Once an afterthought (or not even a thought) for many, the global supply chain crisis is at the forefront of governments, retailers, and consumers’ minds, especially as we head into the holiday season. Spurred by the demand shock caused by the pandemic, our supply chain woes are not going away any time soon. If anything, Covid exposed how deeply broken the chain has been for years and the work that is ahead for the global community as we try to get back to pre-pandemic supply status.

Before founding Eclipse Ventures, Lior Susan founded LabIX, a hardware investment platform for electronics manufacturer Flextronics, and Aidan Madigan-Curtis, now a Partner at Eclipse, was most recently an executive at Samsara and previously worked at Apple, where she helped launch the Apple Watch. Between their backgrounds and now their focus at Eclipse to bring full-stack solutions to legacy industries, Lior and Aidan have studied the intricacies of our global supply chain for years and the role technology can play in transforming it. Lior and Aidan recently spoke to the Bridgewater Daily Observations Podcast on what is driving supply chain shortages, what it will take to solve them, and how long it will take. Below are a few highlights from their conversation.

Bridgewater: Why don’t we start with backward-looking supply-chains? From a very top-down level, we’re seeing a massive demand shock. Looking ahead, we believe demand will stay strong and there will probably be a big desire from a lot of retailers and producers to restock inventories that are pre-COVID. How long do you think it will take for the adaptation process to take place to adjust to this new world of stronger demand and potentially higher inflation?

Aidan: We’re seeing different parts of the chain choke as time moves on. If you look at what’s going on all the way down to the semiconductor level, I think certain products are more likely to recover faster than others. Whether they can get these products to the shore and into consumer hands is a different question. What we’ve seen happening — and what I’ve heard from my friends still in supply chain who are connected deeply to the semiconductor state — is that a lot of the semiconductor manufacturers over the last five to seven years have been moving their manufacturing nodes from the larger, older nanometer — like 20-plus nanometer space — to manufacturing smaller and smaller nanometer semiconductors. That’s really where the investment has been and these investments are totally new ways of doing things, like laser lithography. They’re not fast to change — it’s a bit more like moving the Titanic. You’re buying equipment that costs billions of dollars and you’re retooling your whole line. So, the chance to change the semiconductor manufacturing back to having larger, older process nodes is not likely. Where you see older process nodes is in the manufacturing of components, like bigger microcontrollers or some of the components used for power circuitry. A lot of these older, larger components were used in things like automotive, where you see longer development cycles because automotive needs to be validated for decades on a road, versus a consumer iPhone that’s going to get thrown out every two years.

All this is to say that I think the supply chain for consumer goods — especially the higher tech ones — is likely to recover faster in terms of manufacturing capacity than stuff like automotive components or anything that goes into equipment that is a little bit older and less cutting edge.

Lior: I’m going to take one step back. I think the world is missing a much larger conversation. This is not a supply chain issue, a semiconductor issue, or a China-US issue. I believe what we are currently facing is a once-in-a-century event — an Industrial Revolution-style situation. You have multiple-stream matters that are basically collaged together: e-commerce growth, electrification of the world (basically everything became electronics), product SKUs, macro situation — call is China-US, COVID, lack of CapEx investment in a significant way in most of the world over the last two decades. We can deal with one of those things and we can also probably deal with two. What is happening, I think, is that we are seeing the stress on the world unfolding in a very different way than what we are used to as a modern society. We are used to pressing the button on our Amazon app and expect it to show up at our door the next day. Many of us don’t think twice about how the inventory is being run, what the micro fulfillment automation is, where the cargo is that has been manufactured, if this is an autonomous or not autonomous vehicle, which algorithm on what data center and which chips are sitting inside the data centers, where this is all produced, and how many minerals you need to pull from earth in order to produce.

I think what we are seeing is consumerization of many, many industries that are using totally different volumes of what the world needs to consume, and that’s all sitting on shaky geopolitical ground. This thing is just a big kaboom. It sucks for the world, but it’s exciting in that it’s going to reshuffle a lot of resources and create a lot of interesting opportunities. Yes, the Infrastructure Bill is great, but when do we think we’re going to see the results of that? Decades from now? Five years from now? This is not software — it’s going to take many, many years in order to see results. Everyone is now jumping through hoops to put more capabilities on their manufacturing, trying to figure out their onshore strategy, trying to buy more lines. This is not a one-slice approach — this is many layers of cake coming together and it’s going to be interesting.

