HP’s 2019 Megatrends reveal how technology is shaping a new era

In my role as HP’s CTO, I spend a lot of time pondering questions such as: What will future customers need from their HP products? How do we start the work now to meet those demands?

But there are also bigger, more far-reaching questions that influence how everyone at HP thinks about the work we do and the impact we can have as a company. What will society look like in 30 years? What will be the respective roles of automation and skilled labor? How will Asia’s rise impact our business model? How do we create a more sustainable future?

Each of these questions offers an opening into a heady topic with no simple answer. But we must consider them, because they will directly impact how we move forward as a company and a society.

That’s why we’ve formalized our analysis and forecasting process into a body of work we call Megatrends, a systematic effort to identify the global technological, economic and social currents that are influencing how people will live and work around the world. Since 2016, our annual Megatrends report has explored the hidden patterns that will move markets and people.

This body of work has led to a striking realization over the years:  We are now witnessing the early stages of a technological, socioeconomic and demographic revolution that will likely lead to a new era for humanity. The signal has only grown stronger with this year’s report. People, machines, data and energy are now so deeply connected to each other, that for many but not all in the current world, creating a better future will require finding the balance between them all.

Urbanization unlocks prosperity and new challenges

Let’s start with the people that are at the center of this picture. We pored over volumes of socioeconomic data and consulted with leading economists to uncover the trends that our society is most likely to encounter in coming years. Today, just over half of the global population resides in cities, and the UN predicts that number will grow to 68 percent by 2050. The future will be filled with large and mega cities that will each be home to at least 5 million inhabitants. This year, we explored what the economic situation will be like for those urban citizens and those in smaller urban areas, in all places where people will be accruing more disposable income over the next three decades. The answer tells us where tomorrow’s markets and talent will be.

The data points to more people moving out of poverty and into the middle class — particularly in Asia, North America and Europe. Our analysis finds that Asian workers, on average, will have seen their average household disposable income more than triple from 2001 to 2035 in constant 2015 dollars. In cities of all sizes around the world, we expect the average household income to grow by an average of 2 percent annually through 2030. That’s a big boost in spending power.

What’s driving this urbanization and income rise? Growing economies creating more jobs, especially a constantly growing demand for skilled labor in urban settings.

Yet we already know the high-tech workforce needed to sustain economic growth won’t be sufficient to meet the demand for labor. Research shows there will be a global shortage of around 85 million high-skilled workers by 2030. As an extreme example, Singapore is expected to register a stunning labor shortage of 1.1 million people out of a workforce totaling 4 million by 2030.

And because technology is changing so rapidly, those who are already in the skilled workforce need to embrace lifelong education and constantly develop new skills to keep pace. Up to half of IT-sector workers need retraining by 2022 to remain effective. Ironically, as technology creates a skills gap, it can also help close that gap, creating accessible, scalable upskilling opportunities via gamified educational platforms, VR-based training and more.

But bolstering the number of high-skilled workers isn’t going to fix the problem alone. Companies are aggressively investing in automation to deal with rising demands and our growing labor shortages. It’s possible we’ll see the number of tasks that machines perform increasing by 50 percent in coming years.

Automation is a critical component to boosting global economic productivity, shifting labor from routine tasks to higher-skilled and higher-paying jobs, and opening up the middle class to millions of people. This is where the machines and data come in.

Smart machines that can collect data, learn from it and respond are the heart of automation. Every day, more smart machines are coming online, ranging from power generators and manufacturing robots to smart speakers and autonomous cars, and they are starting to communicate with each other at the internet’s edges. The lifeblood of this world of sensors and actuators is data, which is central to the evolving rate of AI and machine learning, the move from cloud to edge computing, and the role that software and digital twins will play. 

Right now, it takes too long to transmit all the data being generated at the edge to distant cloud servers, process it and then return other data in the form of decisions — and there already isn’t enough bandwidth to do it. That’s why engineers are developing powerful, tiny computer processors called Machine Learning Accelerators (MLAs) that can crunch data on the spot much more efficiently, and act on it immediately. This enables services that have resided in the cloud to move down into devices at the edge, what's known as edge computing.

At the same time, the network itself is getting a major upgrade to boost bandwidth and reduce latency. Fifth-generation, or 5G, mobile cell networks will start switching on this year, and while their rollout may take years, the era of ubiquitous wireless broadband will begin. Data transmission and reception across the network will be nearly instantaneous. Edge computing and 5G will dramatically expand opportunities to adopt automation and accelerate the use of edge-driven automation, especially for critical sensing and actuation events where timeliness is critical.

As just one example from our business, HP’s giant industrial presses that can rapidly print huge runs of packaging need systems that can monitor for defects. Press-mounted cameras built with powerful machine-learning software could detect print-run issues in real time and give the machine the information it needs to automatically correct itself. But for such capabilities to become reality, we need to build powerful chips that can handle computer vision analysis locally. That’s where edge-computing innovation comes in.

