AI may soon order your coffee as blockchain handles instant payments

TLDR

• Kevin O’Leary says AI will automate orders while blockchain completes payments.
• Current blockchains like Ethereum struggle with high retail transaction volumes.
• DAG-based platforms like Hedera aim to process payments faster than blockchains.
• Agentic AI already manages shopping lists and price tracking across retailers.

Artificial intelligence could soon be making everyday purchases, but blockchain might be the missing link to make it happen. Venture capitalist and Shark Tank co-host Kevin O’Leary says AI has the potential to automate tasks like ordering coffee, groceries, or meals. However, for the system to process payments fast and at scale, it needs blockchain. But today’s blockchain networks still have limits, according to O’Leary.

AI Could Automate Retail Orders Through Voice

Kevin O’Leary explained how future AI systems could handle purchases without human involvement. In a video posted to X on October 19, he said, “You simply talk to your phone and say, I want a tall, low-fat latte, please. I’m going to be there in 90 seconds.”

He described a process where AI detects the user’s location and picks a nearby café to place the order. The AI would handle the entire transaction, while blockchain would complete the payment. The retailer would already know the customer’s name when they arrive.

O’Leary believes this model could be applied across various sectors including grocery chains, fast food restaurants, and retail stores. He added, “That’s the next revolution in business. Faster, smarter, fully on-chain.”

Current Blockchain Systems Face Technical Barriers

While the idea sounds simple, O’Leary pointed out the major hurdle: payment networks that can’t handle high volumes. He said platforms like Ethereum still process transactions one-by-one, which creates slowdowns during peak demand.

“It’s a long highway on a way to a toll road for authentication,” O’Leary said. He added that when too many transactions happen at once, they “get stuck at the toll” and cause delays and high fees.

O’Leary emphasized that this limitation stops current blockchain platforms from supporting mass-market AI transactions. He said a more scalable solution is needed, especially for big retailers like Walmart and Target who require millions of transactions per day.

Directed Acyclic Graphs Could Be a Solution

Some blockchain networks are already testing other ways to process payments faster. O’Leary mentioned a data structure called Directed Acyclic Graphs (DAGs), which is used by platforms like Hedera and Nano.

Unlike Ethereum’s block-based structure, DAGs allow transactions to be verified in a web-like system. This setup can handle multiple payments at the same time without slowing down.

These technologies, however, are still in early stages of adoption. Hedera and Nano have not yet reached the user base of Ethereum or Solana, and they have yet to be tested at full retail scale.

O’Leary said he is waiting for a blockchain project that can support millions of transactions per day, all running independently and simultaneously.

Agentic AI Already Handles Simple Tasks

While AI-driven payments are still in development, agentic AI systems are already helping users manage daily routines. Kyle Okamoto, CTO at decentralized platform Aethir, said his wife uses an AI assistant for grocery tracking.

She speaks to it when items run out, and it tracks how often she buys them. “It says, ‘you run out of milk every week to eight days,’” Okamoto said. Her assistant also compares prices across stores like Whole Foods, Target, and Amazon, choosing the cheapest option.

These tools are already saving time, but they still rely on humans to complete payments. O’Leary said the next step is a blockchain solution that can handle the transaction automatically and at low cost.