Sun founders on cloud, Google and AI

recession worries? – Not at Arista Networks, at least not for now. “We got surprisingly good numbers for that a few days ago presented in the third quarterand we anticipate strong demand for our products and services over the next few years,” explains Andrew of Bechtolsheim in conversation. That means something. Because the 67-year-old founder, main shareholder, chairman of the board of directors and chief developer of the American manufacturer of high-speed network components headquartered in Silicon Valley is an old hand in the business with excellent connections and knows what he is talking about.

Dissatisfied with the technological equipment at the University of Munich, the native Bavarian moved to the USA as a young student and finally ended up there from Stanford University. In 1982, due to favorable general conditions, he abruptly abandoned his doctorate there founded Sun Microsystems together with three colleagues. In the years that followed, this company turned the IT world upside down with development modern, cheaper Unix workstations around and grew in the process of competing with sluggish competitors such as IBM, Digital equipment or Hewlett-Packard spectacular.

Early Google investment – ​​in the fast lane with Arista

In the 1990s he founded two other companies, which he was able to sell at enormous profits, so that he did not have to start selling his shares in Sun to Oracle financially independent. Today, Andy, as he is known in America, is operationally in the fast lane with Arista. He owns around 15 percent of the shares in the company founded in 2004 Santa Clara, which has a market value of around $41 billion in the stock market today. And that’s not all. Because the man is not only successful as a passionate technical developer, but also as a skilled investor. Because of the convincing search engine advertising concept, he had already invested the first $100,000 in Google before the company was officially founded.

With an estimated net worth of around 10 billion dollars, he really shouldn’t have worked for a long time, but he is driven by the need and passion to “make the world a better place”, as he admits – that was also the case when Arista was founded . “We had friends on google, and from them we learned that the internal networking of cloud farms is more demanding than in normal company networks.» In the early days of the cloud business, the search engine group had started to develop its own switches, i.e. the building blocks for fast communication between different technical components of a network.

Niche recognized for fast «network switches»

Apparently there was no commercial product at the time that met Google’s requirements. Andy and his colleagues took this opportunity and got into the business with it. “We expected the cloud to transform the IT business and focused our product development on potential customers in this area – and fortunately we were right. This year, software sales in the cloud business should be around $500 billion if you count Sestimates by the consulting firm Gartner believes, and in the next four years it will probably double again,” he predicts.

«The fastest network switches and routers are in high demand because the cloud data centers with their 100,000 or even more servers have to work together efficiently. you need for that Petabits of aggregate bandwidth – and we are the leading provider of Internet switches with capacities of 100 gigabits and beyond. Our market share is currently more than 40 percent, Cisco’s is just under 30 percent, and the rest of the market is divided among other providers,” he continues.

Microsoft and Facebook are among the largest customers

In fact, Arista Networks has been working on the further development of cloud technology for a good ten years. Even as a tiny startup, they began discussing with Microsoft how to build faster networks, Andy continues. «The software giant from Redmond finally responded to our proposed solutions, and Microsoft is still our biggest customer today; but we also have good, long-standing relationships with other companies – such as Facebook.”

What defines Arista’s success? “It’s the combination of the best hardware with intelligent software solutions that has helped us to break through in a mature market,” argues von Bechtolsheim. Around 90 percent of research and development expenditure goes into software development, which in turn is based on tried and tested in-house hardware. When they started the company, it’s highly unlikely that a venture capitalist would have jumped in since the market was almost was completely dominated by Cisco. “Today, we’re not only significantly more profitable than Cisco, we’re even gaining more market share,” he says with a proud smile.

Coordination is a matter for the boss

In fact, he personally takes care of it through regular phone calls to suppliers like that specialized chip manufacturer Broadcom and our own customers that there are no errors during the long lead times on the hardware side and that in the end the right products and services arrive where they are needed.

If you believe his expertise, the trend towards migrating IT applications to the cloud will continue in the coming years. «Conventional server parks of small and medium-sized companies are usually only used to 10 or 15 percent. The associated investments are therefore hardly worthwhile. In the cloud, on the other hand, you only pay for what you actually use. In addition, the data and processes in the cloud are usually much more secure because the providers have greater expertise and also use it.” In any case, it will probably be difficult for smaller companies with limited resources to find suitable IT staff in the future.

So are natural cloud monopolies emerging right now? “Amazon and Microsoft are investing heavily in their software offerings. As soon as they compete with others and are cheaper, it becomes difficult for the competitors. So they have to find niches in which they can prosper.” The two giants are very agile in this business, but Google and Oracle are catching up, while IBM has fallen behind. Although the big three dominated, the competition played in the business models – for example in the selection of the databases.

Silicon Valley remains the center of innovation

Andreas von Bechtolsheim is convinced that Silicon Valley will remain the center of innovation. “What is crucial is the unconditional belief that if the future is going to be invented somewhere, it will be here,” he argues. Culturally and commercially, the Valley has always been looking for the next big opportunity, there is enough venture capital and the expertise of all the people who have run the business in the past. The model works well, and since all the infrastructure is in place to set up new companies very quickly, it is hardly surprising that around a third of all venture capital funding in the USA has been made in Silicon Valley and that many of the companies end up found their way to the stock exchange.

In his eyes, most innovations are incremental in nature. This means that the new smartphone is better than the old one, and the old one was better than the one before it. This is crucial in order to always be able to offer consumers something new. “However, there are also areas such as artificial intelligence and machine learning where incredible progress has been made in the last ten years. What is now possible is almost beyond human imagination,” he explains enthusiastically. And in these specialist areas, massive investments are being made and intensive research is continuing.

Artificial intelligence breaks through

«The latest results, for example at the Further development of the Lambda language model from Google, where you can talk to the computer and it responds like a human being, are just amazing.” if Alan Turing still with us, he would surely agree that this machine passed the Turing test. Ultimately, it is no longer a matter of replicating the abilities of the human brain, but of doing things that humans are normally not capable of. “I’ve always been fascinated by the possibilities of artificial intelligence, and now they are developing much faster than I expected and many predicted a few years ago.”

Of course, he understands some concerns that technology could make a biased decision about anything. On the other hand, he criticizes the proposal of the European Union, which wants to have this decision-making process reviewed in the future, for one simple reason: “With a billion parameters in a program, you never know how a decision really came about, it’s impossible . The purpose of artificial intelligence is to solve problems that defy simple, normal rules – otherwise you wouldn’t need the whole thing,” says von Bechtolsheim.

The self-driving car remains a dream for the time being

In this context, he ventures the prognosis that the dream of a self-driving car will remain such for the time being, because the technology will be overwhelmed with the wide range of possible special cases in normal road traffic for the time being. In addition, artificial intelligence is more likely to endanger the jobs of office workers in banks or in public administration than those of craftsmen, for example. And when it comes to the moderation of content in the future, such as on Twitter, it should objectively reflect the entire spectrum of opinion and not unilaterally suppress some aspects, as tends to be the case in today’s media landscape. “I believe conservatives are right to worry that their opinions are being underrepresented in the media.”

Andy recommends the Europeans to set up a breit network of fast charging stations for electric cars otherwise the planned adoption of vehicles with conventional drives will fail because of this. The European car manufacturers would have to break new ground in the design of their vehicles and production facilities anyway, and Tesla’s agility is a role model here. He is skeptical about cryptocurrencies: “Bitcoin is a Ponzi scheme,” he thinks, and in terms of efficiency, the crypto scene relies on inadequate technologies to develop digital currencies.