Joe Terranova Arista Networks, Inc. ANET-N DON'T BUY Feb 19, 2025

It's fairly priced now. They're Cisco's main competitor. Have done very well. Take profits now, then re-enter around $245, then $225 (note: these values don't reflect the recent stock split). January will be a volatile month; Jan. 20 will be inauguration day in Washington.

It trades in classic stair-step fashion. Are riding the data centre story. Shares are up 52% in the last 6 months.

It reports next week. It needs 25% revenue growth. They are still part of the AI networking build-out and likes their correction after DeepSeek. It's now in a good place. Adequate earnings should see a positive response.

They lead in cloud network solutions, sales +20%, EPS +31%. Innovation in AI will only lift this stock.

Pendulum swings between hardware and software. With the infrastructure buildout for AI, it's also spilling over into the routers and switches from a name like this. Product not unique, but management makes the company stand out. 12-month price target of $123.

Owns many shares. He targets $138.50, so now is a great buying opportunity. Buy a third here, another at $78, then at $73. They supply ever-better switches and routers to the cloud stack.

Buy at these levels. Upstream of the hyperscalers like META, MSFT and AMZN. Came off with concerns on hyperscalers' AI spending and ROI. Still in early innings, demand will come back, and this name will grow quickly.

Trading ~37x forward PE, for 15% EPS growth. Below 200-day MA, but so is much of the market. Customer concentration risk; any capex pullback by hyperscalers can significantly impact results. Be careful. Faces strong competition.

Supplier of hardware like switches and routers for the hyperscalers within the cloud. Good managers. Tremendous company that can compete with the leader, CSCO. 

Very volatile, since it's expensive and grows at a pretty decent clip. Networking solutions for AI, among other things. Coming into 2025, his team was super-optimistic about networking. But questions arose on how much inter-connectivity of data centres the hyperscalers will need, so he's on the sidelines. 

Wait and see. He'll revisit at some point.