Why Penguin Solutions Stock Is Surging After Earnings

Shares jumped after Q4 revenue grew 68%, adjusted earnings beat estimates, and management raised its fiscal 2027 outlook as AI infrastructure and memory demand accelerated.

Penguin Solutions stock was up about 15% to roughly $74 shortly before 11:45 a.m. ET Wednesday, even as the broader market traded lower.

The catalyst is clear, as Penguin reported a much stronger fiscal fourth-quarter result after Tuesday's close and, more importantly, raised its expectations for fiscal 2027.

Penguin Solutions Q4 results

• Revenue: $566.7 million
  • Up 68% year over year
  • Versus roughly $520 million expected
• Adjusted EPS: $1.00
  • Up 133% year over year
  • Versus $0.77 expected
• Adjusted operating income: $89.8 million
  • Up 129%
• Adjusted EBITDA: $93.3 million
  • Up 115%

Penguin's revenue was roughly flat year over year during the first half of fiscal 2026. Growth then accelerated to 48% in Q3 and 68% in Q4, producing 58% growth in the second half. Management tied that acceleration to stronger AI infrastructure demand, Integrated Memory and the company's increased focus on neocloud and enterprise customers.

Fiscal 2027 guidance

Penguin now expects:

• Revenue: Approximately $2.43 billion at the midpoint
• Revenue growth: Approximately 40%
• Adjusted EPS: Approximately $4.45
• Adjusted EPS growth: Approximately 55%

Last quarter, management had provided only a preliminary view calling for approximately 30% revenue and adjusted EPS growth from the midpoint of its then-current fiscal 2026 outlook.

Using that same starting point, management says its new revenue midpoint would represent roughly 45% growth.

Wall Street consensus before the report was around:

• FY2027 revenue: $2.21 billion
• FY2027 adjusted EPS: $3.38

Penguin's new midpoint is therefore about:

• 10% above the prior revenue consensus
• 32% above the prior adjusted EPS estimate

The earnings call gave more context for why management became more confident.

When Goldman Sachs asked what had changed over the previous 90 days, CEO Kash Shaikh pointed directly to AI infrastructure and bookings.

“We have booked more than we expected.”

He said bookings and pipeline strength were the primary reasons Penguin increased its Advanced Computing outlook and ultimately raised guidance for the company as a whole.

The market is repricing more than one unusually good quarter.

AI-driven businesses are now most of Penguin's sales

Penguin sits downstream from the major GPU makers. It designs, builds and manages AI data-center systems for customers while also selling memory solutions used around increasingly memory-intensive workloads.

Management groups non-hyperscale AI infrastructure and Integrated Memory together as its “AI-driven businesses.” Those businesses are now large enough to drive most of Penguin's consolidated results.

Q4 AI-related growth

• Integrated Memory revenue: $340.8 million
  • Up 158% year over year
  • Up 24% sequentially
• Non-hyperscale AI infrastructure revenue:
  • Up 99% year over year
• AI-driven businesses combined:
  • 78% of total Q4 sales
  • Up 141% year over year

During the quarter, Penguin added six new AI infrastructure customers, including four neocloud providers.

Management also described a multi-year project to deploy and operate a 36,000-GPU AI factory in Norway for a neocloud customer that Penguin says already has $10 billion of contracted compute from a leading AI lab.

Penguin is describing a customer that already has a large contracted demand base.

Asked when the recently announced neocloud projects would begin contributing, Shaikh said most had already started. Some data centers are being built, some are already being managed, and others are in procurement.

“They are not future, they are already happening.”

Management said it had been working on some of these projects for several months even though the bookings accelerated more recently.

For these customers, Penguin can do more than deliver hardware. It can help design the architecture, procure hardware, build and deploy the infrastructure, and then operate the AI factory.

Those managed-service relationships typically run three to five years.

That creates a fairly different economic relationship from simply selling a rack of servers. Penguin can participate in the initial deployment and remain involved once the data center is running.

Management says the unit economics of these neocloud projects are similar to its enterprise on-premise deployments, while the broader service relationship gives Penguin additional opportunities to attach ClusterWareAI software and managed services.

The broader non-hyperscale buildout has been visible elsewhere in AI as well. Stock Unlock's recent analysis of Nvidia's earnings showed particularly fast growth from AI clouds, industrial and enterprise customers outside the traditional hyperscalers.

Memory demand may have more runway than the Q4 numbers suggest

Integrated Memory has become Penguin's largest business, accounting for 60% of company revenue in Q4.

