Software, Chips, and Batteries

Let’s take the year 1886 and Carl Benz, who invent­ed the first inter­nal com­bus­tion engine for the car. Or the year 1912, when Henry Ford intro­duced assem­bly-line pro­duc­tion in auto­mo­tive engi­neer­ing. Whichev­er date we start from, more than 100 years have passed since then. Now, for the first time world­wide, new play­ers have entered the stage and dis­card­ed the time-test­ed tra­di­tion­al approach of first design­ing a car, then adding elec­tron­ic com­po­nents, and final­ly devel­op­ing soft­ware intend­ed to unite these very dif­fer­ent com­po­nents under one user inter­face. As a result, the auto­mo­tive indus­try is fac­ing what is prob­a­bly the great­est upheaval in its history.

Today’s mod­ern all-elec­tric vehi­cle is equipped with a mul­ti­tude of elec­tron­ic con­trol units. The func­tions of these devices range from unlock­ing the doors to autonomous dri­ving. And the elec­tron­ic com­po­nents come from a wide vari­ety of sup­pli­ers. It is not only becom­ing increas­ing­ly dif­fi­cult to com­bine all these devices into a sin­gle soft­ware plat­form, but also to update this plat­form at appro­pri­ate inter­vals with rea­son­able amounts of data. It will not be enough to accom­pa­ny this process with a tech­ni­cal evo­lu­tion­ary approach.

In search­ing for a term that best describes what truly needs to hap­pen, a book writ­ten 61 years ago by the emi­nent philoso­pher and his­to­ri­an of sci­ence Thomas Kuhn helps. The title is rather dry: The Struc­ture of Sci­en­tif­ic Rev­o­lu­tions. In it, Kuhn coined the term that so aptly and accu­rate­ly describes how the “old auto­mo­tive indus­try” must now act in order to survive.

The term is “par­a­digm shift.” Accord­ing to Kuh­n’s def­i­n­i­tion, a new par­a­digm can prompt new but rel­e­vant ques­tions about old data. In a par­a­digm shift, you are not only “solv­ing puz­zles” using the pre­vi­ous par­a­digm, but also chang­ing the rules of the game and the “work plan” for arriv­ing at new solu­tions. To explain the basics of the par­a­digm shift, Kuhn had his audi­ence gaze at a duck for 10 seconds …

… and then said: “’This is not a duck, this is a rab­bit side­ways.’ From that moment on, no one saw the duck any­more; every­one saw the rab­bit. Look­ing at the same thing with a fresh new per­spec­tive.” In the con­text of the auto­mo­tive indus­try, this means that it has to approach things differently.

Man­u­fac­tur­ers need to move away from first design­ing the car with per­fect clear­ance, i.e., an even dis­tance between the pan­els on the out­side of the vehi­cle, then adding a seem­ing­ly end­less num­ber of con­trol units, and then try­ing to write the right soft­ware. Instead, soft­ware devel­op­ment must come first. Or in other words, “flaw­less code is the new clear­ance.” This new approach for the auto­mo­tive indus­try has a name: the soft­ware-defined vehi­cle, or SDV.

Why is all this nec­es­sary? Many of the new play­ers have proved that a soft­ware-cen­tric approach makes it eas­i­er to sep­a­rate the soft­ware — which nor­mal­ly devel­ops more quick­ly — from the hard­ware, which advances at a slow­er pace. At the same time, this increas­es flex­i­bil­i­ty in terms of hard­ware replace­ment. The new rules for the SDV par­a­digm, accord­ing to Thomas Kuhn, are there­fore: Explain the prin­ci­ples. Add to this the user expe­ri­ence and the abil­i­ty to use a solu­tion immediately.

A great exam­ple to illus­trate the impact of user expe­ri­ence comes from Tesla, whose cus­tomers fre­quent­ly com­plained that they could­n’t leave their dogs in the car when they went shop­ping because it got too hot in the car; or peo­ple came up to the own­ers and berat­ed them for leav­ing their dog in an over­heat­ed car. So Tesla intro­duced DOG mode. When this mode is acti­vat­ed, the win­dows are rolled down so that the dog can get some air, and the fol­low­ing mes­sage appears: “I’m fine. My owner is out shop­ping and will be back soon.”

