Right now, robots making grocery deliveries are becoming an increasingly common sight in US cities like Philadelphia, Chicago, and San Jose.
Walmart, meanwhile, has said it is expanding its drone delivery service to 150 additional US stores, including in Los Angeles, Miami, St. Louis, and Cincinnati. The program, small relative to Walmart’s roughly 4,000+ supercenters and $700 billion business, is likely to remain focused on lightweight payloads requiring super-fast delivery.
Still, the moves highlight the continued push toward autonomous delivery.
The So What
While fully autonomous delivery is likely many years away, it could be a gamechanger for large grocery retailers in particular. And they can prepare now for the impact of AI, automation, and robotics on the last mile of grocery deliveries.
For years, grocery delivery in the US has been strategically necessary but economically frustrating. Consumers love the convenience, but the math has rarely worked due to low margins (often 2% to 4% of EBIT) and high labor costs, variation in order size, and cold-chain complexities.
That may be about to change. Artificial intelligence, robotics, and autonomous delivery technologies have the potential to transform grocery delivery from a labor-heavy service into a more cost effective, scalable infrastructure platform. “There is ultimately a tipping point when this starts to really scale,” says Brad Loftus, a BCG managing director and senior partner. “When that happens, there are big implications for store networks and formats.”
The shift is likely to occur in three phases:
- Wave 1: AI Optimization (now–5 years) AI improves forecasting, routing, and substitution to reduce inefficiencies by 10%–20%, stabilizing margins without materially changing labor needs.
- Wave 2: Automation and Micro-Fulfillment (5–10 years) Robotics and dark stores (locations closed to customers that instead fill orders) cut picking costs by 30%–50%. Back-of-store space expands and front-of-store space may shrink.
- Wave 3: Fully Autonomous Last Mile (10+ years) Autonomous vans, sidewalk robots, and other technologies reduce last-mile costs by 30%–50% and turn delivery into scalable infrastructure with faster speed, higher utilization, and expanded geographic reach.
As these waves unfold, the economics of grocery delivery change in significant ways:
- Cost per order delivered declines. As labor costs fall, retailers can lower delivery fees—making smaller basket deliveries viable.
- Penetration increases. Grocery delivery penetration in many higher-income markets is quite low today—in the mid-single digits. With lower costs and improved reliability, penetration could plausibly reach 20%–30% overall. In urban markets in the US, penetration could be higher, closer to the levels seen in some cities in Asia that have lower labor costs such as Seoul and Shanghai.
- Competition intensifies. If autonomous delivery scales quickly, grocery retailers could face intensifying pressure in the form of speed, price competition, and delivery fee contraction.
Stay ahead with BCG insights on the consumer products industry
Now What
As the evolution toward autonomous grocery delivery advances, retailers should take six steps today to prepare:
- Redesign network strategy. Conduct a forward-looking network analysis assuming 20%-30% delivery penetration. This includes evaluating what new network nodes will be needed (for example micro fulfillment centers), which existing stores could become primarily fulfillment sites versus focused on in-store customer experience, and what stores may no longer be financially sustainable and may need to close.
- Pilot automation. Invest in autonomous delivery pilots to ensure the company moves rapidly up the learning curve. Some smaller players may not run pilots directly but should closely track how the pilots of competitors—and potential partners—progress.
- Prepare for a new labor mix. Develop workforce transition strategies that would address changes in the employee base including fewer drivers ormanual pickers and more data and AI pros.
- Build partnerships. Few grocers will build autonomous fleets alone. Most will need to link up with AV and robotics firms while maintaining control over customer relationships and data.
- Stress-test economics. Grocery players should assess multiple scenarios—from moderate autonomy adoption to overcapacity with intense price wars—and craft strategies to ensure resilience under each one.
- Develop action plan. This must be done with clarity on near term piloting and a more general long-term direction. That long-term view should be “written in pencil” and adjusted as AI and robotics technology evolves.