The impending commercialization of autonomous vehicles also illustrates the critical importance of edge computing. Picture this: While riding along in one of these robotic cars, a pedestrian holding an umbrella crosses into your path. You’ll need the vehicle’s sensors to instantly transmit data to stop the car, not send it to a distant server that’s running imagery analysis software. Such processing needs to be embedded within the car itself, so it can make decisions without fail and nearly instantly.

Other applications, like autonomous vehicle networking, which lets individual machines communicate with each other and with transportation infrastructure to adjust speed, caravan and efficiently use parking locations, will only become possible with the more robust, flexible and higher bandwidth 5G rollout.

Manufacturing, and work itself, gets a digital upgrade

As our machines are asked to gather more data, make inferences from it and need to complete more complex tasks, they’re going to need more advanced operating instructions. To meet demands like those of smart industrial presses and autonomous vehicles, researchers are advancing machine learning so computers can make decisions, not from hard-coded firmware, but from models within the machines that "learn" from data coming from a machine’s sensors and make inferences that drive their decisions on how to operate.

This development, called software 2.0, is seeing software-based models replace the complex code approach that software engineers have been using for years to write firmware. Software 2.0, in combination with these machine learning models, can produce more robust and flexible solutions to help machines process immense amounts of data and make decisions in more nuanced ways. Along with edge computing, this approach would be much more effective at, for example, helping an autonomous car identify a crossing pedestrian whose head and torso are obstructed by an open umbrella. Before encountering this situation, the software is trained on huge numbers of possible iterations of what a pedestrian looks like. This creates a superfast decision-making loop.

When machines can operate and learn in response to the data they sense and capture, it becomes possible to create virtual models of the machines which enable us to predict, simulate and optimzie their operations in environments before they are there. This concept, called the digital twin, enables virtual machines that precisely mimic their real-life counterparts could dramatically alter the nature of work through virtual prototyping and modeling. Software can run huge cycles of tests on these simulations to understand how an actual object would operate in the real world and make improvements before it’s produced.

Companies and other organizations are already starting to deploy digital twin technologies, and the nascent innovation will only become more useful as more data becomes available and models become better at simulating the real world. Manufacturers are testing and improving new microchip designs using virtual models to maximize heat handling. Turbine makers simulate operating machinery to predict when they need maintenance before they break down. Cities from Cambridge to Singapore are ingesting real-world data into simulations to help deliver services to citizens.

Sustainable energy use becomes an imperative

Higher consumption from a growing middle class and the rise of automation will translate to greater energy demand. Tomorrow’s societies and the machines that make them possible will require massive amounts of power to survive. This requires more advanced technology and putting energy efficiency in compute and our products, as well as sustainability, at the center of our planning.

Advances like additive manufacturing are doing just that. By building objects through adding material — instead of the traditional subtractive manufacturing process that cuts objects from blocks of material — manufacturers and consumers, like airlines, can potentially save huge amounts of raw materials and energy. This method also unlocks new component designs that weren’t buildable with subtractive methods. We reviewed one study that concluded additive manufacturing in the aerospace sector could reduce an Airbus A320’s weight from more than 93,000 pounds to just over 73,000 pounds. Since one pound of aircraft weight reduction translates to 114 pounds of annual fuel consumption savings, additive manufacturing has the potential to deliver huge benefits to companies, people and the planet.

Innovations like edge computing also fit the bill. It takes a great deal of electricity to transmit huge amounts of data across networks. Keeping computational processing local erases that demand. That’s why many start-ups and other companies are developing more capable chips, such as MLAs, that can handle data at the edge and operate at much lower power than the remote server processes they replace in the cloud, or the energy-hungry Graphical Processing Units (GPUs) used today for similar tasks in client computing.

The 5G rollout, meanwhile, will unlock other opportunities for systemic energy conservation. Improved wireless telecommunications will help the Internet of Things achieve what its developers have always envisioned — smart machines that make up autonomous vehicle fleets, smart city infrastructure and streamlined industrial processes featuring predictive maintenance will allow massive energy efficiency upgrades. Think of a future where robotic cars talk to each other and infrastructure over no-latency wireless networks to eliminate the need for stoplights and maximize parking in crowded cities.

Companies that invest in digital twins to virtualize equipment prototyping and deployment will decrease energy and raw material use while driving up efficiency in the real machines that come from the process. Just imagine a jet engine prototype’s digital twin that is tested and improved for operating efficiency, heat management and design that limits vibration before any manufacturing begins. This speeds up production timelines and decreases fuel use and maintenance downtimes once the engine is installed on aircraft. This advance is more sustainable by design. And sustainability is key to creating a future where our technology helps make a better world.

For HP, understanding these sweeping demographic, socioeconomic and technology trends — and how they relate to one another — is incredibly important. People, machines, data and energy are all deeply intertwined in our modern world, and profound changes in each will impact all the others. The interplay between them will determine the direction of HP, the global economy and civilization. That’s why every one of us needs to pay attention.

Read how Singapore is transforming itself into a living laboratory for smart technology.