FedEx Chiefs on AI: Speed isn't the problem, customs is
FedEx’s Raj Subramaniam and Vishal Talwar say AI must be led by the CEO, and that customs clearance, not flight speed, is logistics' real bottleneck.

Raj Subramaniam, CEO, FedEx and Vishal Talwar, EVP & CDIO of FedEx, and President of FedEx Dataworks
FedEx CEO Raj Subramaniam and Vishal Talwar, Executive Vice President, Chief Digital and Information Officer (CDIO) of FedEx, and President of FedEx Dataworks, told the WSJ Leadership Institute Technology Council that artificial intelligence at global scale cannot succeed without the CEO at the helm, and that the industry's biggest drag on efficiency is not how fast a package moves, but how long it sits still. On September 14 in New York, the two leaders sat down with Alan Murray, founding president of the WSJ Leadership Institute.
AI has to be a CEO's job, not just the CIO's
Asked bluntly who runs AI at FedEx, Subramaniam did not hesitate to hand himself the responsibility. "I don't think any company today can succeed in AI without the CEO taking active leadership role in AI," he said, arguing that the line between business strategy and technology has all but disappeared. "The gap between business and technology is fast moving away. We can't have a business strategy without technology and the reverse is also the case." He called AI "the defining technology of the era" and said the strategy "has got to be top-down." Talwar, for his part, offered no resistance to the framing: "I'm 100% cool with anything Raj says here."
Two petabytes a day, and a network built to notice
Subramaniam described FedEx's data as the fuel behind its AI ambitions, generated by a network that moves roughly $2 trillion of commerce a year. "The fuel for AI is data, and two petabytes of data every single day represents the movement of the world's most premium supply chain," he said. He traced the shift in how supply chains are perceived, noting that after the pandemic they became "a cool thing" and a "boardroom conversation," pushing FedEx to expand its ambitions beyond internal efficiency. Predictability for customers is one payoff he pointed to: FedEx's healthcare transportation business, he said, has grown 30% over the past couple of years as AI-driven forecasting lets healthcare customers "schedule their workflow significantly better."
Talwar broke the strategy into three layers: efficiency, customer differentiation, and, what he called the "true transformative" piece, new business models. On efficiency, he cited concrete gains: air maintenance research time cut "from 30 minutes to three minutes," a 90% improvement that gets aircraft back into service faster, and "tens of millions of dollars of annual revenue" recovered by using AI to catch mis-tracked, non-standard packages that should have carried a surcharge. On the customer side, he said certainty is what matters most: "The one thing our customers care most about is don't surprise us. They want certainty," pointing to FedEx's ability to guarantee delivery within a two-hour window on shipments moving thousands of miles from Asia to the US.
The real friction isn't speed; it's what happens when a package stops moving
Pressed on whether faster delivery is even possible anymore, Subramaniam suggested FedEx is near the physical limits of aviation. "We are leaving China and delivering to every zip code in America overnight," he said. "Now we are at the speed of aviation technology and that's kind of it at this point." Asked, half-jokingly, whether teleportation was next, Talwar redirected the conversation to where he says the real opportunity lies; not in motion, but in the gaps between it. "The focus that we have is not about the speed while an object is in motion," he said. "I can put anything from China to US in 13 hours, 14-hour flight time. Then it sits in customs for clearance for another four days. The amount of time that's spent in something waiting when it's not in motion, that's where the friction in supply chain is."
Both executives were emphatic that this is not a FedEx-specific failing. "This is a problem in industry at large. It's not a problem that's just for FedEx," Talwar said, putting a number on the scale of the issue: "We are leaking about $1.9 trillion of annual value just because of these inefficiencies and friction points within supply chain. That is the problem we are out to solve." Subramaniam echoed the framing when asked how far along FedEx's AI transformation is. "I would say we are early innings," he said. "We have been building infrastructure... but there is a long way to go. The $1.9 trillion of supply chain inefficiency that's in the global system; we want to be able to have a crack at that."
Robots in the warehouse, autonomous trucks on the highway
On physical AI, Subramaniam described FedEx's giant sortation hubs as "an inventory in motion system" that is almost entirely automated except at two stubborn points: loading and unloading trucks. "Packages of different sizes, shapes, weights coming through. You don't know what's coming exactly. And then it has to figure out how to actually play Tetris on the truck," he said, adding that FedEx has deployed 60 robots in one facility to test solutions ahead of peak season. He also pointed to autonomous trucking on highway middle-mile routes, developed with Aurora, connecting facilities without putting unmanned trucks on interior roads.
Talwar tied the robotics push back to data, arguing that FedEx is positioned to build something the logistics industry has never had. "A real-world model for supply chain or logistics does not exist today. And who's going to create it? It has to be one of the physical operators of that network, which is us," he said, noting that even FedEx's shipping labels remain "passive" and not yet smart devices.
Physical AI deployment at the FedEx Hagerstown hub in Maryland
Growth, not job cuts, for now
Asked whether automation would eventually mean fewer employees, Subramaniam pushed back, noting FedEx added 25,000 new US employees last year. "I think you're not going to see AI delivering packages anytime soon," he said. "The last mile, the last 10 feet, those are even more complex problems to solve. And these robots are going to work hand-in-hand with our team members." He summed up the logic in three steps: "AI as a driver of growth. Growth drives packages. Packages drive people."
Geopolitics: "advanced calculus" every day
Operating in 220 countries, Subramaniam said geopolitical shocks are now a daily operating reality, from rerouting around Russian and Ukrainian airspace to suspending its Dubai hub in favour of a more southern route through the Middle East. "Every twitch and turn that happens anywhere in the world has an impact on FedEx," he said. "As the world has got more complex, before basic math would have done. Now we need advanced calculus, and there are very few people who do advanced calculus, and we are one of them." Talwar said AI has not yet been applied to geopolitical risk directly, but that FedEx Dataworks is increasingly advising customer CEOs on sourcing, tariff scenario modelling and lane strategy.
On AI's risks, no daylight between the two
Closing on the debate over whether AI will drive prosperity or catastrophe, Talwar said there was no disagreement between him and Subramaniam. "We both think that it's in our hands and it's up to us to apply it for good," he said. Subramaniam was equally direct: "We're on the optimistic side, let's call it that." Talwar pointed to FedEx's role in the newly formed "AI for Good" commission, which brings together public figures and industry leaders to address equitable access and responsible deployment.

