Today chip giant Intel turns 50 years old! To celebrate this event, chief engineering officer Murthy Renduchintala wrote a blog post that looks at what's ahead. I have a feeling this sort of post would usually be written by the CEO, but Brian Krzanich got kicked out last month. Officially due to "fraternization" with a co-worker, unofficially due to the whole 10nm mess.
Today we celebrate the 50th anniversary of Intel, a company born at the dawn of the technology industry – the advent of the integrated circuit. Since that day, Intel's impact has been felt through a progression of tech waves, including the personal computer, the internet and the cloud. We stand now at the starting line of a new and even more profound digital transformation where virtually every activity will interact with computing.
Computing is about to become infinitely more diverse. It will evolve into new form factors. It will adapt to extreme cost and environmental constraints. And it will power software and algorithms that are always-on, always-learning and able to excel at specialized tasks. Think of systems to prevent fraud on a blockchain or to concurrently anticipate and prevent diabetic events for a million individuals.
Computers will work in concert, invisibly across devices, nearby networks and distant cloud data centers in service of individuals. For example, the car may be both the next big game-changer in computers and the planet's biggest data collector. Intel's Mobileye is already changing cars with driver-assistive technology installed in 27 million vehicles around the world. An additional 2 million cars this year will continuously crowdsource very precise high-definition maps for enhanced sensing and localization as part of Mobileye's evolution to higher levels of autonomy. To help build confidence in these new autonomous cars, we have proposed an open, industry-driven formal model for safe decision-making. Mobileye's example is a new kind of integrated technology platform that speaks to the future of innovation.
In his original 1965 "Moore's Law" paper, Intel co-founder Gordon Moore wrote of integrating many similar devices to great benefit:
"The future of integrated electronics is the future of electronics itself. The advantages of integration will bring about a proliferation of electronics, pushing this science into many new areas."
In our generation, we will have the freedom to integrate devices dissimilar in almost every respect: function, architecture, cost, power and manufacturing process.
As a company, we have been working toward this future for over a decade. Our manufacturing and engineering expertise produces the products and technologies that are the foundation for the world's innovation. It is critical that we continue to drive breakthroughs in five areas:
Innovative Technologies - Mixing and Matching to Create Something New: In a world of heterogeneous integration, where value is created by combining powerful but sometimes disparate technologies, invention and investment in diverse intellectual property (IP) portfolios will be increasingly important. One new pervasive area of technology where we are investing is computer vision – from graphic processing units and vision processing units to domain-specific integrated platforms like Mobileye. Seeing is difficult for computers, yet advanced computer vision and other sensing technologies and software will enable more natural human-computer interactions. At the same time, we continue to invest in the next generation of foundational technologies including central processing units across the spectrum of high-performance and low-power; very fast memory; and 5G communications needed for low-latency access, analysis and transportation of data.
Advanced Architectures and Ecosystems: Artificial Intelligence-related computing tasks will permeate virtually all data-rich processes over the next decade. We are inventing technologies and open software tools that will advance the nascent AI ecosystem, making it possible to gain insight, anticipate needs and continuously learn from data at enterprise scale. This includes new neural network processors, customizable FPGAs in the cloud and embedding AI technologies into existing platforms. Over the past year, Intel has worked with Google on Tensorflow*, Amazon for MXNet*, and Caffe* to enhance deep learning performance with optimizations for Xeon in the data center resulting in over 100 times performance gains for training and nearly 200 times for inference across frameworks.
Packaging in New Dimensions: We are making huge progress to broadly enable heterogeneous integration of computing, memory and communications. We will connect and stack diverse technologies in tiny footprints tuned for specific power envelopes, providing unique cost and performance characteristics with much greater flexibility. One processor we are developing connects "chiplets" built in different manufacturing processes using new 2D and 3D assembly techniques to deliver powerful PC performance with the energy usage of an ultra-efficient mobile device. Testing and verification breakthroughs were also required to begin to scale these new packaging techniques. Imagine the possibilities as we combine into the most efficient of packages the more diverse capabilities – even data center-class technologies – once miles apart in computing terms or based on incompatible processes.
New Models for Computing: Intel Labs is working with academic partners around the world to look over the horizon and rethink computing itself. Quantum computing promises to penetrate complex problems with seemingly infinite variables if it can be scaled reliably. And neuromorphic technology mimicking the function of neurons and operating on feedback from the environment could be a new model of adaptable, always-on, ultra-efficient computing at the edge. We are making strong progress in both.
Securing the Future: Consistent with our Security-First pledge, we are working on deeper collaboration to identify and address vulnerabilities in increasingly complex technologies. And we are engaged in industry and academic initiatives like the RISE program at the University of California, Berkeley on new frameworks and technologies to help protect millions of connected people and things relying on assistive technologies and software.
All of this requires commitment to long-term strategies and sustained investment in people and innovation platforms. At Intel, we have acquired innovative companies but we have also increased our research and development spending. Last year, Intel accounted for over one-third of the world's semiconductor R&D, or $13 billion, to ensure that we are in a position with the right people and the right technology to lead in this future.
On Intel's 50th anniversary we celebrate the past – icons like our founders, Robert Noyce and Gordon Moore – but mostly we look forward, as they did, with our colleagues and our broader industry and academic communities to create the future, improve lives and solve the world's biggest challenges.
Dr. Venkata (Murthy) M. Renduchintala is group president of the Technology, Systems Architecture & Client Group and chief engineering officer at Intel Corporation.