Arista Networks' $8.9 Billion Procurement Pile-Up Spooks Investors Despite Record Quarter

Arista Networks reports record Q1 revenue but stock falls 18.6% as supply shortages and margin pressure overshadow strong AI demand and raised guidance.

The arithmetic governing supply and demand has turned unusually lopsided for Arista Networks. Clients building out artificial intelligence clusters are placing orders faster than the company can source the silicon, optics and memory chips needed to fulfil them, leaving the networking specialist in the uncomfortable position of booking record revenue while watching its shares get hammered. Over the past week, the stock shed 18.6 percent to close at €120.14 in Frankfurt — still up 44.4 percent over twelve months, but well below the April peak of over €150.

Revenue in the first quarter of fiscal 2026 climbed 35.1 percent to $2.709 billion, easily beating the company's own forecast. Adjusted earnings per share of $0.87 also came in ahead of the $0.79 analysts had pencilled in. Yet the market's reaction was anything but triumphant. The sell-off owes less to past performance than to a sobering outlook: management's guidance for the current quarter points to revenue of roughly $2.8 billion and adjusted EPS of about $0.88, a trajectory that failed to justify the lofty expectations baked into the stock after its extended rally.

The bottleneck sits squarely in the procurement pipeline. Arista's management described “broad-based industry shortages” spanning wafers, CPUs, optical components and memory. In some cases, lead times of 52 weeks are now considered routine, and the company warned that the constraint could persist for one to two years. To lock down critical parts, Arista has dramatically boosted its purchase commitments, which swelled from $6.8 billion to $8.9 billion. That stash of inventory-in-waiting is a necessary evil — without it, many of the orders flooding in from hyperscale clients would never convert into revenue.

Higher procurement costs, however, are squeezing margins. Gross margin slipped to 62.4 percent in the first quarter, down from over 65 percent a year earlier. The company's exposure to a handful of giant cloud customers — Microsoft accounted for roughly 26 percent of sales and Meta for 16 percent — gives those “cloud titans” considerable leverage to negotiate favourable pricing on large deals. For the current quarter, management guided gross margin in the 62 to 63 percent range, underscoring the trade-off between volume and profitability.

Despite the supply headwinds, Arista lifted its full-year revenue forecast to about $11.5 billion, representing growth of 27.7 percent. The AI networking revenue target was raised from $3.25 billion to $3.5 billion. CEO Jayshree Ullal expects the ramp-up to shift from pilot phase into a broad-based demand surge across three networking tiers: scale-out, which is already contributing meaningfully; scale-across, which is developing faster than anticipated; and scale-up, which the company sees as a 2027 story.

In the meantime, Arista is looking to win on efficiency. At the Optical Fiber Conference, it unveiled XPO, a new optical interconnect format backed by more than 100 module suppliers. The technology is designed to slash rack-count in data centres by up to 75 percent and reduce floor-space requirements by as much as 44 percent. Tighter clusters mean shorter cables and lower power consumption — exactly the kind of density boost that becomes critical when every chip and every watt must be used sparingly. But XPO does nothing to alleviate the immediate chip shortage, and its impact on scale-up applications will not materialise for several years.

Arista sits on a cash pile of over $12 billion, providing ample cushion to ride out the supply-chain storm. The resumption of price stability — or at least a ceiling on further declines — now hinges on the company's ability to reassure investors that margins have stabilised and that delivery timelines are improving. Upcoming presentations at the Needham, J.P. Morgan and William Blair conferences offer the next opportunity for management to flesh out its roadmap on lead times, margin trajectory and the evolution of AI networking products.