Arm, Ltd. is synonymous with IP. The company has done more to shape the semiconductor industry and to enable the growth of modern electronic gadgets than any other IP company. In this section, Arm tells its story.
It was on the 26th of April 1985 (at 3 p.m. to be precise) that the first Arm silicon sprang into life—it was a 25K transistor design implemented in 3 µm technology with just two layers of metal.
However back then the “A” in Arm stood for Acorn—Arm, the company, had yet to be formed. Acorn sold computers to schools, and so cost was a prime concern. This meant that when it came to replace the aging 8-bit 6502 in the BBC Micro with a more powerful microprocessor it had to be cheap.
Unfortunately, the commercially available alternatives at the time were simply not cheap enough, nor did they give sufficient performance improvement. So Hermann Hauser, the Managing Director of Acorn, decided that Acorn should build its own 32-bit microprocessor.
However, he gave the Arm design team two distinct advantages over other microprocessor design teams—no money and no people! So, the design had to be simple and straight forward; indeed the first Arm reference model was written in just 808 lines of Basic.
Interestingly, although the Arm silicon worked the first time, it appeared to be consuming no power at all, at least, that is what the ammeter said. It turned out that the test board had a fault, which meant the Arm core was effectively unpowered and was running solely on leakage from the I/Os. This low power consumption was a valuable side effect of making the Arm core cheap and turned out to be the key to its success in the emerging mobile electronics market.
1990: Arm Ltd. Founded
In early 1990, Apple was developing a “Personal Digital Assistant” called Newton and was looking for a low power processor to power it. Apple was very interested in the Arm RISC core but was reluctant to base a product on Acorn’s IP. The result was the foundation of Arm Ltd. on the 27th of November 1990 as a joint venture between Apple, Acorn, and VLSI Technology.
The first Arm office was established in a beautiful 17th century converted barn just outside Cambridge, UK. Apple invested £1.5 million, Acorn put in the 12 engineers who had worked on Arm and VLSI provided the design tools. VLSI also became the first licensee, manufacturing the devices for the end customer, Apple.
Arm set about extending the architecture to meet Apple’s requirements for bit addressing and endianness support. In January 1992, the ARM610 was complete and the Apple Newton launched in 1993.
Unfortunately, the Newton was not a great success. In hindsight, many think it was ahead of its time. Yet, Robin Saxby, Arm’s CEO, knew Arm had a great product and to take advantage he steered the business in a new direction, one that could quickly scale the spread of the new Arm technology globally. That led to Arm’s IP licensing business model, a rather unusual move at that time. The semiconductor industry was in its pre-Moore’s Law phase, where most microprocessors were designed and built as discrete chips. This was mainly due to their size as they were not yet small enough to be formed into an SoC.
Later in 1992, UK-based GEC Plessey Semiconductors and Sharp (in Japan) became the first two official licensees. The following year they were joined by Cirrus Logic and Texas Instruments, the first US licensees.
The Arm processor model meant it was licensed to semiconductor companies for an upfront license fee and then royalties were received on production silicon. This effectively incentivized Arm to help its partners get to high volume shipments as quickly as possible.
One interesting feature of the Arm IP licensing business model is that the pipeline is very long—it can take years from the time a license is signed until the royalties really start to kick in.
When Arm started, it had an internal software group producing compilers, assemblers and debuggers. But it was still a small company with a niche processor architecture, and so companies such as Wind River that produced real-time operating systems needed to be paid to port their product lines and support the architecture.
As the Arm architecture became more and more widely licensed, Arm put a lot of effort into building a partner program so that anything that an Arm licensee might need would be available from an ecosystem of third party suppliers. As Arm’s licensee base grew, the economics of supporting the Arm architecture changed and selling into Arm’s base of licensees became a huge opportunity, so nobody then needed to be incentivized to support the architecture.
1994: “Thumb”—the Big Break
In 1993, Nokia approached Texas Instruments (TI) to produce a chipset for an upcoming GSM mobile phone. Arm proposed an Arm7-processor-based system to meet Nokia’s performance and power requirements. However, Nokia rejected the plan as the memory footprint of an Arm7-based solution made the system cost too high because dealing with a 32-bit processor meant each instruction took 4 bytes. To counter that, Arm came up with a radical idea to create a subset of the Arm instruction set that required just 16 bits per instruction. This improved the code density by about 35% and brought the memory footprint down to a size comparable with 16-bit microcontrollers.
Thumb, as it became known, was a major breakthrough that won Nokia over, and it is arguably the innovation that propelled Arm into its subsequent dominance of the mobile phone market. The first Arm-powered GSM phone was the hugely popular Nokia 6110. The Arm7TDMI that powered it went on to become one of Arm’s most successful products with more than 170 licensees who have shipped more than 10 billion units since its introduction in 1994.
