Keynote speech by William Xu at HAS 2019
Good morning, everyone! Thank you for joining us for the 16th Huawei Global Analyst Summit. This event is a time to revisit the changes that we have made over the past year, and an opportunity for us to exchange our thoughts and ideas.
Our products have stayed ahead of competition over the past three decades. This can be attributed to the engineering and technological innovations we have made in response to customer needs. Moving forward, we will pursue innovation based on not only customer needs, but also our shared vision for the future. We will also invest more in research of basic technologies and theories. We will adopt a new innovation strategy that enables us to achieve technological leadership. As we continue to explore the future, we seek to not only light the way forward for Huawei, but also the world at large.
As we all know, the information industry is running into many bottlenecks after 50 years of high-speed development.
First, there are theoretical bottlenecks. Today's innovations are mainly built on theories from decades ago, and are mainly about technological and engineering innovations. For example, Shannon's Limit was proposed 70 years ago in 1948. Now, in this 5G era, we are about to hit the Shannon limit. Another example is CDMA, which was invented by the famous actress Hedy Lamarr in 1941. We are facing a bottleneck in ICT development. We need new theoretical breakthroughs and inventions.
Second, there are engineering bottlenecks. Moore's Law has driven the development of ICT. In the past, CPU performance improved by a factor of 1.5 every year. Now, it's only 1.1. So what's next for Moore's Law? This is the bottleneck we are facing in the development of ICT.
Third, for Huawei, our founder Mr. Ren Zhengfei said in 2017 that Huawei's innovation is now mainly about engineering science, such as engineering mathematics and physical algorithms. He said that the way forward is unclear, and we don't know where we are heading. We are in uncharted territory. How will Huawei address these bottlenecks?
To address the industry's bottlenecks and challenges, we will adopt a new innovation strategy: we are going from Innovation 1.0 to Innovation 2.0. In Innovation 1.0, we have focused on technological and engineering innovations to meet customer needs. In Innovation 2.0, we will focus on theoretical breakthroughs and inventions driven by our shared vision for the future.
I'd like to talk about the core concepts of Innovation 1.0 first. In this stage, we have focused on innovation in technology, engineering, and products and solutions to address customer needs and challenges. It is about going from 1 to N. The key is to help our customers and partners become more competitive, increase their revenue or reduce their costs, and enable them to achieve greater business success. In the past, Huawei made a lot of engineering and technological innovations in wireless, optical networks, and smartphones, and these efforts have generated a great amount of business and social value.
So what's Innovation 2.0? Based on our assumptions about and visions for the intelligent future we are entering, we will aim to break the bottlenecks in theories and basic technologies that have hindered the development of ICT. In this stage, we will focus on theoretical breakthroughs and inventions, which means going from 0 to 1.
First, we have developed the right network development strategies based on our insights into the general direction of the industry. This has ensured that our product R&D has stayed on the right track. What's more, it has enabled our products to stay ahead of the industry and guided industry development. For example:
Huawei pursues innovation in all product domains to provide innovative products and solutions. Our aim is to continuously lead industry innovation.
Take the wireless domain as an example. Huawei developed the first distributed base station in the industry in 2005. In 2007, Huawei launched the world's first SingleRAN base station, which integrated 2G and 3G into one piece of equipment. These innovations have not only helped reduce customers' TCO by 30%, but also significantly lowered the barriers to network rollout. The entire network construction process has become simpler.
Huawei's leadership in the wireless domain comes down to its ongoing technological and engineering innovations.
Huawei pursues innovation in all domains – not just in wireless solutions, but also in the optical domain. From OTN to all-optical networks, Huawei has been leading the development of industry standards. For example, with Huawei's OXC optical switching system, the 3,000 optical fibers traditionally required can be replaced with a single optical backplane. With photoelectric integration, each cabinet is reduced to the size of a small board. All these innovations have created immense value for customers. The results are a 50% reduction in power consumption, a 90% reduction in floor space requirements, and deployment and maintenance efficiency hundreds of times better than before.
Of course, innovation in mobile phones is a key part of Huawei's innovation strategy.
Huawei was the first company to release dual-camera phones. Its latest smartphone, the P30, has four cameras, establishing another new benchmark for phone photography. Without 36 months of persistent efforts by our 300+ R&D employees, it would have been impossible to realize a micrometer folded optical path, 10x hybrid zoom, and optical image stabilization that ensures movement of less than 0.00024 degrees for both main and zoom lenses.
Huawei was also the first company to launch an AI-powered phone. This phone knows what users need, wherever they are. These phones are smarter than their predecessors, even approaching genuine intelligence.
Huawei's most popular phone is its foldable phone Mate X, which features a rotating shaft. This is thanks to innovations in multiple disciplines, such as mathematics, materials, machinery, and design. It took three years to solve this problem. After dozens of iterations, it has achieved stable performance in 200,000 tests, ensuring that every part of the mobile phone lies flat.
What we actually see are products. But the basic technologies behind the scenes are what truly keep our competitive edge sharp. These technologies include mathematics, chip design, materials, and cooling technology.
Huawei has more than 60 labs focusing on basic technologies, more than 700 PhDs in mathematics, and more than 200 PhDs in physics and chemistry. Breakthroughs in mathematical algorithms are the basis for the birth of SingleRAN.
