The past two years have seen unprecedented upheaval in Tech hardware supply chains. Component shortages have forced purchasing executives to rethink supplier relationships in unimaginable ways prior to the pandemic. Indeed, there is much more interest in attaching the "strategic" prefix to supply chains than ever before. (Will it last through a pending downturn?)
Right on cue, experts have published several books in the last year describing the best practices for managing strategic supply chains. It is hard not to chuckle at catchy phrases like "full stack" and "360-degree spotlight view." Nor can one help sense a bit of self-serving rhetoric about the importance of supply chains during component and material shortages. After all, supply chain theory is not new. There is a rich history in business literature on strategic supply chains. It begins with the Nobel prize-winning economist Ronald Coase. More recently, Kate Vitasek and her team at the University of Tennessee have done yeoman's work to advance the discipline.
But why are all these "new" supply chain books written for buyers managing their suppliers and not the other way around? Ultimately, supply chain management boils down to bargaining power, and buyers usually have more of it. As a result, it is much harder to discuss supply chain strategy from the supplier's perspective. How do suppliers of commodity parts and processes become "strategic"? What steps does one take to get to a strategic supplier's presumably higher margins?
RCD Advisors has spent over 22 years working with clients on the supplier side of the bargaining table. We have worked with clients to help them understand their competitive position. There are no three-point plans, no catchy titles, and no books pending.
Improving the standing and margins of a commodity supplier to a strategic supplier is hard. But there are repeating business patterns where suppliers deliver higher value. These blog posts are a reference to the firm's perspective on navigating the challenges in the tech hardware supply chain. So, it is worth taking a break from the project work to explore how modularization can help a supplier cross the continuum to a strategic partner.
Before diving in, it is worth defining the sourcing continuum in better detail. Again, the insights from the research team at the University of Tennessee are exceptional.
Most downstream purchasing managers strive for a transactional view of their supply chains on the left side of the graph. It is how buyers typically purchase electronic components. Supply and demand imbalances drive pricing dynamics.
At the far-right end of the spectrum is when the supplier becomes such an essential partner that it forces the buyer to integrate operations. There have been many examples in the Tech hardware industry where a buyer has taken an equity stake or acquired one of its suppliers. For instance, Apple acquired Intel's RF business in 2019 to reduce its dependencies on Qualcomm for 5G RF front ends (It hasn't worked out so well yet). In 2017, Qualcomm formed the RF360 joint venture with TDK for RF filters so that the company could integrate them with their RFFE ICs. In 2019, Qualcomm acquired the rest of TDK's interest to secure its capabilities ahead of the 5G market transition.
Between the extremes of transactional to investment are relational business models. In this domain, strategic suppliers have a vested interest and are partners in the success of their customers. While that may sound a bit too pleasant and harmonious, strategic suppliers do generate higher returns on invested capital through partnership arrangements that fortify the buyer's captivity. The buyer benefits from reduced switching costs (because partnerships are hard to establish). At the boundary between transactional and relational is the differentiated "preferred provider" model. This relationship is a sweet spot for electronic component suppliers. And it forms the investment thesis for developing modules as part of an extended product portfolio.
What is a module?
Engineers define a module as a partitioning tier of an electronic system. Modules improve assembly efficiency, interchangeability, expansion, and upgradability. The canonical example is the memory "modules" plugged into PCs or server motherboards.
But products like memory modules provide the lowest baseline value. For a supplier, taking components off the main system board and repackaging them into a subassembly is a "value"-added service. The supplier trades value by reducing coordination costs associated with assembling off-the-shelf components. Supplier operational efficiencies, usually the product of economies of scale or scope, can result in lower costs. The buyer assumes a higher risk through external sourcing for this lower cost.
A module supplier crosses the barrier into more strategic relationships through differentiation. A differentiated merchant module is an optimized electronic assembly. These modules deliver a complete electronic function with optimized performance enabled by the segregation from the main assembly. Optimization is critical and distinguishes a high-value module from the benefits of value-added services.
The most apparent avenues for optimization in a module are in the interconnections. Placing components closer together on a dedicated substrate reduces stray coupling and increases signal and power integrity. It is why voltage regulator modules, RF modules, and high-speed interconnects on modules (between memory and processors, for example) are already significant parts of the electronics supply chain. But the optimization parameter can vary. For example pluggable optical modules optimize mechanical alignment.
