Source: Shifting gears: How the cloud drives digital transformation in the automotive industry from Google Cloud
It’s undeniable that technology has improved many facets of modern life. Transportation and mobility, however, continues to be an area where more can be done. For example, look no further than your daily commute. Research by INRIX shows that time spent in traffic has more than doubled in many major cities around the world since 2015.
But five megatrends hold the promise of bringing modern technology’s full benefits to the mobility sector, to improve the commuter experience and more: autonomous driving, shared mobility, deeper customer insights, digital manufacturing, and connected cars. These trends have the potential to fundamentally change the way the automotive industry works, and cloud technology will play a major role in getting there.
Let’s look at each of these trends and see how we’re helping drive them.
Given its impact on both technology and fundamental business models, autonomous driving is perhaps one of the most dramatic revolutions in mobility. Building the infrastructure and providing the tools for companies to build autonomous driving solutions is an area where the cloud can provide great value. At the same time, building an artificial intelligence (AI) system that can be a safe driver is one of the most demanding machine learning (ML) problems to solve at industrial scale.
To validate autonomous driving models, companies need to test cars both on the road and in complex digital simulations. These simulations involve massive computational demands and volumes of data, and are often best served with a combination of GPUs and Cloud TPUs (Tensor Processing Units). We custom-built Cloud TPU for AI workloads, and its high-speed network offers over 100 petaflops of performance in a single pod—essentially making it an on-demand supercomputer. So, no matter what your workload is, we have the world-class AI infrastructure to run it as efficiently and inexpensively as possible.
Almost any vision of improved transportation involves optimizing ridesharing services. Google Maps Platform helps ridesharing and delivery companies improve driver navigation and the overall efficiency of their fleets in a number of ways.
At the driver level, developers can embed a Google Maps-powered turn-by-turn navigation experience into their applications. This means drivers don’t have to switch between apps to get directions or information about their next job. It also lets companies retrieve data about a driver’s journey. With this programmatic control and insight into drivers’ behavior, organizations can better allocate drivers, ultimately decreasing drivers’ idle time and customer wait times. One early user of our ridesharing solution reported a 4% reduction in drive times and up to a 48% increase in the accuracy of estimated times of arrival.
While automotive manufacturers create some of the most iconic consumer products, gaining customer insights can be very difficult. That’s because customer-related data is often fragmented; some resides at dealerships, while some is with the manufacturer itself—potentially distributed over many non-connected systems.
Siloed data can lead to a number of inefficiencies, especially around automotive incentive spends. According to McKinsey, these incentives are among the largest expenses that car companies have, but are the least understood and managed. Additionally, discounts and rebates introduce enormous variation and complexity for dealers and original equipment manufacturers (OEMs) around customer money, dealer money, lease discounts, bonuses for members of the National Funeral Directors Association, and rebates for realtors, as a few examples.
This complexity can result in real pricing confusion. In a recent study, Cox Automotive’s rates and incentives unit compared interest rates, cash, and incentives through dealer service provider tools and found pricing fluctuations of up to $6,750. And one thing consumers don’t like is uncertainty. J.D. Power’s 2018 U.S. Sales Satisfaction Index shows that 14% of customers who shopped but didn’t buy a vehicle at a dealership said they rejected the store because they had difficulty getting a straight answer on price.
With Google Cloud, automotive incentives can be optimized in a way that helps OEMs and dealers control critical costs and eliminate confusion. Using BigQuery, they can ingest, store, and analyze data that can connect the dots between the OEM, dealer, and customer. This process enables incentives to be tracked accurately and consistently. In addition, Google Cloud’s advanced AI/ML tools can investigate patterns in past rebates, so OEMs and dealers can learn whether they’re applying rebates effectively, and even provide automatic triggers to point out potential overspending due to factors like overlapping/duplicative incentives and conflicting rebates.
To pull it all together, Google Cloud APIs let dealers connect to their dealer management systems (DMS) for “one source of truth” for data across their systems-of-record. Finally, Google Cloud has a network of systems integrators and other consulting partners in the automotive industry who can help implement these incentive optimization solutions.
Despite improvements in robotics, AI, and other digital technologies, the automotive manufacturing shop floor remains mired in decades-old systems and siloed data. IDC notes that 77% of manufacturers view digital transformation as an opportunity, making it one of their top priorities. And McKinsey says the new era of automated production and data exchange opens a broad range of use cases that can cut costs, increase yields, and support new manufacturing methods.
Automotive companies, however, often want to run manufacturing workloads on premises. This could be for a number of reasons, such as latency, data residency requirements in some countries, and customer preferences to keep data local. Anthos, our managed, cloud-native platform, addresses these challenges, while still allowing customers to develop and operate their systems as if they were in the cloud. As a software-only stack, Anthos runs on customers’ choice of hardware. Moreover, Anthos workloads can seamlessly move to Google Cloud, or other cloud vendors, at any time.
Quality inspection is another area where our customers are seeing tangible operational improvements with the help of cloud and machine learning technologies. Edge TPU—our ASIC (application-specific integrated circuit) designed to run at the edge with a small power and physical footprint—can let manufacturers run inspections on the shopfloor. And outside the auto industry, LG CNS is already using our technology to detect defects in LCD panel production.
Today’s modern automobile is a supercomputer that generates enormous amounts of data—data that’s captured from as many as 60-100 sensors, and often in real-time. Unfortunately, much of this data is unstructured, in siloed systems, and vulnerable to hacks. At the same time, the electronics and software inside the vehicle are getting exponentially more complex. Between 2010 and 2016, the lines of code required for an average vehicle increased 15-fold, while the complexity of suppliers and processes has also multiplied.
We want to make automotive software simpler and help companies draw insights from the data their fleets generate. We’ve already taken a big step with Android, which has seen significant adoption among automotive OEMs. With the cloud, we can offer a fully integrated approach to telemetry, making it much easier to extract and run analytics on vehicle data.
Of course, whenever data is moving between organizations, trust and security are of utmost importance. That’s why, with our method, all businesses, including OEMs, retain full control and ownership of their data. We also build our own chips to ensure the integrity of our data centers and your data.
Transportation and mobility is a complex field, and its digital transformation has the potential to touch almost every single one of us in some way. Our solutions aim to help automotive companies continue to make that transition, and we look forward to seeing the innovative ways they use cloud technology to reach their goals.