With two of our portfolio holdings staring down growing competition, I’m selling them both for gains:

• Arista Networks (NSDQ: ANET): Up 53% since it was recommended in March.
• Xilinx (NSDQ: XLNX): Up 17% since recommended in June.

Arista Networks is an old-school hardware company that’s benefited from the rise of cloud computing. When operating a cloud-based platform, a company’s network might have to move terabytes of data every minute, if not every second, and that requires blazing speeds. By marrying its hardware switches with a customizable Extensible Operating System, Arista created fast, low latency switches that are basically custom-made for those needs.

That’s driven a blistering pace of growth for the company, with 63% three-year average revenue growth and net income growth better than 78% over the same period. That’s translated to big gains for the stock, which is up 53% since we added it to our portfolio back in March.

Success always attracts competition though, just ask the folks that opened frozen yogurt shops a few years ago who now find competitors on every corner. It gets harder and harder to make a buck as competition heats up, and that’s true whether your product is old school or new school.

Even as more competition is entering the field, Arista’s valuation is starting to look a bit stretched. Currently trading at more than 40 times trailing earnings, its valuation is well above the industry average. That would be fine if Arista had the industry to itself, but it doesn’t. It’s also now well above our target price and I can’t justify bumping it up at this point.

Sell Arista Networks.

Speaking of competition, I’m also selling Xilinx (NSDQ: XLNX) because of growing competition.

Most processors are designed to serve a specific function, but Xilinx’s main claim to fame is its programmable logic devices (PLD), which can be reconfigured to meet whatever a user’s needs are. Xilinx’s main competitor was Altera, which was acquired by Intel (NSDQ: INTC) almost a year ago. The pair up hadn’t result in much competitive change in the industry, but a couple of weeks ago Intel announced that it has married its Xeon server CPU and an Altera chip using a super high-speed optical interconnection. That marriage could potentially be a game changer, putting Xilinx at a bit of a disadvantage.

As with any other advancement there’s a fair bit of debate amongst industry watchers about what sort of impact it will have on the industry. Both sides have fair points and, to be quite honest, this newest Intel product is a bit of an apples-to-oranges comparison to Xilinx’s PLDs. But it’s a good indication that Intel plans to bring it on new product development on that front.

And just like Arista, Xilinx has gotten overvalued. It’s trading at more than 24 times trailing earnings, well above its peer group and its own five-year average of 20 times. Revenue and earnings growth also appears to be set to slow as management predicts 6% revenue growth for its current fiscal year, predicated on a strong fourth quarter.

While I still like Xilinx for the long haul – it doesn’t rest on its laurels when it comes to product development either – at this point I think we’ll be able to find better opportunities. At this point, I think our best move is to lock in our almost 18% gain.

Sell Xilinx.