I recently came across this great article: « 日本の「半導体産業」は復活しない…台湾の最先端企業を誘致しても「ムダ」なワケ 技術ではない、経営が問題なのだ ». Or “Japan's Semiconductor Industry Will Not Revive... Why Attracting Taiwan's Most Advanced Companies Is a Waste of Time: It's Not the Technology, It's the Management, by Professor Noguchi or Hitotsubashi University. https://lnkd.in/g7DTNn9
But my thought is that it is beyond just management. The decline of the Japanese semiconductor industry is prime example of the effects of an ill functioning governance system. I am not thinking of corrupt practices but more from the perspective that management, stakeholders, shareholders and government did not have balanced discussion throughout all these years. There were a couple of inflection points that I think determined the downward trend of the Japanese semi industry from a dominating 50% share.
I believe there are many areas ripe for revival but some of the traditional Japanese practices must be tossed aside proactively to generate positive chaos and to have things to realign properly. So, it’s not just management. Japan still has major champions and leaders in the semiconductor eco-systems. Leaders in semiconductor process equipment, test equipment, manufacturing equipment, lead frames, package compounds, wafers. Instead of trying to dictate industrial policies I suggest that politicians and bureaucrat should clear the road for change: incentivize employees to jump ship, encourage specialization, real tax breaks into start-up investments and simple things like banning banning sex based school and company uniforms. Last time the government tried to create a Japan foundry, they were focusing on a manufacturing node which was already there…..
1 Comment
The US congress should be commended for its efforts to ensure that connected and autonomous cars are safe and secure by requiring minimal cyber security standards. We have heard time and time again about cars being hacked and control taken over from the outside. This is a serious threat when you consider these 2 ton machines being converted into weapons. But authorities should also not forget that unsecured home IoT devices are also a serious threat. We have seen cases of home web cameras used to violate privacy but also attempts at physical harm or kidnap children. We also hear of terrorist organizations trying to interfere with infrastructure connected to IoT. There was one instance where water dam measurement and control systems open, even without password protection. Though they may not be a 2 ton weapon they still are dangerous There is little doubt that the semiconductor industry has become a key enabler of the IoT revolution. Integrated chips and reference designs with all the necessary computing, sensing, communicating, encrypting functions necessary to supporting the IoT are flourishing and are enabling new services to come to the market With forecasts of many billions of these devices being connected in the network, the future seems very bright for the industry. But when you take a closer look, IoT is not an industry, rather it is an enabler for business to become smarter and more efficient in capturing value. With so many activities to accelerate this, it has become easier for smaller players with innovative ideas to assemble the technologies without necessarily having to develop the knowhow in-house. The semiconductor industry is largely a volume-based industry requiring heavy capital investment. The potential dilemma is how does this match with serving the increasing long tail, and fragmented customer base? Some initial attempts have started by taking the manufacturing outsourcing EMS model and stretching it to include shipping and end-customer services on the back-end with ideation to prototyping on the front end. Some chip companies have ramped up free web-based digital design tools, and a few have reverted to using AI. Some have even moved more radically and are repositioning themselves as machine learning software companies. Are we seeing the start of radical change in the semiconductor go-to-market model? In the 90’s, just as digital cell phones started surfacing in Europe, I remember being impressed with the DSP technology and the complex algorithms to compress and encode voice into efficiently transmissible code. Also impressive was the modem and radio transmission ICs that encoded the information in such a manner to transmit as far as possible with the lead amount of data loss and power consumption to enable smaller and smaller phones. However, today no one even talks about these anymore, as the function only comprises a small corner of a chip. It is also a given that the chip also contains all possible functions needed in a single chip devices which contains all function of a phone, including the modem and radio transmission. Also thrown in is the software needed to control the communication protocol to enable the phone to communicate with the network, the operating system and most probably tools to easily create smartphone apps. Granted there is always technology advancement because this is what moves humanity forward. The challenge is keeping the value in the value chain. It is also interesting to note that the top players of the mobile phone industry have underdone radical change over the last 10-15 years. There maybe only one name, which still remains in the top 10. Most have dropped out to give way to new players. This is certainly not for the lack of trying to fight it out in the marketplace or intellectual property war chest. For most it was due to an abrupt change in the value chain model. With the advent of electric cars, broadly specialized integrated chips and components are being developed for the automotive market. How disrupt the automotive landscape and value chain in the next 10-15 years? User benefits and business opportunities arising from IoT seem boundless. Yet there are many issues which need digesting. Though most talk about security, a far larger and more present issue is the massive transformation of the semi-conductor industry.