Back in 1991, Huawei designed its first ASIC chip and established its chipset design office.
Over its 30 years of development, Huawei has conducted research in corrosion resistance in materials to make Huawei products adaptable to various environments. In addition, our research into graphene has significantly increased efficiency in battery cooling. Fan-free cooling design has made base stations 30% smaller.
So far, I've been talking about engineering and technological innovations that we have done to meet customer needs. We call it Innovation 1.0. However, businesses today are stagnating because theoretical innovation is not catching up. Without new theoretical breakthroughs, it will be very difficult for us to break technological bottlenecks.
To respond to the demands of the times, Huawei is going from Innovation 1.0 to Innovation 2.0.
So, what is Innovation 2.0 at Huawei?
At the core of Innovation 2.0 are theoretical breakthroughs and inventions driven by our shared vision for the future. Academia is one of the birthplaces of theoretical breakthroughs and inventions. To make it happen though, academia needs to work together with businesses. Businesses are expected to articulate the challenges they face and fund university research to crack these challenges.
Theoretical breakthroughs and inventions involve a great amount of uncertainty, so a closed approach to innovation will not work. The results and capabilities must be shared. Huawei's Innovation 2.0 is about open innovation and inclusive development. We will work together with universities, research institutes, and businesses alike to drive breakthroughs.
How can we make Innovation 2.0 a reality?
We will make it happen through initiatives like funding university research, building labs, and making technological investments along multiple paths. We need to consider industry challenges and academic insight, and also confidence of venture capitalists to conduct joint innovations.
We established the Institute of Strategic Research to help make Innovation 2.0 a reality.
The Institute of Strategic Research mainly engages in the research of cutting-edge technologies for the next five or more years. Each year, we invest 300 million US dollars to fund academic research in basic science and technologies.
The most important mission of the Institute is to look into the future and carve out a clear path for Huawei's technological development over the next five or more years. It will help ensure that Huawei does not lose direction or miss opportunities as it moves forward. It will also explore technological and business models surrounding Huawei's core business to ensure lasting competitiveness.
We will invest heavily in the following areas:
The Institute of Strategic Research is poised to explore and research future-proof technologies throughout the entire information lifecycle, from information generation and storage all the way to computing, transmission, presentation, and consumption.
For example, in the field of information presentation, we have light-field displays. In computing, we have brain-like computing, DNA data storage, and optical computing. In transmission, we have visible light. In basic materials and process manufacturing, we have metamaterial, atomic-level manufacturing, and so on.
Next, I'd like to introduce several basic research and invention areas that Huawei is engaging in to give you an idea of what future-proof technologies look like.
The first area is optical computing. As you may be aware, massive data of different types are being generated every day. Due to the limits of Moore's Law, it is very costly to process all data with only one type of computing architecture. Heterogeneous computing is an effective way to address this challenge, pushing the limits of Moore's Law.
Huawei researches optical computing and has taken advantage of the analog nature of optics to enable complex logical operations for data processing.
For example, in the field of artificial intelligence, 80% of computing is a matter of matrix transformation and optimal solutions. If it is done using the traditional CPU approach, it can be very inefficient. With optical computing, the performance will improve by a factor of hundreds due to optical features such as diffraction, diffusion, and interference. These mathematical features give optical computing a natural advantage, eliminating the need to undergo the complex process of analog-to-digital conversion.
Just imagine: As AI computing becomes increasingly prevalent, new computing architectures will be more suitable for 80% of computations. This will deliver performance hundreds of times faster than before. The limits of Moore's Law will be broken to a large extent.
The second area is DNA data storage. In the information world, the volume of data is growing exponentially and piling up, much faster than the growth of Moore's Law. There is an increasing demand for storage capacity, leading to higher cost and limiting the continued growth of capacity. In fact, storage has become the most costly element in the IT industry.
To address this, you either keep discarding some data, or search for new technologies that offer more capacity.
We all know that genes contain huge amounts of information. A human gene stores dozens of gigabytes of data. Now that DNA is so good at storing information, can we possibly use NDA to store data? A cubic millimeter of DNA can store 700 TB of data, equivalent to 70 hard disks of 10 TB each. According to this calculation, a kilo of DNA is sufficient to store all of today's data. The capacity is staggering.
However, this DNA data storage technology is far from commercial use. This is because the data read and write speeds are still very slow. It takes 4 days to write 5 megabytes of data. So we need to explore new methods and technologies that can help us break these bottlenecks.
The third area is atomic-scale manufacturing. Today, the precision of manufacturing has reached the nano level – say, 10 nanometers. However, the advancement to the "micro level" still happens at the level of macro manufacturing. For the more sophisticated precision manufacturing of today, the macro approach will no longer work properly, hitting the Moore's Law ceiling.
Let's look at this problem from another angle. Can we manufacture directly at the atomic level? There is a manufacturing concept called "atoms to products". In this model, atomic-scale items are directly assembled into nanostructures and then integrated into larger micro components. This is how an atoms-to-products model works. The size of an atom is 0.1 nanometer, which means atom-scale manufacturing will lead to a 100-fold capacity improvement over Moore's Law.
There is still a long way to go in our journey toward an intelligent world. We need new inventions and breakthroughs in theory. We will continue to explore the unknowns that the future holds for us and boldly go where no one has gone before.
In uncharted territory, we have to make our own trail.