Earnings Recap — Q3 FY2026
CY Q3 2026 · Reported July 7, 2026 · N/A
Penguin Solutions, Inc. reported Q3 FY2026 revenue of $479M, a beat of 17.5% against consensus, and EPS of $0.84, a beat of 55.8%.
The Buildout is a website built on agentic AI tools, currently in beta, and could have factual errors.
Penguin's record quarter and raised guidance underscore the accelerating demand for AI infrastructure beyond hyperscale, particularly for inference and agentic AI workloads. The surge in Integrated Memory revenue and backlog highlights memory as a critical bottleneck in AI factory buildouts, reinforcing the need for memory-centric solutions. The company's expanding enterprise, neocloud, and sovereign AI customer base signals a broadening AI infrastructure buildout across multiple segments.
Penguin Solutions delivered record Q3 FY2026 net sales of $479M, up 48% YoY, with non-GAAP EPS of $0.84, up 79% YoY. Integrated Memory surged 111% YoY to $275M, while Advanced Computing grew 4% YoY to $138M, with non-hyperscale AI infrastructure up 81% YoY. The company added four new AI infrastructure customer logos and expanded engagements with existing customers, including Deepgram and a Tier 1 financial institution. Non-GAAP gross margin was 28.1%, down 3.6 points YoY, and operating margin improved to 13.4%. CFO Nate Olmstead announced his departure, with Aaron Johnson named interim CFO.
Management raised full-year FY2026 net sales and non-GAAP EPS guidance, citing strong AI-driven demand and a growing backlog. They highlighted continued strength in memory and AI infrastructure, and noted that the demand environment has strengthened as AI transitions to production-scale inference and agentic AI. Management also mentioned that they expect Q4 gross margin pressure from less favorable memory pricing and noted extended lead times and higher memory costs as potential headwinds.
“We believe that for enterprises, sovereign AI initiatives and new cloud customers, AI factories are becoming essential for optimizing time to first token performance and token economics for inference.”
on AI factory demand
“Our strong results reflect a fundamental shift. Agentic AI is driving sustained structural demand for memory. Reinforcing our view that this AI-driven demand is more durable than a traditional cyclical memory upturn.”
on Memory demand
“We believe we are still in the early stages of a significant long-term profitable growth opportunity.”
on Growth outlook
How much of the Integrated Memory guidance raise is driven by pricing versus new products like CXL cards? And what are you shipping into hyperscale?
Kash Shaikh: The higher outlook reflects both volumes and pricing. CXL is a growing new product line but a part of the business; the majority is data center-focused products for OEM customers. One hyperscaler customer is buying memory modules for data center products, but it's a portion of the overall business. Nate Olmstead added that the hyperscaler is not a significant piece of the FY27 view.
Can you provide guardrails on Advanced Computing growth for FY27?
Nate Olmstead: Preliminary view suggests mid-teens growth for Advanced Computing, with the Edge wind-down and Meta transition largely behind, so growth will come from non-hyperscale AI infrastructure. Kash Shaikh added that bookings-to-revenue lag of 3-6 months provides visibility into the first half.
What are you seeing in services engagement across the broader customer base, and are you doing anything different in the memory business as it becomes more strategic?
Kash Shaikh: Services are a strategic advantage, providing full system integration and management for 3-5 years. In memory, we are focusing on data center demand, which is exploding due to agentic AI, and we are strategically focusing on data center products.