Search Captions & Ask AI

AI Could Make Supply Chains Smarter But Not Disaster-Proof

October 22, 2025 / 10:27

This episode discusses the current state of supply chains, the impact of a recent aluminum plant fire in New York, and lessons learned from the pandemic. Gad Alon, a professor at the Wharton School, shares insights on how automakers are adjusting to supply shortages, particularly in aluminum, which is crucial for vehicle production.

Alon explains that 40% of aluminum for US manufacturers comes from the affected plant, leading to long recovery times and potential price increases for vehicles. He emphasizes the importance of resilience in supply chains, noting that many companies still prioritize cost efficiency over redundancy and risk management.

The conversation also touches on the role of artificial intelligence in improving supply chain management. Alon mentions that while AI cannot prevent disasters, it can help in planning and predicting supply chain needs more effectively.

Alon reflects on historical lessons from World War II regarding supply chain efficiency and how companies should adapt their strategies for the future. He highlights Amazon as a model for integrating resilience and agility in supply chains.

The episode concludes with a call for companies to rethink their supply chain strategies to ensure long-term effectiveness and stability.

TLDR

Gad Alon discusses supply chain challenges post-pandemic and the impact of an aluminum plant fire on automakers.

Episode

10:27
00:00:00
Recently, an aluminum plant in New York State caught fire, destroying a large section of the facility. That plant was
00:00:08
a supplier to several automakers in the development of their vehicles. Events like that and others have an impact on
00:00:15
the supply chain and force companies to have to make adjustments. Of course, the
00:00:19
pandemic really challenged the supply chain and forced many entities around the supply chain to rethink how they
00:00:27
would operate on a daily basis. As for where we are at the moment when we think about supply chain, pleasure to welcome
00:00:33
in Gad Alon who is a professor of operations, information and decisions here at the Wharton School. Gad, great
00:00:40
to catch up with you again. >> Yeah, great to be here. Then >> what do you think is the general state
00:00:45
uh of supply chain right now and and what did the pandemic teach companies about their supply chains?
00:00:53
>> It's a it's a great question and two different question and one question. I
00:00:55
will say most supply chains are are I won't say deep trouble that that's that
00:01:01
will be probably an exaggeration but there are still there are suffering from some implication of what started even
00:01:07
before the pandemic which is you know more and more issues that are more and more frequent from climate to
00:01:12
geopolitical risk to demand fluctuations to suppliers having issues to now you you add tariffs to that and then so most
00:01:21
firms are are struggling to just get their supply chain to stabilize but then you have continuously the other
00:01:28
issues that just happened and that's of like the the fire that you're describing
00:01:31
is just an illustration of one more issue that one needs to prepare for. >> Okay. So for the automakers that uh did
00:01:39
get aluminum supply from that plant, how are they adjusting to not having that access that they've had or and are used
00:01:46
to having for such a long period of time? >> Yeah, I think so for many of them I
00:01:50
think that's going to take a very very very long recovery and and I think that's to some extent really the the
00:01:56
first question. So just to get the numbers right so 40% of the aluminum to US manufacturers come from that plant.
00:02:03
So we might ask why do we have such a concentration in one place but then my understanding is only Q1 in 2026 they
00:02:11
will be able to recover back to the full capacity and even then it's questionable. So here multiple things
00:02:18
need to happen. So the first thing is many places many of these manufacturers will have to think about we like what we
00:02:23
like to call time to recovery. How long will it take until they can recover? And
00:02:27
then you start rationing. You start saying, "Okay, which models are going to
00:02:32
get the aluminum they need and which models not so much." But you remember in
00:02:36
during the COVID days, one of the things that we had shortage of were semiconductors
00:02:41
chips. And there what you saw is that most manufacturers prioritized more expensive models. So Audi prioritized Q7
00:02:49
and and Toyota prioritized the Lexus and the like. And and so that's potentially
00:02:54
what we're going to see. We might see cascading effects of suppliers that are
00:02:58
now going to get, you know, fewer orders and now they're going to reduce their
00:03:04
manufacturing only to get caught back again where the aluminum is going to be back at capacity. Firms are currently
00:03:10
probably scrambling to find spot markets that as we know has implication on price
00:03:16
as well. So what you'd expect is expensive models will become potentially more expensive, cheap models to either
00:03:21
not exist or becoming expensive. It will take a while for for this market to recover or what we're going to see is
00:03:28
is those who actually were impacted will have to shrink their margins while those
00:03:32
who are not impacted will be able to actually gain market share. Let me go back to something you said at the
00:03:37
beginning of that answer and about the fact that such a a large percentage of the supply is in one location. So
00:03:45
obviously that's something that needs to be looked at as well when you think
00:03:49
about we're talking about a period of time of growth in this economy right now
00:03:54
that there needs to be a consideration of not only repairing the facility that was damaged by fire but looking at
00:04:01
another facility the potential of a larger buildout uh in this sector. >> Right. So I think that goes back to what
00:04:09
something you asked me in the beginning which is what did we learn from the the pandemic and and I'm sad to say nothing
00:04:16
but it's not that we didn't try to learn there's just a inherent trade-off between resilience
00:04:23
and cost cost efficiency and one of the things that many CEOs came out of COVID is with the idea that we need to build a
00:04:32
more resilient supply chain but most of them do not understand or don't appreciate the fact that to build a more
00:04:38
