DoorDash’s $17B Instacart Deal Is Really About AI’s Hardest Job

An AI assistant can plan your dinner. It still needs someone to pick the avocado.

On October 29, 2026, DoorDash and Instacart announced a definitive agreement for DoorDash to acquire Instacart at $65.00 a share, an equity value of about $17 billion. It is a proposed deal, not a completed one. It needs Instacart shareholder approval and regulatory approvals in the United States, Canada and Israel, and the companies expect to close at the end of 2027 or in the first half of 2028. Until then, they operate as separate and independent companies.

DoorDash has already explained the bet. While other companies compete to build your personal AI assistant, its shareholder letter says it wants to be the physical network that gets tasks done in the real world.

Follow one grocery order

Imagine asking an assistant for dinner ingredients. It builds a list, and a few taps later the order is on its way. On the screen, that looks almost finished. In the store, the work is just beginning.

A shopper has to find the right products, judge whether an avocado is ready for tonight and decide what to do about the item missing from the shelf. A convincing answer in a chat window cannot tell you whether the replacement that reaches your kitchen will work for dinner. The promise only becomes real when the bag reaches your door.

More than a shopping app

DoorDash’s third quarter shareholder letter calls this the battle for atoms. The company still builds AI ordering tools, including Ask DoorDash, Text DoorDash and a Model Context Protocol interface for external AI agents, including workplace tools for corporate ordering. Its stated strategic priority is the physical network those tools can reach.

Instacart would bring more than a second shopping app. Its Enterprise technology runs more than 380 grocers’ ecommerce sites and adds fulfillment, in-store tools and Carrot Ads for retailers’ own sites. DoorDash brings its delivery network and DashMart Fulfillment Services. DoorDash says it expects to keep Instacart’s brand, app and enterprise technology products after closing.

So imagine the same dinner order starts somewhere different tomorrow: in a grocery app, on a supermarket’s website or with an assistant. The front door changes. The physical jobs behind it do not. I make the longer case in DoorDash, Instacart and the physical layer of AI commerce.

Why the money might work

A warehouse needs space and people before the next order arrives. Running more orders through the same operation spreads those fixed costs across more baskets, and delivery routes become denser. DoorDash’s letter names that scale as one reason Instacart appeals, and says that when it manages a grocer’s inventory itself through DashMart, fulfillment accuracy exceeds 99%.

Scale is an attraction and a challenge. Combining two companies does not automatically make a difficult grocery order easier to fill.

The synergy target and reinvestment

DoorDash says it sees a multi-year path to at least $500 million in annual synergies, net of investments. That is a target, not savings already delivered. It also expects to reinvest more than 100% of the synergies in the first year after closing, with positive net synergies from the second year. The early plan is to spend on improving the business rather than treat every projected efficiency as immediate profit.

The local grocer’s choice

For a local grocer, renting online ordering, delivery and advertising can be the practical way to compete with retailers that build their own. But when more of the software, advertising and fulfillment comes from one provider, switching becomes a bigger business decision.

DoorDash has committed that grocers will remain free to work with any delivery platform, with no grocery exclusivity. That commitment matters. So will the terms and performance retailers actually experience.

What still has to happen

Shareholders and regulators have decisions ahead. The merger agreement allows extra time if antitrust reviews run long, and DoorDash would owe Instacart $1.25 billion if antitrust obstacles end the deal. Until it closes, a proposed combination is a plan.

My test is whether orders arrive more reliably, whether retailers get useful tools at economics that work for them, and whether outside AI assistants send meaningful business through the network. Those results will say more than a launch presentation.

The business behind the screen

Come back to the avocado. The assistant may help you decide what to buy. The valuable business could be the one that makes sure the right thing arrives. The shopping screen can keep changing while the physical work becomes more important.

Would you let an AI assistant choose your grocery substitutions?