Take one global supply chain crisis. Add in a semiconductor shortage, flavored with a trade war between the US and China. Blend together and the end result is some serious challenges facing companies such as Qualcomm.
As Brent Wilson, the firm’s SVP Global Supply Chain and a 34 year veteran of the sector, points out, semiconductor shortages and cycles are nothing new to the industry. But this time around it’s different:
Typically, what happens is there's some type of a global event. And there's a dramatic downturn in demand, and therefore, shipments, and then there's a reduction in inventory, there's a reduction in labor within the factories, all these things that go with it. Then there's typically a very sharp rise on the other side of that and that's when we see the shortage where production cannot keep up with the demand for semiconductors. That's exactly where we are today.
One of the things that's not getting talked about very much was in the semiconductor industry, just before COVID, there was a trade war going on with semiconductors between the US and China. Most of the capacity that had been built in the 2000s for semiconductors was built in China. The Chinese Government gave a lot of money to companies to build factories there, and they took advantage of it and built them. But then when the trade war started, a lot of the access to those factories was cut off, particularly for US-based companies. There was a huge shift in what suppliers could be used. That caused a bit of a downturn in that.
The other thing that happened was the US Government put some key buyers of semiconductors on an a restricted list, so they were no longer able to deal with US companies. Typically, US companies do a lot of the design work for the industry. So that was part of the reason that we were already in kind of a downturn, started to come out of it. And then COVID hit and we kind of went into more like a W curve. That hasn't happened before.
For Qualcomm, this all poses problems that need to be overcome, partly to ‘keep the lights on’ and partly to meet the firm’s strategic ambitions. Wilson explains:
Qualcomm is rapidly becoming much more than just a handset company. We do focus on handsets, that's about two thirds of our revenue today, but we also have what we call RF front end. RF front end is the radio-frequency part of the chipset that helps the device talk to networks, either a WiFi or a cellular network. We're also very much into automotive. We want to build the chassis of the future so that cars can communicate with each other and with the infrastructure in your city. But we also want to make sure that we're driving the infotainment, that we're driving a lot of the advanced safety systems, all from one centralised chipset. And then last on the list is IoT.
We want to have everything connected and have it have intelligence out at the edge. Why do you want intelligence at the edge? It's because you want to have things calculate faster, you want to have improved functionality. One of the best examples of this is your smartphone. It's the most advanced connected device at the edge. What it has is a lot of smarts that can do a lot of its own computations before it actually sends data back to the cloud. That's really the wave of the future, having smarter devices that can do more of the calculations in real time, and then send the results of those calculations back up to the cloud.
The challenge for the provider is how to know more and sooner about its supply chain by using a device that is connected to a network wirelessly that has low power consumption. Wilson explains:
What we do is when we develop a product for other markets, we just scale down that same chip, that same design. We use the exact same design, but we may take a few computing cores out of it, or we may take the graphics out of it, because that device doesn't require it. One of the reasons this becomes a complexity is now we have a lot of shared resources. The chip has been designed on the same platform, so it runs in the same factory, in the same technology node, as many of our other products.
The other complication this brings about is that all of these different business units we're trying to support have different levels of products that are important to them. So we may have a mid-to-low tier baseband set that is directly competing with what we see from one of our IoT groups as their premium product that they're trying to roll out in their strategy. So as a supply chain group, we have to make this balance between all these competing resources and understand what's the best strategic return for all the business units across the corporation.
Complexity number two lies in chipsets, he adds:
It's a multi-functionality chip. The baseband can have power chips to go with this as well as RF connectivity chips and other chips, but depending on where the baseband is being sold, by what country, by what customer, by what product they're producing, the chips that go with that baseband can be different because there can be different requirements, either forced by the government or by a different product that they're trying to produce.
Finally there can be situations where the firm is designing an RF chip that might be sharing a node with, for example, an automotive product:
The strategic bottleneck might actually be in the RF section. So we've got a complicated situation here, that sometimes the dog, which is the baseband, is wagging the tail, and sometimes the tail is wagging the dog. We have to understand that and understand what's the best strategic outcome for the company, and make sure we're executing that across the supply chain.
What to do?
Qualcomm has tapped into solutions and services from Kinaxis and Genpact, says Wilson, the latter of which is playing a large part in setting strategic direction:
When the current semiconductor shortage started, Qualcomm was not in a position to handle this, so we didn't understand all the ramifications of constraints on chipsets and how they might actually impact baseband. We, for sure, did not understand the strategic intent of all of our businesses, and how these intersections played out in the chipset. So we contacted Genpact and hired them. They helped us diagnose where we were at today and built some better systems on just what we could do with our current tools. They also gave us some augmented labor to help us just with the amount of workload that was created by that. And then they've also helped us kind of ideate the future system - how could this be better in the future? How could it be more automated? And they're staying there to help us build that.
As for Kinaxis, Qualcomm has been attracted by the completeness of the solution, says Wilson:
We're going to use the completeness to do demand planning, including chipsets of both the dedicated and the competitive chipsets. We're going to do SNMP [Simple Network Management Protocol], so we can understand and see that resource load based on all of the chipsets and make sure we optimize it. And then we're gonna move right into execution, where we can make sure that we're pacing the chipset build at the same cadence as its baseband. And then making choices and changes as things move through the execution window.
The other thing that's going to be super important to us is concurrency. This moves very quickly. There's lots of moving parts to it. So if there's going to be any changes, we need to know about them very soon, and we need to be able to react. We need to be able to see the full picture very quickly from end-to-end. So we will be fully utilizing all of Kinaxis’ capabilities in order to help us solve this problem.
For more diginomica stories from Kinexions22 visit our Kinexions22 event hub. The event was live in San Diego from May 9-11th 2022 and many sessions, including the above, are available to view on-demand until the end of June. Click here to register and view now.