Semiconductor supply chain, ecosystem, value chain: whatever you call it, its become a hot topic of discussion, and one that is near and dear to my heart as a supporter of 3D IC commercialization. Well over a year (maybe even going on two) we’ve been hearing suggested changes like a collaborative ecosystem, co-design and development, and a system level approach will help speed up time-to-market and reduce costs.
Internally, this need for a supply chain overhaul has been driven by next-generation technology development; specifically as a way to commercialize 3D ICs more efficiently, and figure out liability issues. But on a higher level, the mobile supply chain market is also driving this need for change.
In a post on RCRWireless, Mattias Lewren, Global Managing Director, Accenture’s Electronics & High-Tech Group, writes that the semiconductor supply chain is completely umprepared to meet the demands of the mobile market for more powerful, innovative semiconductor chips. The reason he gives is that the semiconductor industry is still stuck in the desktop PC world, where the supply chains were stable and predictable, and chip manufacturers could take a leisurely two years to come up with the next great technology innovation. The mobile market is much more fast pace, with a six-month turnaround from development to delivery to keep up with needs of mobile device manufactures.
But this is only part of the story. Jennifer Baljko makes a good point in her EBN blog post, The Mobile Supply Chain: Not Just About the Chip Guyswhen she says that “supply chains are never just about improving one link in the chain. Real supply chain improvements come from addressing the entire end-to-end process from part and board design up to when a consumer buys and recycle a device.” She says the responsibility lies equally with the OEMS, who need to be willing to collaborate closely with their suppliers, sharing critical customer data, and even share some of the financial burden. We have come up with a word for this. It’s called “Interconnectology.”
This theory is similar to the supply chain model for 3D IC products supported by Qualcomm, where the OEMs procure the parts and own the liability, and foundries handle the TSV processes while OSATS do the assembly.
Qualcomm’s motivation for this is a reduced system cost. However it doesn’t necessarily mean a faster time-to-market for developing and delivering new chips, which seems to be the most critical pain point for the mobile market. For that to happen, the OEM has to be involved in more than just procuring the parts. They need to be involved at the R&D level.
Calling OEM demands “impossible” Balijko paints this scenario of OEM expectation: “Go figure out how to design a chip we’ll need for a mobile device we want to launch in six months, and make sure you have it in stock when we need it, but we can’t really tell you a lot about when we’ll need it or how many we might need.”
We’ve learned in the semiconductor space to stop compartmentalizing “ends” (front-end, middle-end, and back-end) and eliminating the “throw it over the wall” mentality. We clearly need to take this strategy to the next level, and engage in better communication between the customers and suppliers themselves beginning at the R&D stage. Only when we truly understand the customer’s needs and expectations, can we be expected to make the impossible possible. ~ F.v.T