Last Friday, Arista Networks (ANET) went public. The nine-year old equipment maker saw its shares soar from $43 to $55 on the first day of trading. At its peak, the company was worth almost $4 billion. Is Arista worth the hype?
Arista makes Ethernet switches and is quickly taking market share from the networking kingpin Cisco Systems (CSCO). Arista switches are disrupting Cisco's business for several reasons.
First, Arista's equipment is less expensive, because Arista uses more off-the-shelf semiconductors, instead of custom-designed parts. Cisco's equipment is filled with application-specific integrated circuits (ASICs). These parts take time to design and test. Arista uses lots of generic parts from a wider range of vendors, and that enables the company to ride the cost curve lower. To drive silicon costs lower, Cisco has to sell more boxes stuffed with its own custom circuitry.
Second, Arista's equipment runs an operating system called EOS (Extensible Operating System), which is based on Linux. This means software developers can more easily connect into Arista's switch fabric using a simple application-programming interface (API). By using Arista's APIs, developers can build all sorts of software programs on top of the switch. Software developers have built many open-source tools on top of Arista's switches, and that helps customers save money. Cisco's equipment runs Cisco's proprietary code, which can be difficult to build on. Cisco customers complain that they can't "see" into the switch and don't know how well it is functioning.
Third, Arista's switches connect together seamlessly through something called software defined networking. Arista and Cisco approach networking very differently. Cisco approaches networking from a hardware prospective, while Arista approaches networking from a software perspective. Cisco's hardware requires lots of custom set-up and programming.
When a large data-center needs more capacity, adding more Cisco switches can be difficult, expensive and time-consuming. Cisco switches can take hours to configure, because they are not provisioned out-of-the box. The IT staff has to set up each switch.
Arista's equipment plugs into the network and is instantly provisioned through software. Furthermore, as you add equipment, that additional capacity is available throughout the network virtually.
When a company runs Cisco equipment, an administrator may have some equipment running at half capacity. But that capacity isn't available to the rest of the network. You end up with wasted resources. Arista's software takes that unused capacity and turns it into virtual capacity in the cloud.
By moving the functionality from hardware into software, network administrators can manage an army of switches in the cloud instead of having to physically manhandle them when they need to move capacity around. These features come in handy when you have thousands of switches to manage in dozens of data centers around the world.
Recently, Cisco has made a lot of noise about how it is going to bury Arista, but it's all bluff. Arista's equipment turns the whole networking industry on its head. To add more capacity, Cisco wants to sell you more expensive boxes. Arista wants to sell you less expensive generic hardware that runs its software.
Additional capacity is added to the cloud and provisioned dynamically through software. If the Chicago data center needs more networking capacity, you can provision it in the cloud, instantly. With Cisco's approach, you have to buy a new switch for Chicago, ship it over there and have someone tinker around with it until it works correctly.
The ease of use, the raw speed and the open-source nature of Arista's equipment have made Arista the only company that has been able to chip away at Cisco's 70%-plus market share in a very long time. Analysts estimate that Arista has about 6% of the Gigabit Ethernet market.
According to Cretan Research, the market for Gigabit Ethernet switches was worth about $6 billion in 2013 and is estimated to grow at a 19% annualized rate to $12 billion by 2017. Arista makes 10, 40 and 100 Gigabit equipment.
Networking junkies have been feasting on Arista's equipment for some time. For example, from 2010 to 2013, Arista grew revenue at a compounded annual growth rate of 71%. Revenue grew 87% in 2013, to $361.2 million.
While analysts haven't begun publishing financial models, it would not be crazy to assume that Arista can continue this hyper growth. The company should be able to easily get to $1 billion in revenue by 2016. A mid-60%s revenue growth rate (down from 71%) gets you to a $1 billion revenue estimate.
I believe that Arista can follow in the footsteps of VMware (VMW). In 2005, VMware had revenue of just $387 million. Today, VMware's fiscal 2014 revenue estimate is $6.02 billion, and that company sports a market cap of $41 billion.
Arista is profitable, but the momentum hounds will focus on the revenue growth rate to drive the shares. Any slowdown in that rate will be met with severe punishment. In 2008, VMware saw its stock fall 78% because of the recession and a slowdown in sales.
While Cisco struggles to get its act together, I believe that Arista can grow unchecked for the next three years. For example, VMware is still taking market share from the big server providers such as Hewlett-Packard (HPQ) and Dell (DELL).
I would look for opportunities to build a long-term position in the shares of Arista Networks. I would view dips and scary drops as entry points.