A new and open
chip culture

The crisis in the semiconductor sector threw a wrench into the automotive industry’s plans for production and sales. How could this happen? And how can supplies be secured in the future?


Semiconductor production is a complex and highly sensitive key industry. Shortages can lead to serious problems in many sectors.Getty Images | Image Source | Monty Rakusen

For decades the auto­mo­tive indus­try was known for the sta­bil­i­ty of its sup­ply process­es. Every vehi­cle has thou­sands of com­po­nents, which are made by hun­dreds of sup­pli­ers around the world—and each one of these com­po­nents was right there on the assem­bly line when it was needed.

In 2021, how­ev­er, these oth­er­wise so robust sup­ply chains sud­den­ly began to crack. The prob­lem­at­ic com­po­nents were pre­cise­ly those expect­ed to play a key role in both the cars’ inner work­ings and the mak­ers’ brand identities—namely, semi­con­duc­tors. They con­tin­ue to be in short sup­ply, which means that many cars can­not be pro­duced on time or at all. Accord­ing to the Ger­man Auto­mo­tive Indus­try Asso­ci­a­tion (VDA), around 700,000 fewer vehi­cles will be pro­duced in 2022 by Ger­man car­mak­ers alone. That is a decrease of around 13 per­cent from the year before. A sig­nif­i­cant improve­ment in semi­con­duc­tor sup­plies seems unlike­ly to set in before the sec­ond half of 2023.

Multiple trips around the world

The short­age is due to a num­ber of fac­tors. Covid-relat­ed fac­to­ry clo­sures in Asia have played a role, as has the fire at a major Japan­ese fab­ri­ca­tion plant in spring 2021. In addi­tion, the extra­or­di­nary com­plex­i­ty of microchip pro­duc­tion makes it cor­re­spond­ing­ly vul­ner­a­ble. Dr. Hagen Rad­ows­ki, Senior Part­ner at Porsche Con­sult­ing, has more than two decades of expe­ri­ence in the IT indus­try. “Semi­con­duc­tor pro­duc­tion is one of the most sophis­ti­cat­ed value-cre­ation process­es ever invent­ed,” he says, “with more than 130 indi­vid­ual steps from the ini­tial grains of sand to the fin­ished chips.” In the course of pro­duc­tion, the nascent semi­con­duc­tors trav­el mul­ti­ple times around the world. More­over, the indi­vid­ual pro­duc­tion steps them­selves are very com­pli­cat­ed. For the lat­est chip gen­er­a­tion with a pack­ing den­si­ty of 5–9 nanome­ters, the pro­duc­tion process­es take place on a near­ly atom­ic scale with the asso­ci­at­ed complexity.

What strate­gies and mea­sures are the car­mak­ers using to relieve the cur­rent short­age prompt­ly and then to ensure it will not hap­pen again? First of all, they are try­ing to under­stand the req­ui­site process­es bet­ter and to gain greater con­trol over the sup­ply chains. That has proved very dif­fi­cult thus far, because most of a vehicle’s semi­con­duc­tors are found in com­po­nents from a mul­ti­tude of dif­fer­ent sup­pli­ers. As Rad­ows­ki explains, “Car­mak­ers are seek­ing more direct con­tact with their ‘sup­pli­ers’ sup­pli­ers’ in chip production—and at the high­est lev­els. Chip pro­cure­ment is now a board-level mat­ter and the top pri­or­i­ty on the agen­da just about everywhere.”

Dr. Hagen Radowski, Senior Partner, Porsche Consulting GmbHPorsche Consulting

Building their own chip expertise

Anoth­er goal of equal or greater impor­tance for car­mak­ers is to enhance their own lev­els of chip exper­tise. Does that mean they are chang­ing their approach to the ques­tion of “to make or to buy”? “No,” says Rad­ows­ki. “Car­mak­ers are not aim­ing for in-house pro­duc­tion. Instead, they’re seek­ing to devel­op chips joint­ly with the semi­con­duc­tor pro­duc­ers, which means becom­ing fab­less chip design­ers. That’s what the world’s lead­ing mak­ers of com­put­ers, tablets, and smart­phones do. The actu­al pro­duc­tion is then per­formed by inde­pen­dent ser­vice providers known as foundries.”

Build­ing in-house chip exper­tise is not just a mat­ter of seek­ing to reduce depen­dence on sup­pli­ers. It’s also about devel­op­ing tech­no­log­i­cal know-how as such. “Chips are ulti­mate­ly going to be the dis­tin­guish­ing fea­tures of both the brands and the cars. They’re what will con­nect cars with the cloud, coor­di­nate self-dri­ving capac­i­ties, and enable attrac­tive ser­vices to be added or booked on a tem­po­rary basis. Car­mak­ers don’t want and can’t afford to del­e­gate exper­tise about the prod­ucts that make all this pos­si­ble to sup­pli­ers, or for that mat­ter to sup­pli­ers’ suppliers.”