resilient supply chain you need to to increase costs. You need to have excess capacity. You need to have redundancy.
00:04:43
You need to have many a small overlapping plans rather than one big ones. The main reason that we have such a
00:04:50
concentrated risk in one place is that firms optimize for economies of scale as a way to reduce cost. They try to
00:04:57
optimize for quality and compliance and it's very hard to qualify multiple suppliers. You try to be just in time.
00:05:04
So you try to concentrate everything in one location. All of these are best practices if what you try to do is
00:05:09
reduce cost. This is not what you try to do when you try to be more resilient. But as long as the customer still wants
00:05:15
cost and unwilling to pay for resilience, that's what we're going to get.
00:05:20
>> How will artificial intelligence play a role in supply chains as we move
00:05:24
forward? >> Yeah. So it's a great question. As we know, AI is the solution for
00:05:30
everything. Well, not so much. So, AI cannot prevent the the fire. I don't think so at least.
00:05:38
But it can prevent concentration. It can help. You know, one of the things that has been always the big promise is this
00:05:44
notion of digital twin. And digital twin is the idea that you can actually run continuously simulation real time and
00:05:50
understand the implication of everything you do. These are what very very hard to
00:05:54
build and underutilized because most humans just do not understand the implication of what they see. And but
00:06:00
now we're going to see more and more AI agents run supply chains and we're still
00:06:03
a few years away from that. But there are already firms that are developing AI agents. Then these AI agents are are
00:06:09
actually capable of saying that. So we might see inventory that's more strategic and more in better places.
00:06:14
It's trying to understand the implication of this longtail implications. So you know we're not
00:06:19
preparing for a fire, but we should prepare for the fact that 40% of our supply is going to come from one
00:06:25
location and and maybe we we may want to hedge that. maybe. So I think AI is going to make us more predictive so
00:06:33
easier to predict things. It's going to be easier to try to orchestrate things
00:06:37
and potentially also do more preventive maintenance and predictive maintenance and and and so that's and and that
00:06:43
potentially may prevent some of these longtail events. We cannot prevent tariffs but we can prevent some of these
00:06:48
things. And I think the other side to this is that you know so many of the stories around supply chain obviously
00:06:54
because of the pandemic and and recovering from the pandemic have been negative and the excess cost and and the
00:07:00
negative impact that you know we've seen because of supply chains. If you flip it
00:07:05
the other way, thinking about how positive a strongly run, efficiently run supply chain, what that can mean to not
00:07:15
only economies but to cultures around the globe. >> Yeah. So that's a it's a great question
00:07:22
to some extent. I mean I think it's a there there are two sides again to that
00:07:26
story. One is that of course we read only negative things because it's a little like okay you don't read about a
00:07:32
great defense. you read about great difference only when about difference only when they many there are many goals
00:07:37
scored on them that's a story of of supply chain but you're right I mean I
00:07:41
think firms cultures that managed to actually build robust supply chain meant that they actually you know I mean let
00:07:48
me just tell a very very simple I mean you know very simple story which is you know you have to go you can go back to
00:07:53
World War II and and you realize that the countries that had better industrial base with better supply chain those are
00:07:59
the ones that managed to ultimately win the war. It was not just about having being able to have the best army, but
00:08:05
the best army that can actually replenish very very quickly its supply and and so I think we we are many years
00:08:10
away from that and so we forgot many of these lessons and and and I think we we are start trying to relearn some of that
00:08:16
now. COVID was a route awakening for many but but not apparently not for everybody and and definitely at that's
00:08:23
at least where we see at the federal level some awareness now for that topic but again it's it's exactly as you say
00:08:29
it's usually through negative stories not through positive stories. So then how should companies think about supply
00:08:34
chains right now having gone through everything that they've gone through in
00:08:38
the last few years but thinking about the next 15 to 20 years about being as effective as they possibly can because
00:08:46
it obviously is something that's going to impact their bottom line. >> Yeah. So I think I'm I'm it's
00:08:51
interesting you know the the best example for that and and I think the firm that you'd expect everybody to try
00:08:56
to copy is Amazon right and Amazon understands supply chain almost like better than anyone else Amazon and
00:09:01
Walmart and but Amazon specifically Amazon has a good website but it's not such a great website but the main thing
00:09:07
is that when you order from Amazon something today you will get it today and if you order it and you're willing
00:09:12
to pay you might get even within a few hours and within a few minutes and Amazon is definitely a firm that I mean
00:09:20
the amount of cash in their supply chain is then used to fund the rest of the business and every customer Amazon
00:09:26
Amazon customer is is an Amazon customer not because how well the website is run
00:09:31
that's great that they have everything but you have they have everything and
00:09:33
you'll get everything and so you'd expect more firms to understand that you
00:09:37
can actually be both resilient and agile and cheap at the same time if you just continuously think about how do you
00:09:44
improve everything, how do you integrate everything, how do you make sure that there
00:09:48
constant loop and you continuously try to optimize for both proximity and cost and and and it's possible but it
00:09:55
requires a constant effort to do that and we just don't see firms most firms
00:09:59
have the share of mind of leadership into doing that Gad great to talk to you again thanks very much for your time and
00:10:06
your insight today all the best always great to be here thank you Gatalon professor of operations information and
00:10:12
decisions here at the Wharton