Bridgewater: How do you think this is all going to play out? Where do you think the biggest shortages are going to show up across the supply chain and what does that mean in the short term?

Lior: I think people are coming at this thing with an approach like, “Oh, we’re just going to put a few more lines in. That will fix it, right?” It’s not going to unfold like that. I think onshore manufacturing is real and that we will see decoupling — essentially what happened in the last three to four decades. As Western consumers, we got obsessive about paying 20 cents less on a shirt. Because of this, we don’t know how to design shirts anymore in this country. We don’t know how to do the material science or where the supply chain is coming from. We don’t even have automation around knitting to be able to produce the shirt. So, if Lululemon wanted to do domestic manufacturing, it isn’t just a discussion about moving the factory. You’re talking about cultural change, and we can’t copy what China is doing in the U.S. because we will fail.

The U.S. needs to implement a very different strategy and I believe that strategy needs to be technology first. We are going to deal with a new type of inventory management and people are going to source in a very different way, including retail and e-commerce. This notion of inventory and cash flow cycles are going to change significantly. That will result in a higher interest rate because we are going to use much more cash to put the product inside warehouses due to not trusting the lead time we were able to trust before. I think we will see onshore manufacturing ramp up in a significant way with a lot of that cost flying back to consumers. So, as a consumer, you will be paying more than 30 cents, but guess what, we’ll be able to manufacture it in Salt Lake City vs. Tijuana and be able to deliver it with a truck in the U.S. versus putting it on a boat for three months.

Aidan: To answer that question from a different, more bottoms-up point of view, I think we’ll end up in a inflationary spot for quarters. If you think about where the bottlenecks are, even though there’s a 24/7 operation happening at ports, they can’t get enough trucks out. Why can’t they get enough trucks out to take inventory out of the ports? It’s because there aren’t physically enough trucks and there’s not enough drivers. What does it take to get the drivers into seats to get them both trained and classified into the commercial classifications? It takes both time and financial incentive because the people who are doing truck driving could also be doing warehousing and/or food processing. It becomes this interesting labor arbitrage game for a lot of those people and for certain aspects of the chain — like truck driving — you do need to get licensed, which is not an overnight thing.

I feel like we’re talking on the scale of at least quarters — if not a year or two — before some of the inflationary pressures die down that relate to some of the supply chain bottlenecks and some of the capacity that will get built out both in terms of labor, manufacturing, shipping, and processing.

Question from Bridgewater: Do you have a sense of how much economics have shifted? Because that will certainly be an important factor on how quickly this transition will happen.

Lior: We faced in the last 30–40 years one thing and one thing only in my mind, which was labor arbitrage. That was the game. You go where there is low-cost labor. If I can show you a technology that takes low-cost labor out of the equation, plus move everything I said before: The design knowledge, the process knowledge, the long tail supply channel knowledge and processes, China is not better than US. I won’t need an army of people because I can have an army of robots or an automation line. A great example is Tesla and how they are building a car in a way that people are building computers. They will change the design on the fly and it’s fully automated on the body shop. They’re doing reconfig silicon, so they can use much more general purpose. On the flip side, when Ford or GM commission a line, nobody can touch the line for five years. Tesla is changing the line three times a year. So, you could do it only because there is an economical benefit. You can show the Dow rise there and the technology went into a mature phase where you can use it.

Question from Bridgewater: As this onshoring happens in the U.S., how much of it is going to be purely U.S.? Because again, it’s not a labor question — it’s mostly infrastructure — so can it be domestic production versus shifting some of the productions from China to Mexico, Canada, or Latin America?

Aidan: I’m not sure exact numbers, but in the context of industries where cybersecurity is a major risk or the reliability needs to be high, I think we’d see those move back into U.S. proper and maybe subcomponents will be sourced from cheaper sources elsewhere. Anything that’s lower value without much software technical risk, I think it would make sense for that to be more of a Mexican manufacturing base. Open question about Canada, but I think that the changing landscape from a cyber perspective means that things produced within the country will be at a premium.

Listen to the full conversation on Bridgewater’s “Inside the Research Engine”

Follow Eclipse Ventures on LinkedIn and Twitter for the latest on the Industrial Evolution