The 158% year-over-year growth already stands out, but management's comments on backlog make the near-term setup more interesting.

Despite another record revenue quarter for memory, Shaikh said memory backlog was higher than the revenue recognized during Q4 and now extends at least four quarters. Management cited this backlog as one reason it felt comfortable raising its outlook.

Penguin is also seeing both higher memory volume and higher pricing.

Management argues that its data-center exposure may make some of this demand more durable than a normal memory cycle. The company is increasingly selling into inference workloads, where AI applications require more memory as usage scales, rather than relying primarily on consumer memory demand.

Newer products such as Penguin's CXL memory expansion cards are also beginning to contribute to the pipeline and bookings. Management said demand for those products is increasing as inference deployments expand.

There is still cyclical risk in memory. But the combination of a four-quarter-plus backlog, higher volumes and pricing, and increasing data-center exposure gives more substance to the 50% fiscal 2027 growth guidance for Integrated Memory.

Why earnings are growing faster than revenue

Penguin expects revenue to grow around 40% at the fiscal 2027 midpoint while adjusted EPS grows approximately 55%.

The operating leverage was already visible in fiscal 2026.

Full-year operating leverage

• Revenue growth: 26%
• Operating expense growth: 4%
• Operating income growth: 44%
• Adjusted EPS growth: 51%

Operating expenses fell by 3.4 percentage points as a share of revenue during the year.

Management expects the same basic mechanism to continue in fiscal 2027: sales and gross profit dollars grow rapidly while operating expenses increase much more slowly.

The earnings call provided useful detail on why management thinks it can scale that way.

More than 95% of Penguin's engineering staff is located in India, with memory engineering also located in Taiwan and manufacturing in Malaysia. Management says those lower-cost locations allow engineering and operations to scale without expenses increasing proportionally with revenue.

Penguin is also using AI internally. Shaikh said all of its software engineers who perform code generation are using AI tools, and the company has deployed AI across functions including finance and legal.

That does not guarantee 55% EPS growth. But it provides a clearer mechanism for why management believes expenses can grow materially slower than revenue even while the company continues investing in new products.

Cash flow is the part to watch

The income statement is showing substantial operating leverage.

Cash flow is less clean.

According to Penguin's cash flow statement:

• Q4 cash used in operations: $163 million
• Full-year cash used in operations: $152 million
• Full-year capital expenditures: About $12 million

The operating cash outflow was primarily driven by additional working capital needed to support growth in memory and AI infrastructure.

The balance-sheet changes show where the cash went.

Working capital

• Accounts receivable: $796 million
  • Versus $308 million a year ago
• Inventory: $749 million
  • Versus $255 million a year ago
• Accounts payable: $817 million
  • Versus $267 million a year ago

Inventory increased as Penguin bought components for its record backlog and expected customer deployments, while higher memory costs also contributed.

This is an important distinction. Penguin is not burning cash because it is spending billions building its own data centers.

Capital expenditures were only $12 million for the entire year, or well under 1% of revenue. Management describes the model as asset light.

Instead, rapid growth is tying up cash in inventory and receivables before that investment converts into cash from customers.

Stock Unlock's guide to operating cash flow and free cash flow explains how working-capital movements can cause reported earnings and cash generation to diverge sharply for a period.

The next step is seeing whether Penguin can convert its record backlog into revenue while bringing working-capital needs back down. Management explicitly called improving working-capital efficiency a priority for fiscal 2027.

Analysts raised targets, but earnings started the move

Several Wall Street firms raised their price targets Wednesday morning.

The changes included:

• Rosenblatt: $100
• Goldman Sachs: $85
• Needham: $85
• Stifel: $85
• Barclays: $60, while maintaining an Underweight rating

The earnings report had already started the move. Penguin shares jumped in after-hours trading Tuesday immediately following the results, before Wednesday morning's analyst notes.

Those target increases likely added some support once regular trading began, but they were reactions to the same fundamental development investors were already responding to.

Penguin entered the report with expectations for rapid AI growth. It left the report with:

• Roughly 40% expected revenue growth
• Roughly 55% expected adjusted EPS growth
• Record company backlog
• Memory backlog extending at least four quarters
• AI infrastructure bookings running ahead of management's prior expectations
• Large neocloud projects that are already underway

The quarterly beat started the rally. The larger reason for the move is that management now has substantially better visibility into the growth behind fiscal 2027 than it had only 90 days ago.