But soft­ware always runs on hard­ware. This is where the semi­con­duc­tor indus­try comes into play. Just as for writ­ing good car soft­ware, proven process­es are used for hard­ware and chip archi­tec­ture as well. Experts call this the elec­tri­cal and elec­tron­ic archi­tec­ture, or E/E archi­tec­ture for short. It is the soft­ware-based E/E archi­tec­ture that makes the soft­ware-defined vehi­cle pos­si­ble in the first place.

The most impor­tant pre­req­ui­sites for soft­ware-defined vehi­cles are cre­at­ed in the semi­con­duc­tor indus­try. Sev­er­al thou­sand chips are used in a mod­ern bat­tery-pow­ered vehi­cle. If we com­pare this with the 60 or so high­ly indi­vid­u­al­ized chips in an Apple iPhone, we can see that we are still a long way off. It may be pos­si­ble to reduce the num­ber of con­trol units to around a fifth, but with a high­er pro­por­tion of cus­tomized chips. But why go to all this effort? Because dig­i­tal­iza­tion and elec­tri­fi­ca­tion will con­tin­ue to drive the demand for chips in cars. The auto­mo­tive indus­try is the sec­tor for which the semi­con­duc­tor indus­try needs to grow the most quick­ly in the years to come. And car­mak­ers need more direct part­ner­ships with the semi­con­duc­tor indus­try to man­age this growth.

An apt exam­ple comes from the mobile com­mu­ni­ca­tions indus­try: By work­ing par­tic­u­lar­ly close­ly with the semi­con­duc­tor indus­try and mar­ket­ing direct­ly, phone man­u­fac­tur­ers were able to move from a “stan­dard­ized” approach along the lines of Nokia to what we see today in mod­ern Apple iPhones: greater speed and flex­i­bil­i­ty in the devel­op­ment process, cost sav­ings that should not be under­es­ti­mat­ed, and the abil­i­ty to trans­late avail­able chip tech­nol­o­gy into fea­tures that con­sumers are will­ing to pay for. These are a few of the ben­e­fits that come from a direct part­ner­ship between the semi­con­duc­tor and auto­mo­tive indus­tries. But there are dif­fer­ences as well. Above all, the auto­mo­tive indus­try needs chips that are “car-proof” under very exact­ing con­di­tions. This means cre­at­ing trans­paren­cy and mak­ing strate­gic invest­ments in tech­nolo­gies that the mar­ket demands.

All car man­u­fac­tur­ers will need to put in con­sid­er­able effort to pre­pare for and embrace this par­a­digm shift. For this rea­son, Porsche Con­sult­ing has devel­oped the Strate­gic Semi­con­duc­tor Man­age­ment Frame­work. This model helps com­pa­nies to gain com­pre­hen­sive semi­con­duc­tor know-how and at the same time to visu­al­ize the impact of semi­con­duc­tor pro­cure­ment. This, in turn, helps in devel­op­ing a long-term strat­e­gy for semi­con­duc­tors. More­over, it facil­i­tates the exchange with the semi­con­duc­tor indus­try by agree­ing on a lan­guage that both sides under­stand. This model applies to many dif­fer­ent sec­tors, as a semi­con­duc­tor strat­e­gy is not only urgent­ly need­ed in the auto­mo­tive indus­try. How­ev­er, it is the auto­mo­tive indus­try that is fac­ing the great­est challenges.

The time is now for the auto­mo­tive and semi­con­duc­tor indus­tries to work even more close­ly togeth­er. For the auto­mo­tive indus­try, three fac­tors will deter­mine whether it is com­pet­i­tive: soft­ware, chips, and bat­ter­ies. Car man­u­fac­tur­ers want to lead in these three areas. This does not mean that they want to pro­duce only their own chips. How­ev­er, they must col­lab­o­rate with the semi­con­duc­tor indus­try in a man­ner that leads to a bet­ter end product.