Arm’s timing turned out to be very fortunate. The Arm7TDMI was released just as the cellphone market started its explosive growth. Arm became the standard processor in mobile, as it still is today. Not only did this mean that a lot of cores were shipped, it meant that every semiconductor company needed an Arm license if they were to sell semiconductors successfully into the cellphone market.
One licensee of the Arm architecture was Digital Equipment Corporation (DEC). But rather than licensing a particular core, they instead bought an architectural license and built their own core on their own process, highly optimized for even lower power and higher performance. It led to the amusingly-named StrongARM which debuted in 1995. There was an interesting twist here too as many of the team members who developed StrongARM moved on when that part of DEC was acquired by Intel (which developed a more powerful successor to StrongARM called Xscale, before eventually selling its entire communication business to Marvell). The team that left DEC would go on to create PA Semiconductor, which designed very low power PowerPC cores. In 2008, Apple acquired PA Semiconductor and took out an architectural license from Arm. Today that team still forms the core of Apple’s processor design team working on…Arm cores. In 2013 they produced the first 64-bit Arm inside the Apple A7 that powered the iPhone 5 and the iPad Air.
By the end of 1997, Arm had grown to become a £27 million business with a net income of £3 million. To continue its growth, it was decided to float the company and on April 17th, 1998, Arm completed a joint listing on the London Stock Exchange and NASDAQ with an IPO stock valuation of £5.75. Luckily for those early investors, Arm’s stock soared, and the company became a billion-dollar success story almost overnight.
The Move to Synthesizable Cores
Chips were now small enough so that a microprocessor only occupied a part of a chip and it was possible to build software-based systems on a single chip, the so-called SoC. The microprocessor was one of the first elements of anu SoC to be sold using an IP business model as most design teams didn’t have the knowledge or desire to build their own microprocessor. They also mainly lacked the skills to build the toolchain of compilers and debuggers necessary to make it usable. As a result, Arm was designed into more and more SoCs, especially in the rapidly-growing cellphone market where Arm was quickly becoming the de facto standard architecture.
However, the Arm core was technology-specific “hard IP” and it became clear that porting it to so many different technologies was causing a bottleneck, and something had to change. A synthesizable core was required that could be licensed to anyone without needing a technology-specific port of the core.
In 2001, the ARM926EJ-S was announced. It was fully synthesizable with a five-stage pipeline and a proper MMU, as well as hardware support for Java acceleration and some DSP extension. It went on to be licensed by over 100 silicon vendors worldwide and has shipped over 5 billion units to date.
2001 was also the year that Robin Saxby, the original CEO of Arm when it was spun out of Acorn, passed the torch to Warren East who became the new CEO.
Artisan Components was a company that designed and marketed standard cell libraries, memory compilers, and interface components. These are the basic components of any synthesizable design—the Lego bricks out of which complex designs are built. In 2004 Arm acquired Artisan and so added a physical IP business line.
In recent years, the physical IP business has also been extended to add special cells called Performance Optimization Packs (POPs) that further optimize the process of synthesizing Arm cores for particular processes, most notably for the big foundries that actually manufacture many of the Arm-based designs.
The subsequent development of Arm9 and Arm11 families had extended the capability of the Arm architecture in the direction of higher performance with the introduction of multi-processing, SIMD multimedia instructions, DSP capability, Java acceleration, etc. However, there were other potentially larger market segments, which these processors did not address. So, in 2005, Arm introduced a change of direction and the Arm architecture was split into three “profiles,” the upwards and to the right path continued with the Cortex-A, a new range of high-performance real-time processors was introduced as Cortex-R while the Cortex-M profile targeted microcontrollers.
By 2008, the smartphone market was booming and the demand for increased performance, while at the same time maintaining a long battery life, presented quite a challenge. Arm responded with the Cortex-A9 MPCore, a multi-core processor which was better able to address the huge dynamic range in processing power from idle or playing music to full bore 3D gaming. This was further improved with the introduction of the heterogeneous “big.LITTLE” architectural extension in 2011. This highly innovative design innovation enabled switching between a high-performance core and a lower performance core as compute demands shifted depending on what the chip was required to do.
In a smartphone or a tablet, there are two main processors: the application processor, which was already dominated by Arm, and the graphics processor, a specialized core that drives high-resolution screens and is required to run videos and games on such devices. In 2008, Arm introduced its Mali graphics processing unit (GPU). Like previous Arm processor cores, Mali would go on to become the world’s most widely licensed GPU architecture.