New capacities in Europe

Car­mak­ers are there­fore address­ing the problem—but does that mean they will man­age to elim­i­nate the risk of semi­con­duc­tor short­ages in the future? Relief in pro­duc­tion is pos­si­ble over the medi­um term at least, and pro­duc­tion sites are in fact mov­ing clos­er in geo­graph­i­cal terms. Bosch has announced sig­nif­i­cant expan­sion of its microchip pro­duc­tion for the auto­mo­tive indus­try, includ­ing at two sites in Germany.

Gov­ern­ments have also taken action. With its Euro­pean Chips Act, the Euro­pean Union (EU) is seek­ing to bol­ster the continent’s semi­con­duc­tor ecosys­tem. The most strik­ing project at the moment is the chip fac­to­ry that Intel will be build­ing in the east­ern Ger­man city of Magde­burg, with an invest­ment of around 17 bil­lion euros. It will take some time, how­ev­er, before large-scale pro­duc­tion can begin. Experts also note that the small num­ber of high­ly spe­cial­ized man­u­fac­tur­ers means that short­ages can still con­tin­ue to plague the sector.

The name of the game: “Smaller but stronger”

The risk of sup­ply short­ages is there­fore still present—especially given the fact that demand from car­mak­ers and their sup­pli­ers will rise sharply over the com­ing years. Fur­ther­more, the com­plex func­tions involved in process­es such as autonomous dri­ving will require even more sophis­ti­cat­ed microchips. “Right now the car industry’s needs are pri­mar­i­ly in the main­stream range of 20 nanome­ters or more per tran­sis­tor,” says Rad­ows­ki. “When autonomous dri­ving becomes stan­dard, more pow­er­ful chips will be needed—which means small­er ones—not only in the cars them­selves but also in infra­struc­ture, and we’ll also want con­sid­er­ably greater quan­ti­ties of them.”

Most car­mak­ers expect their demand for semi­con­duc­tors over the com­ing years to rise at about twice the rate of all other sec­tors com­bined. Chip mak­ers will there­fore have to expand their capac­i­ties enor­mous­ly. Com­pe­ti­tion for the prod­ucts would appear to be inevitable, in part because it takes at least four years to build a new chip fac­to­ry. This is a field in which short-term respons­es to new short­ages are hard­ly possible.

Collaboration instead of competition

We’re approaching a paradigm shift in supplier management: The collaborative model involving carmakers, first- and second-tier suppliers, and semiconductor makers will change in fundamental ways.

 Dr. Hagen Radowski Dr. Hagen Radowski
Senior Partner at Porsche Consulting

How is the auto­mo­tive indus­try address­ing this prob­lem? “We’re approach­ing a par­a­digm shift in sup­pli­er man­age­ment,” says Rad­ows­ki. “The col­lab­o­ra­tive model involv­ing car­mak­ers, first- and sec­ond-tier sup­pli­ers, and semi­con­duc­tor mak­ers will change in fun­da­men­tal ways. Dri­ven by neces­si­ty, car­mak­ers will be play­ing a greater role, also in light of com­pe­ti­tion from other sec­tors.” What will that look like in con­crete terms? One exam­ple: “A num­ber of Euro­pean car­mak­ers and sup­pli­ers have found­ed the Catena‑X infor­ma­tion­al plat­form to ren­der their sup­ply chains more trans­par­ent and to com­mu­ni­cate their medi­um- and long-term needs for spe­cif­ic semi­con­duc­tor prod­ucts. Chip pro­duc­ers can use it as a guide and base their pro­duc­tion plans on the demand.” The new par­a­digm will there­fore also bring high­er lev­els of open­ness and trans­paren­cy among rivals as well as greater col­lab­o­ra­tion. Com­pe­ti­tion alone will no longer be what dri­ves the indus­try forward.

Not over yet: Moore’s law

Since 1965 microchip development has been following Moore’s law, which says that the number of transistors in an integrated circuit doubles roughly every one and a half years. Further such increases are hardly possible because wafer production with a packing density below 10 nanometers is highly complex and nearly on an atomic scale. But there is still room to advance by other means. Chip makers are working with new types of housings and producing “systems on a chip” as well as “chiplets” with supplementary functions such as storage and antennas. They are also using new contacting technologies. Additional substantial increases in performance for the same installation space can be achieved by stacking multiple chip layers. There are a number of reasons, therefore, why Moore’s law will continue to apply, even when transistor units themselves cannot be miniaturized further.
Read the next topicKaraoke Electrifies China’s Drivers