Executive Cases: Interviews with Senior Executives of Early-Stage Companies:
Arteris – France
Prepared by George Foster
In 2003, three French engineers founded Arteris to develop the first commercial network-on-chip (NoC) technology for use on chips by the semiconductor industry. NoCs are used to link the various IP cores on a system on a chip (SoC) in a somewhat similar way as the Internet connects computers. The founders, Philippe Boucard, Alain Fanet and César Douady, first worked together at Matra and then founded T.Sqware, a developer of network “edge” processors. When T.Sqware was closed in 2002, the three decided to apply network processor technology to packetized on-chip communication, founding Arteris in 2003. In 2006, Arteris shipped its first product, NoCSolution, which was licensed by Texas Instruments for its OMAP processor line in 2007. In 2009, Arteris released its second-generation product called FlexNoC, which made it easier for makers of semiconductors to adopt NoC technology. Starting in 2010, Samsung and Qualcomm licensed FlexNoC for use in the majority of their mobile and wireless chips. Arteris became profitable in 2011 and, by 2012, its products were being used in the majority of the world’s mobile and wireless SoC designs.
Wayne Cantwell is a partner at Crescendo Ventures, a key venture capital investor in Arteris. Prior to joining Crescendo in 2003, he held operating roles in software and semiconductor companies. He previously served as president and CEO of Soisic SA (a French start-up in the semiconductor intellectual property licensing business that was sold to ARM Holdings) and president of inSilicon Corporation. Cantwell has been involved in over 20 early-stage companies in Board of Directors and Advisory Board roles over the past 25 years.
Q1: What was the source of the initial idea, and how did that idea evolve into a viable growing company? How did it change over time?
Janac: “Arteris was formed to pioneer the idea of using packet transport networking techniques for moving information inside SoC-type semiconductors. This idea was based on technologies delivered in network processor chips designed by a predecessor company called T.Sqware. T.Sqware was acquired by GlobeSpan/Virata, which in turn was acquired by Conexant. Conexant shut down T.Sqware as part of the communication/internet meltdown of 2001. The founders of Arteris then sought to apply network processor technology to on-chip communications within semiconductor chips. While the company considered using this technology to build new types of chips or new types of field-programmable gate arrays, Arteris decided to pursue what was at the time a newer business model by becoming an IP licensing company to the world’s semiconductor manufacturers. While NoC technology has broad applicability, Arteris chose to focus on SoCs for the booming smartphone market in order to focus our engineering, sales and marketing efforts on major global mobility SoC-makers such as Qualcomm, Samsung and Texas Instruments.”
Q2: What were the major growth accelerators for your company in the early years of high growth?
Janac: “The primary accelerator was the increase in SoC complexity, making internal interconnect R&D more expensive. The industry was putting the functionality of what used to be on two to three chips into one large chip, and customers were in pain. Successful use of the Arteris NoCSolution interconnect IP by Texas Instruments in their OMAP4 application processors gave the industry the confidence that our technology worked and that our company could deliver. Growth accelerated in 2009 with the delivery of Arteris’ second-generation NoC interconnect IP product with improved latency and ease of use. Finally, competition that failed to adapt to the technology changes in the semiconductor industry provided a vacuum for Arteris to fill.”
Q3: What role did key aspects of the entrepreneurial ecosystem surrounding your company play in the growth of your company?
Janac: “The market for on-chip networking technology was small in France and not large enough across Europe to support the heavy R&D required to build NoC technology. Therefore, we either had to accept being a small company with a limited product capability, or go global by entering the US and Asian markets in order to be able to fund the considerable investment required for a broad technical solution. Within France, workforce availability was and is quite good. French engineers are well trained in maths, they are taught to work in teams and they are taught to think through complex problems. Arteris needed all three capabilities. A number of our key employees came from the Ecole Polytechnique and the Laboratoire d’Informatique de Paris 6 (LIP6), universities which produce very well-trained graduates with the uncommon mix of network, hardware and software engineering skills required by Arteris. We were also able to get early adoption of our products from several European research organizations such as Ecole Nationale Supérieure de Techniques Avancées (ENSTA) and the Interuniversity Microelectronics Centre (IMEC) because NoC was a favourite research topic in these universities. Funding in Europe was available but limited in size. We did obtain interest-free loans from the French government’s Agence Nationale de Valorisation de la Recherche (ANVAR) and [French Innovation Agency] OSEO programmes and were beneficiaries of the generous R&D tax credits available in France. At the same time, French social charges are so high that these costs and benefits approximately offset each other. There are a number of experienced European semiconductor executives advising Arteris, and we particularly benefited from the advice of Philippe Geyres, one of our board members, who originally ran about half of STMicroelectronics.”
Q4: What key aspects of the entrepreneurial ecosystem surrounding your company that were absent (or existed only in a weak form) created the greatest challenges for growing your company? Please describe and discuss how you met/were impacted by these gaps in the ecosystem and their resultant challenges.