One issue creeping up is the disruption caused by broadener incumbent players offering full reference designs who while serving the industry inadvertently accelerating the arrival of the changes. The semiconductor industry is continually hungry to devour more product functions, and performance requirements as technology nodes improve performance and cost. We have seen cases where companies have been buying out their customers with value addition shifting from the equipment manufacturer to the chip supplier. When Softbank buys ARM, Avnet buys Hacktser, Amazon buys Annapurna, one starts to wonder what is happening and what will be the pecking order of value chain. We have seen the EMS (electronic manufacturing services) companies take over the process from prototype to manufacturing to shipping. Likewise, the process of ideation, design, prototyping is seemingly on the same path by-passing the traditional “design-in” performed by the chip manufacturer . At least that seems to be objective of the alternative players. Interesting changes ahead…..again Cars are still take many human lives but have become an essential part of human life. That technology will make this safer and thereby reduce the number of deadly accidents, through advanced driver assistance systems (ADAS) or through even autonomous driving, is indeed good news. However, one needs to be aware of issues which are still yet to be resolved.
There is much excitement about the electronics and in car-networks in the latest electric cars. It is also important to remember that cars running on internal combustion engines, which constitute most of the cars still sold, also are very dependent on similar electronics and networks. Unlike the electric cars, the latter are still pretty much based on legacy networks. Imagine a small to medium enterprise with a 100 PCs, all networked in one way in a closed system. Then imagine that there are some Apple PCs, some Windows PCs and even some game consoles like Playstation, Xbox. The imagine that not all of them have the latest version of operating system: some have Win95, Windows 7, OS9. Some of these are also connected together using legacy cables and switches. Many of these do not have any updated virus protection since they never were intended be connected to the internet. All of a sudden they need to connect to the internet. This is in a sense what happened to cars when they were “required” to become intelligent. No one thought that people would pay money to get data generated between your brake pedal and the wheel or that someone would attempt to remote control it to take over driving your car. It is key to understand that smart cars without cyber-security are not safe. The trend towards connected cars and autonomous driving is inevitable but better not to become a guinea-pig for the industry. With the advent of the connected vehicle, car cyber security, autonomous driving, my company hosted a conference to engage OEM and tier-1 customers, partners and specialized media to give thought to all the potential technical and engineering challenges, issues with the hope of finding solutions. We managed to land great speakers from industry as well as the government. We were very satisfied to find out that we were sold out. We were however hit with a surprise. The attendee list had a very large number of attendees from the insurance industry, especially given the specialized focus of the event. It is relatively clear that autonomous driving was a threat to business loss if the number of cars were to decrease in the future. However, more interesting was the other concerns that surfaced:
Clearly the potential across industries for disruptions of IoT is larger than what meets the eye initially. Last time I was at CES was back in 2013. I remember a neat laser car headlight demo in the North Hall by a European car manufacturer. I distinctly remember this as this as the rest of the hall was mostly audio entertainment and other specialized aftermarket car products. The impression was that CES was no longer the show case event of years past but that it was still a useful gathering place.
Skip to January 2016. The North Hall was now bulging with the who’s who of the automotive world with car manufactures, tier-1 ECU vendors and selected component vendors who morphed into systems software players. Despite the dominance of the car vendors, it was distinctly different from those car shows in Frankfurt or Detroit as it did not display cars per se but was redefining what kind of services cars could offer customers. Also noteworthy were the entire new halls devoted to smart heath and wearable devices. The main hall was redefining the household appliance by making it smarter: smart TV of course but smart washing machines, smart fridges, smart everything. In the short three years, my take was that CES had morphed from a latest and greatest gadget showcase to a demonstration platform of gadgets that enable services connected through the internet. The electronics is no longer center stage. Welcome to the new reality. Implementing multiple business functions across different geographies, cultures, corporate structures and priorities takes thought and planning. There is not one right answer. Some organizations seem to have corporate cultures whose local organizations are resilient enough to handle things well with a minimal set of core values Others have very detailed and stringent regulations from HQ that requires significant handholding. Yet no matter how much you plan ahead, managing business structures across borders will require constant adjustment as the requirements change and the organization matures or evolves. Setting up common, agreed-upon expectations on roles and responsibilities is probably the most critical start to efficient and effective operations. This is even more crucial when distances are great. Each function (sales, marketing, supply chain, business development) by definition has different prisms of vision on the business and that can create some conflict or disagreement. When handled properly these can enable the organization as a whole to flourish for the better. Here are key points to consider:
|