Episode Highlights

  • Aluminum Plant Fire Impact
    A fire at a New York aluminum plant disrupts supply chains for automakers.
    “40% of the aluminum to US manufacturers come from that plant.”
    @ 02m 00s
    October 22, 2025
  • Lessons from the Pandemic
    Gad Alon discusses what the pandemic taught companies about supply chains.
    “Nothing but it’s not that we didn’t try to learn.”
    @ 04m 16s
    October 22, 2025
  • AI's Role in Supply Chains
    Gad Alon explores how AI can help manage supply chain risks.
    “AI cannot prevent the fire, but it can prevent concentration.”
    @ 05m 35s
    October 22, 2025
  • Historical Supply Chain Lessons
    Gad Alon reflects on how supply chains influenced World War II outcomes.
    “The countries with better supply chains managed to win the war.”
    @ 07m 59s
    October 22, 2025
  • Amazon's Supply Chain Strategy
    Gad Alon highlights Amazon's effective supply chain management as a model.
    “You can be both resilient and agile and cheap at the same time.”
    @ 09m 40s
    October 22, 2025

Episode Quotes

  • Most supply chains are suffering from implications of what started before the pandemic.
    AI Could Make Supply Chains Smarter But Not Disaster-Proof
  • To build a more resilient supply chain, you need to increase costs.
    AI Could Make Supply Chains Smarter But Not Disaster-Proof
  • AI cannot prevent the fire, but it can prevent concentration.
    AI Could Make Supply Chains Smarter But Not Disaster-Proof
  • The countries with better supply chains managed to win the war.
    AI Could Make Supply Chains Smarter But Not Disaster-Proof
  • You can be both resilient and agile and cheap at the same time.
    AI Could Make Supply Chains Smarter But Not Disaster-Proof

Key Moments

  • Aluminum Plant Fire00:03
  • Supply Chain Challenges00:12
  • Pandemic Lessons00:19
  • AI in Supply Chains05:22
  • Historical Context07:59

Tension Over Time

Words per Minute Over Time

Vibes Breakdown