Janac: “The French government policy was to support national champions, which were invariably large companies like Airbus, STMicroelectronics and Bull. There are relatively few successful start-ups coming out of France so the entrepreneurial culture is not very strong. This is particularly evident in the area of middle management, where we had trouble finding people experienced with the global sales and marketing issues facing emerging companies. We also faced a limited domestic market, scepticism about whether there was a market for NoC technology in semiconductors, and difficulty in raising capital once Arteris became larger. Our solution was to keep engineering in France, where we have an excellent workforce, and put in a global sales and marketing structure that would look ‘European to the Europeans, American to the Americans, Chinese to the Chinese and Korean to the Koreans.’ Another aspect of Arteris’ international expansion was to attract investors from target international markets such as the USA and Japan.”
Q5: At what stage did you invest significant resources seeking to grow your company internationally/beyond your domestic country or region? What factors were pivotal in deciding when to seek growth internationally and where to seek that growth?
Shuler: “2009 was the big year when we decided we needed to focus on international growth and add local employees in our target regions. Our primary market was the mobile and wireless chip market and, by 2009, American and Korean companies had evolved to become the market leaders. In addition, the Chinese government started funding domestic companies whose charters were to quickly grow to become national champions. All these companies needed our technology, but would prefer to license from or accept product support from domestic sales and support people.”
Cantwell: “We hired Janac to put a global face on the company. We knew that we had development operations in France, customers throughout the world and we needed an experienced executive that had the ability to be a global citizen and provide the customers with the confidence to do business with Arteris. This single point created an inflection point in the customers’ and employees’ worldly view. We soon followed by shifting the company’s headquarters to Silicon Valley, which provided even more comfort for US customers.”
Q6: What were the biggest challenges in building growth internationally? How did you meet or adapt to those challenges?
Shuler: “The biggest challenge was how to balance the use of country representatives and distributors versus permanent employees within non-US and non-French countries. Our products are highly technical and require a lot of communication with and support to our customers after the sale. We found that although distributors in our target countries were well connected, it was difficult for them to provide the amount of post-sales resources to each customer that we wanted. What emerged is a strategy where we initially used local distributors to learn about a target market within a country. As we gained more knowledge about country-specific norms and who were the great salespeople and application engineers in that country, we then switched to a direct sales model staffed by Arteris employees. We followed this approach in Japan, Taiwan, Korea and China.
“Building local offices in these countries created a lot of operational challenges for us. It is very difficult for a small company to manage various foreign legal entities, employment laws, tax laws and financial reporting requirements. Although there are consultants that can help, we found that only we had the information required to formulate and manage this complexity. We hired an experienced general counsel and a VP of finance with lots of international accounting experience to help manage this complexity.”
Cantwell: “There were cultural challenges with having most of our technical resources in France. The French development team was very capable but nervous that development and primary architecture would shift from France to the USA. Janac was able to overcome this issue and created an environment where architecture development was led from the USA, engineering stayed in France and customers were supported globally. He realized it was critical to support customers both where they are geographically and culturally.”
Q7: What major role, if any, did key aspects of the ecosystem in the country (or countries) you first sought international growth either promote or impede your ability to grow in those international markets?
Janac: “The US ecosystem is not particularly kind to emerging European-based companies trying to penetrate the domestic market, and we had to make several attempts before we succeeded in the USA. In Asia, we initially started with distributors in Japan, Korea, Taiwan and China but the product proved too complex and engineering support-oriented, so we had to build direct organizations in these markets. We had challenges understanding the significant differences between Chinese, Japanese and Korean cultures. As an example, you really do not want a Korean country manager running Taiwanese sales no matter how well he is doing in Korea. Even large corporations prefer to deal with suppliers who are culturally compatible with them. Today, Asia represents almost 60% of our revenue.”
Q8: Seeking international growth often has both high moments and dark (low) moments. Briefly describe one high moment and one dark (low) moment in seeking international growth.
Janac: “A high point was when Texas Instruments successfully delivered the OMAP4 application processor to the market. A more recent high point was this year (2013) when the Samsung Galaxy S4 smartphone shipped with an Arteris-connected application processor SoC (the main chip in a smart phone), designed by Samsung’s System LSI Division.
“Low point: It is often difficult for European companies to penetrate the US market. Arteris went through three generations of sales people over three years before we hired the right personnel, delivered the right product features and developed a customer track record so that large US customers felt comfortable licensing our NoC IP technology. To accomplish this we had to ‘look American to the Americans’. At that point, the fact that we had all of our engineering in Paris became a non-issue.”
Cantwell: “There was a point when it became clear that we needed to hire a global sales force as the revenue was just too lumpy to build a business with. After putting in place a sales force that was culturally compatible with local customers, the business really began to take off. Janac continued to be customer focused, but mainly as a resource for ‘executive’ sales.”