In 2011, Arm announced the Armv8 architecture, which took the architecture up to 64-bit without losing backward compatibility with all the existing 32-bit software. This was targeted at expanding Arm’s footprint into the data center market. Arm had natural advantages over Intel in the data center as a significant part of the cost of a data center is from the electricity needed to power all the computers and cool them. Arm’s low power design is very attractive in comparison to Intel, the current market leader, as Arm cores can deliver high performance but at far lower power, silicon size and therefore cost. You can see how far this has now moved by what Arm announced in 2018 with their Neoverse range of technology.
In July 2013, Warren East retired as CEO of Arm, and Simon Segars, his deputy, took over. Segars, an engineer by training, had been hired in Cambridge as employee #16. He immediately began the process of positioning Arm to succeed in the emerging Internet of Things (IoT), where billions of tiny, highly efficient processors and IP blocks would be required. That expansion included the acquisition of companies such as Sensinode, an IoT startup from Finland that led the creation of the 6LoWPAN and CoAP standards for low-cost low power devices. Sensinode had also been a key contributor to the IETF, ZigBee IP, ETSI, and OMA standardization efforts. Other Arm acquisitions targeted security (Sansa Security and Offspark (IoT security software) and connectivity (Wicentric).
At the same time, Arm was bolstering its position in areas such as tools (PolarSSL, Carbon Design Systems, Allinea Software).
Product lines continued to develop with a spread from the Cortex-M0 microcontroller up to 64-bit multi-core processors aimed at the data center and communications networks, and with cores in between targeted at attractive new markets such as low-end low-price smartphones.
ARM had become the standard microprocessor for mobile computing, especially for smartphones such as the iPhone or Samsung Galaxy, and tablet computers, including the iPad. Its architecture also powers Qualcomm’s Snapdragon, Apple’s series of Ax application processors, Mediatek’s chipsets, and most high-volume low-cost feature phones. By the middle of the second decade of the 21st century, after 25 years of constant innovation, the company is also very well poised to deliver solutions spreading well beyond the mobile market, and the IoT into servers, the most advanced vehicles and even laptops (another Intel stronghold).
But then in the summer of 2016, everything changed for Arm. They had been considered almost acquisition-proof because of their independent positioning in the semiconductor sector. Very few companies could even consider buying Arm and without risking a partner revolt that might destroy the company’s value. One company that could do this though was Japan’s SoftBank, led by the highly enigmatic Masayoshi Son.
Masa, as he’s often called, rang Simon Segars to say he wanted to meet him and Arm chairman, Stuart Chambers. Segars had first met Mr. Son a decade earlier and thought he had a technology partnership in mind.
One thing he later said was that he thought whatever Masa was thinking it would be worth listening, such was Son’s reputation. The only problem was Chambers was vacationing on a yacht off the Turkish coast. So, Son sent his private jet to collect Segars and take him to Turkey, where Segars, Chambers, and Masa lunched at a restaurant near the town of Marmaris. They were the only guests, as Son had arranged to clear the place just for them. To the two Arm guests, the real reason for the meeting soon became clear: Son was proposing that SoftBank acquire Arm.
Not long thereafter, SoftBank bought Arm for $32 billion, a figure that was 40 percent more than its market value at the time. Son was convinced that Arm was the company with the only foundational technology capable of delivering his IoT vision for a trillion smart devices deployed by 2035.
Once more a private company within the SoftBank universe, Son had given Arm the ability to invest more and faster, and to realize its founders’ vision of Arm technology spreading into all markets and all geographies.
In June 2018, Arm acquired Stream Technologies for its connectivity-management solutions for IoT. Two months later, Arm acquired enterprise data management company Treasure Data, which positioned Arm’s IoT Services Group as a leader in data, connectivity and device management.
As 2019 dawned, Arm had more than 6,000 employees, a far cry from the 12 employees it had started with as it was spun out of Acorn. The company is acknowledged as the leader in IoT chip technology and is making inroads fast with rolling out its IoT software platform, Pelion. Alongside maintaining its hold on mobile computing, Arm is also pushing hard and fast into the infrastructure market as the traditional cloud evolves and 5G begins. It is also in the vanguard of true edge intelligence, driving advanced computing, such as artificial intelligence, into many more devices. Perhaps even more interesting, it is looking to solve the many complex challenges presented by the needs of fully autonomous vehicles and what may replace silicon in the heart of a compute chip.
Arm’s global ecosystem of partners is made up of more than 1,000 companies. These partners add value to the Arm architecture and make it extremely difficult for others to compete with Arm’s IP business at scale.
When Arm was founded in 1990 it had just one licensee, VLSI Technology, which had shipped a total of 130,000 cores. Today, Arm has more than 500 licensees who have collectively shipped more than 130 billion cores with that number increasing at a rate of 20 billion cores each year.