Arista Networks, a maker of networking equipment for high-capacity data centres, could hardly have picked a worse time to launch.
Conceived in 2004, the idea for the company was partly inspired by Google, says founder Andy Bechtolsheim. “Cloud computing” had yet to become a buzzword in the tech industry, but the search company’s massive computing needs convinced Mr Bechtolsheim, a veteran Silicon Valley hardware engineer, that a new computing architecture was about to emerge — creating an opening for a fresh breed of start-ups.
By the time Arista was ready with its first product, however, the cloud market was still in its infancy and the company instead targeted banks involved in high-speed trading. Its first customer was Lehman Brothers. The investment bank imploded almost immediately, marking a low point in the 2008 financial crisis, and one of Arista’s first jobs turned out to be reclaiming equipment from a bankrupt customer.
“We were patient — we didn’t build a venture capital-backed company and sell out early, we took five years to build the product,” says Jayshree Ullal, president and chief executive of Arista. “The market didn’t proceed as expected, it took longer to pan out.”
Besides patience, which led to Ms Ullal and Mr Bechtolsheim being named as EY’s US Entrepreneurs of the Year at the end of 2015, Arista’s emergence owes a lot to focus, experience and a willingness to bet against tech fashion.
It was a daunting beginning. At the time, Cisco’s dominance of the networking market seemed to leave little room for start-ups. Venture capitalists “thought the networking business was mature. That benefited us, we were internally funded and we had no new competition,” says Mr Bechtolsheim.
Arista was designed to take advantage of a shift in the dynamics of the networking market. Cisco had built its lead on selling integrated equipment, based on its own chips and software, which was a model that was hard for newcomers to emulate.
For his new company, Mr Bechtolsheim — who had been one of the founders of Sun Microsystems, one of the star hardware companies of the 1990s — was able to tap a new wave of independent chipmakers, effectively building a networking software company that took advantage of new, low-cost hardware for a new market of cloud data centres.
An intense focus and organisational simplicity were built in from the beginning, he says. Part of the aim was to avoid the internal rivalries he had seen building up in other companies. “We don’t have multiple business units that compete with each other. You just do things to make yourself look good and make someone else look bad.”
Arista also avoided another common feature of fast-growing tech start-ups: a pursuit of revenue that takes companies into new markets, even when these are not closely related to its existing products and customers. “It’s easy to get distracted,” says Mr Bechtolsheim.
Staying focused and moving quickly were key. Like many disruptive new tech markets, the emergence of the cloud created the kind of opening that only occurs with secular shifts in computing. “While big companies can spot a trend, they are slow to react and find these trends to be niche markets they can’t invest in,” says Ms Ullal.
If Arista started out by swimming against the tide in a market other start-ups were shunning, then the experience and credibility of its management group were crucial to its future success. Much of its senior management team came from Cisco. Ms Ullal had joined in 1993, after the company she was working for, Crescendo, became Cisco’s first acquisition.
Mr Bechtolsheim arrived at Cisco for a spell in 1996 after selling his latest start-up, Granite Systems, after just a year.
Having a management group that has spent years working together before is “a tremendous advantage”, says Mr Bechtolsheim, who is now chief development officer and chairman. “You have a model for working together.”
The chief executive, who joined after Arista was founded, says she took some persuading to make the leap: “I had to shed some skills, rather than acquire them.” It required an intense pragmatism. “You have to learn to be hands-on and not build castles.” It also brought a sense of urgency and precariousness not usually encountered in a big company: virtually every decision became a matter of make or break, she says.
Having a management group with a record also brought much-needed credibility.
“One of the things we learnt early on is that great technology is essential, but it isn’t enough,” says Ms Ullal.
The personal record and relationships of Arista’s management were key in breaking through this bias that favours the status quo. “They had worked with us before, they believed in the product,” she says. Customers were also looking for an alternative to Cisco, says Mr Bechtolsheim — one reason he says the company would never have sold out to the networking giant, as so many others have done.
Stepping out of Cisco’s shadow may be easier said than done. Arista’s success has brought a patent lawsuit from its bigger rival and Cisco scored an important victory earlier this year with a ruling in its favour from the US International Trade Commission.
Wall Street, however, does not expect that setback to discourage customers. Arista, which went public two years ago, is forecast to have a 30 per cent sales jump this year, lifting its revenues above $1bn for the first time. For a 12-year-old company, it may not count as overnight success, but it shows the rewards that can come from getting in ahead of the next big tech trend.