PENG Q4 CY2025: Enterprise AI Uptake and Memory Growth Offset LED Weakness
Penguin Solutions (PENG) delivered Q4 CY2025 results slightly ahead of expectations, signaling steadier demand in enterprise AI and memory while the LED business stayed soft. Revenue came in at $343.1 million (1.2% above estimates) with adjusted EPS of $0.49 (10.6% beat). Operating margin held at 5.7%, inventory came down sharply, and billings grew double digits year over year. Shares moved to $22.54 (from $21.65 pre-earnings), and management kept full-year adjusted EPS guidance at $2 (midpoint).
Quick numbers managers care about
- Revenue: $343.1M vs $339.1M estimate (flat YoY, 1.2% beat)
- Adjusted EPS: $0.49 vs $0.44 estimate (10.6% beat)
- Adjusted EBITDA: $45.24M vs $43.21M estimate (13.2% margin, 4.7% beat)
- Operating Margin: 5.7% (in line YoY)
- Inventory Days: 79 (down from 96 last quarter)
- Billings: $262.2M (+23.1% YoY)
- Market Cap: $1.13B
What management signaled
Enterprise AI demand is scaling beyond hyperscale pilots. CEO Mark Adams highlighted hands-on workshops and custom system design that are converting into wins across advanced computing and memory. Operations tightened up with lower inventory days, though LED weakness continues-especially in China and among large U.S. OEMs.
Looking ahead, growth is expected from broader enterprise and sovereign AI deployments, stronger memory demand, and a more diversified customer mix. CFO Nate Olmstead stressed supply chain execution in memory and noted current outlooks assume no near-term hardware sales to hyperscale customers.
Where demand is moving
- Enterprise AI: Move from pilots to production. These deals now make up a larger slice of the pipeline.
- Advanced computing diversification: New clients in defense and education; active discussions in financial services, oil and gas, and telecom-reducing concentration risk.
- Integrated memory: Healthy demand with early interest in CXL and continued investment in optical memory appliances for AI performance needs.
- LED: Ongoing pressure; focus on specialty offerings and cost discipline to protect profitability.
- Ecosystem partnerships: Deeper work with NVIDIA, AMD, and CDW to offer fuller infrastructure stacks and expand managed services/software.
Why this matters for management and sales
If you sell into data-heavy enterprises, timing is improving. Buyers are shifting budgets from trials to real deployments, and they want complete solutions-compute, memory, software, and services-sourced through trusted partners. That favors packaged outcomes, not just hardware quotes.
Memory is a swing factor. Demand and pricing look favorable, but supply may be tight. Set clear delivery assumptions in your pipeline, qualify deals by required memory configurations early, and build options with multiple SKUs. Meanwhile, steer LED conversations to specialty and margin-protected niches.
Action checklist
- Prioritize accounts with near-term AI deployment roadmaps (IT + data science + security stakeholders in the room).
- Lead with outcome-based bundles: compute + CXL-ready memory + software + services.
- Qualify memory needs up front; add supply lead times to proposals and SOWs.
- Target sectors showing momentum: defense, education, financial services, oil & gas, telecom.
- Co-sell with channel partners (NVIDIA/AMD ecosystems, CDW) to reduce friction and speed deployment.
- For LED, emphasize specialty SKUs and service wrap to protect margin while volumes stay soft.
What to watch next
- Pace of enterprise and sovereign AI bookings and deployments in 2H.
- Memory component availability and delivery performance amid supply constraints.
- Customer mix diversification across advanced computing and memory.
- LED segment profitability trends and product mix shifts.
- Depth of new partnerships and expansion of managed services/software offerings.
Context and helpful resources
CXL is becoming a key ingredient in AI-era architectures. For background, see the industry consortium's overview of Compute Express Link here. For executives planning workforce upskilling around AI deployment, browse role-based programs on this page.
Bottom line
Enterprise AI and memory are doing the heavy lifting, and tighter operations are helping. LED remains a drag, but specialties and cost control can contain it. If you align your pipeline to production-grade AI projects, secure memory early, and sell with ecosystem partners, the next few quarters set up well for closing larger, integrated deals.
Your membership also unlocks:
