chart of US retail sales broken down by category and likelihood of online spend. Red categories are those that I subjectively labeled as unlikely to be purchased online, blue are those that are more likely to have some share online. I've faded the color on categories that may lean one way or the other. By my analysis, nearly 3/4 of US retail sales are unlikely to be made online. Only a portion of the remaining 25% is likely to be spent online too.
US retail sales are ~$4T. WMT sells for 0.5x sales, implying that AMZN would need to do $200B worth of sales to justify its current market cap with WMT's multiple. AMZN has ~25% of e-commerce market share with $40B+ in sales, so if it keeps that market share, online sales would need to be about $800B to justify AMZN's current market cap. That's 20% of current US retail sales.