Vicor, a pioneering power module company in the US, has been using this playbook for decades. The company's latest ChiP™ (Converter housed in Package) power modules offer the highest power conversion density in the market. Vicor is unique in that the company's competitive advantages lie in its unique understanding of system architecture, power electronic topologies, power semiconductors, passives, and package assembly integration to deliver an optimized solution inside a package. The company has developed a custom panel-level overmolding process for the ChiP module and manufactures everything internally.
Likewise, Murata, Qualcomm, Broadcom, Skyworks, and Qorvo are major RF front-end module suppliers for smartphones. There is a long history of RF integration in handsets enabled through the supply of dedicated modules in the RF signal chain. IC makers often offer modules as a way to simplify the design process for their customers.
Power, RF, and optical module businesses are already a $30bn-plus market within the electronics supply chain. Depending on how wide one casts the net, the value of modules in the electronic supply chain can easily exceed over $100Bn.
There are other insights about module business development (all insights from Robert Coase):
Modules have standard interfaces. Without standardization, a supplier is fabricating a custom subassembly. Any tangible or intangible investments must be amortized over a single customer and usually a smaller volume. There is nothing wrong with manufacturing a custom subassembly for one customer. Smartphone makers assemble handsets with dozens of custom subassemblies from suppliers (although the OpenRF Association is trying to change that). But the economics are different. A successful module business is designed once and made for many. In contrast, a custom subassembly is designed once for one customer and redesigned again for every other customer.
The merchant module value has to be a small percentage of the end system value. The whole thesis for a merchant module business breaks down if its value is too large. In that scenario, a buyer would either be absorbing too much sourcing risk or lack the strategic bargaining leverage to matter. A buyer with high bargaining leverage would make investments to integrate capability. This situation would move the relationship to the right in the sourcing continuum model. On the other hand, a component supplier would already be earning high margins selling a product to a buyer with low bargaining leverage, and the whole discussion would be moot.
Modules should contain components that manufacturers cannot integrate into silicon because it is too costly or not workable. One of the prominent historical use cases for modules is to jump ahead of Moore's law. For example, the old multichip modules on ceramic substrates integrated several ICs because it was too costly to incorporate in one large silicon die. This same scenario is happening today with more advanced packaging (Apple's M1 Ultra uses two M1 Max chips stitched together to create a more powerful processor). The rationale for heterogeneous integration is to deliver complete assemblies from different fab processes.
The number of integrated components is large enough to discourage a discrete solution. There are plenty of simple two or three-chip analog ICs. For example, sensor makers often integrate a control IC with the sensor. If the buyer could assemble the solution discretely without any loss in optimization, the value added from integrating these components boils down to reducing the coordination costs described earlier. In contrast, a module with a complex subcircuit that incorporates many more elements provides more avenues for optimization and likely a more compelling value.
Merchant SiP Modules
Modularization is a worthwhile product expansion strategy for discrete component suppliers to explore. These organizations have more opportunities for delivering value through modularizing their components in the downstream assembly. The rationale is simple. The technology maturity of typical electronic discrete components limits the pace of innovation and lowers the return on investment. But the return on intangible (and scarce) engineering capabilities is much higher if there are avenues to optimize downstream performance through the interconnection with other components. Moreover, that optimization can create technical differentiation and value above and beyond the baseline coordination benefits.
This engineered optimization is equivalent to a buyer renting a part of the supplier's engineering talent (i.e., intangible investments) to create value. It moves the sourcing relationship into a preferred supplier model where buyers and suppliers have vested interests in each other success.
But one has to be sober about who is creating value. For example, if a component supplier uses a merchant SiP process, then part of the optimization value lies outside its internal capabilities. Presumably, that capability is available to competitors and customers as well. You don't need a degree in supply chain management to know that scarcity increases value and ubiquity reduces it.
And this is why the emerging merchant market for SiP assembly processes is narrowing the traditional avenues for component vendors to capitalize on module product expansion. A merchant SiP process makes integration processes accessible to any organization in the supply chain. It allows organizations to design their custom semiconductors, their custom electronic assemblies, and their custom products. In addition, SiP isolates and reduces the value contributed by a module supplier to the intellectual property associated with that optimization.
But this also brings up an interesting alternative and parting thought. The thesis for component makers is to capitalize on the internal intellectual property by expanding product lines with optimized module products. But, in a SiP world, component suppliers may be able to deliver that value by providing IP and design services. Suppliers would create value by offering readymade IP and services for embedded components in a standard SiP process.
Contact RCD Advisors for more and customized analysis on growing value in the electronics supply chain.