Over the last few years Intel has had numerous delays of its 10nm manufacturing process and now the same problem is effecting the bringing up of its 7nm process. A year's delay can often be reason enough to drop a node altogether but the leading-edge is now so expensive that it is almost impossible to recover from a dropped node. It effectively spells the end of participation at the leading-edge. The most obvious reference here is Globalfoundries (see GloFo rethinks its future, drops 7nm FinFET).
While Intel's revenues and profits right now are being boosted by the Covid-19 pandemic, the recessionary 'morning-after' is yet to come and Intel's problems with manufacturing will at some point start to hit its bottom line.
Intel has partly adopted the fab-lite business model by default as it acquired fabless chip companies that were working with TSMC. But CEO Bob Swan is drawing on that playbook more and more heavily. And that can quickly become a self-fulfilling prophecy.
This is because margins are reduced on foundry-produced chips giving the fabless or fab-lite chip vendor less funds to spend on R&D and capex. If Intel starts to lean heavily on foundries for 7nm manufacturing so that it can get competitive products out the door in a timely manner, it will be even less likely to be able to afford to develop the technology for future nodes. In addition, it will have less experience of manufacturing at 7nm putting it at a disadvantage against the competition at a future node. That is unless it can leap on a discontinuity in the miniaturization roadmap and suddenly show an advantage over TSMC.
But CEO Bob Swan said Intel is "pragmatic" about manufacturing: "We will continue to invest in our future process technology roadmap, but we will be pragmatic and objective in deploying the process technology that delivers the most predictability and performance for our customers, whether that be in our process, external foundry process or a combination of both," he said in the 2Q20 financial conference call.
Next: Going lighter