Astera Labs and Arista: Why Higher Gross Margins Have Not Produced Higher Operating Margins

Astera Labs leads Arista in gross margin but trails in operating margin. R&D expenses as a share of revenue and stock-based compensation reveal differences in the two companies’ current profit structures. Scorpio’s ramp-up is already supported by its position in revenue, while Orion and Etherlink SU-144 still require validation through subsequent deliveries and revenue disclosures. Current financial reports support a comparison of company-wide operating efficiency, but are insufficient to determine which rack interconnect approach will ultimately be more profitable.

Astera Labs and Arista: Why Higher Gross Margins Have Not Produced Higher Operating Margins: concept illustration
Original concept illustration; not a data chart or a news photograph

Beyond Gross Profit, Expense Structures Change the Ranking

In the third quarter ended September 30, 2026, Astera Labs reported approximately $585 million in revenue, compared with approximately $3.658 billion for Arista Networks. On the same GAAP basis, Astera Labs had a gross margin of 72.3% and an operating margin of 30.6%, versus 63.6% and 44.9%, respectively, for Arista. My judgment is that comparing these companies’ ability to generate profit requires looking at both the gross profit their products generate and the operating expenses needed to sustain the business. A high gross margin provides room for investment, but cannot by itself establish that a company ultimately retains more operating profit.

Evidence: [1] [2]

R&D Spending Differs in Both Absolute Size and Relative Burden

Astera Labs spent approximately $171 million on R&D during the quarter, or 29.2% of revenue; Arista spent approximately $487 million, or 13.3%. Arista spent more in absolute terms, yet R&D consumed a smaller share of its revenue. This supports one explanation: revenue scale and expense structure jointly affect the operating-margin gap. But consolidated financial statements alone cannot establish how much of that gap comes from product maturity, R&D efficiency, or business mix, much less establish that a particular interconnect technology is inherently more profitable.

Evidence: [1] [2]

Stock-Based Compensation Matters, but Must Not Be Double-Counted

Astera Labs recorded approximately $82.6 million in stock-based compensation during the quarter, equivalent to 14.1% of revenue; Arista’s corresponding ratio was 3.8%. This is another important clue to understanding the GAAP profit gap. Some of Astera Labs’ stock-based compensation is already included in R&D expenses, so adding all stock-based compensation on top of R&D and treating it as an additional cost would be incorrect. Nor does it represent an equivalent cash outflow in the period. R&D expenses as a share of revenue and stock-based compensation therefore need to be read together, rather than added as non-overlapping expense items.

Evidence: [1] [2]

Astera Labs Already Shows Signs of Expenses Being Spread Across More Revenue

Astera Labs’ first-quarter GAAP gross margin was 76.3% and its operating margin was 20.1%, compared with 72.3% and 30.6%, respectively, in the third quarter. A declining gross margin alongside a rising operating margin shows that these metrics can move in different directions. As an inference, this is consistent with operating expenses declining as a share of revenue as revenue expands, but comparing two endpoints is insufficient to establish that the improvement will continue. Scorpio has become the largest product family by revenue, and the 320-lane X-Series has entered high-volume production, providing concrete evidence of the product ramp-up. Its standalone profit remains undisclosed, so the change in the company’s margins cannot be attributed entirely to Scorpio.

Evidence: [5] [1]

Cash Flow Adds Evidence but Cannot Establish Business-Level Attribution

Using a consistent definition of operating cash flow less purchases of property and equipment, Astera Labs generated approximately $268 million in free cash flow in the first three quarters, equivalent to 20.9% of revenue over the same period; Arista generated approximately $3.793 billion, or 40.3%. Arista’s company-wide cash flow performance was stronger over this window. However, these figures cover the first three quarters and cannot be treated as the same measure as a single quarter’s operating margin. Cash flow is also affected by working-capital changes, and Arista’s consolidated figures alone cannot establish the returns generated by its AI interconnect business.

Evidence: [1] [2]

Rack-Scale Offerings Expand the Opportunity, with Limits on What Can Be Inferred About Revenue

Orion is a rack-scale reference architecture designed and validated by Astera Labs. The company plans to make it available through OEM/ODM partners starting in Q1 2027, with complete racks and systems manufactured, licensed, and sold by those partners. Arista’s announced Etherlink SU-144 proposes an open Ethernet rack solution designed to support up to 144 XPUs within a single rack over a single hop. This supports the inference that both companies want to expand their roles in rack interconnects. But reference designs, topology capabilities, and partner lists establish only that solutions and partnership arrangements have been announced; they cannot directly serve as evidence of committed purchase volumes or recognized revenue.

Evidence: [3] [4]

The Key Is How Much Profit the New Business Retains

The conclusion supported by current evidence is that Astera Labs has demonstrated product progress through Scorpio’s changed position in revenue, while Arista leads in company-wide operating margin and cash flow metrics. The unresolved question is how actual deliveries of the new offerings will change each company’s revenue and expense structure. Subsequent disclosures of product revenue, related profit contributions, and deliveries would allow further assessment of whether the new business improves operating efficiency. At this stage, turning the potential of a technical solution directly into a verdict on profitability still goes beyond what the evidence supports.

Evidence: [1] [2] [3] [4]

What to watch next

  • Whether Astera Labs’ R&D expenses as a share of revenue and GAAP operating margin continue to improve after Scorpio ramps up; if revenue grows while the expense ratio rises again, the explanation that expenses are being spread across more revenue will need to be reassessed.
  • Whether Orion is delivered through OEM/ODM partners as planned, and whether Astera Labs discloses the resulting product revenue.
  • Whether Arista discloses actual deployments, committed purchases, or revenue contributions for Etherlink SU-144, allowing its design capabilities to be connected to commercial results.
  • How much of each company’s subsequent cash flow changes comes from operating profit, and how much comes from working-capital changes such as receivables, inventory, and deferred revenue.

Sources and verification

Topic originated from 潘大的投研笔记, published on 2026-11-04 16:59 (UTC+8). This article was independently organized using the primary sources below. Read observations separately from interpretation.

  1. Astera Labs 2026 Q3 results — released 2026-11-03; quarter ended 2026-09-30
  2. Arista Networks 2026 Q3 results — released 2026-11-03; quarter ended 2026-09-30
  3. Astera Labs unveils Orion — 2026-10-12
  4. Arista announces rack-scale scale-up/scale-out architectures — 2026-10-07
  5. Astera Labs 2026 Q1 results — released 2026-05-05; quarter ended 2026-03-31
  6. ALAB与ANET:两份财报,看清AI互联的赚钱逻辑

← All research notes · Research sources and review method · Editorial standards