Dell's AI Backlog Hits $95B as Server Sales Blow Past Estimates

Dell just posted a quarter that makes six-month-old AI forecasts look almost quaint. The company reported fiscal second-quarter results Tuesday that blew past every estimate on the Street, and the guidance hike that followed tells you Wall Street is still underpricing how fast AI infrastructure spending is moving.

Adjusted earnings came in at $7.04 a share, more than 40% above the $4.92 analysts had penciled in. Revenue jumped 58% year-over-year to $46.97 billion, topping the $44.92 billion consensus. Net profit nearly quadrupled to $4.13 billion from $1.16 billion a year ago — that's the kind of margin expansion investors want to see alongside top-line growth, not just a revenue headline.

The engine behind it all is Dell's Infrastructure Solutions Group, which pulled in $31.78 billion in revenue, up 89% from last year. AI-optimized servers alone accounted for $16.4 billion of that, ahead of the Street's $16.07 billion forecast. Traditional servers and networking gear, often overlooked in the AI story, jumped 122% to $10.53 billion as companies upgrade CPU capacity to support AI workloads riding on top of GPUs.

Storage revenue climbed 26% to $4.85 billion, and ISG operating income surged 225% year-over-year to $4.78 billion, with margins expanding 620 basis points to 15%. Dell now carries $95 billion in AI-related backlog, and management says its pipeline of future orders is a multiple of that number. As Forrester Research analyst Naveen Chhabra put it in comments following the report, the shift is bigger than one company's product cycle.

"Dell's results suggest that enterprise AI investment is moving beyond experimentation and driving a broader infrastructure modernization cycle." — Naveen Chhabra, Analyst, Forrester Research

That modernization angle matters for anyone holding or considering DELL. Six months ago, management was guiding for AI server growth of just 13% for the fiscal year. Now the forecast calls for more than 200% growth, with AI-optimized server sales projected to reach $74 billion in fiscal 2026 — a stunning revision in a short window.

The full-year guidance bump reflects that shift. Dell raised its outlook to $25.50 in earnings per share on $192 billion in sales, up from a prior forecast of $17.90 and $165 billion. For comparison, Wall Street had only been modeling $18.92 in EPS and $172.67 billion in revenue for the year, as Yahoo Finance's coverage of the earnings noted — Dell just leapfrogged its own analysts' expectations by a wide margin.

Here's the catch for anyone chasing the stock now: DELL shares are already up roughly 236% year-to-date, dwarfing the S&P 500's 11% gain over the same stretch. That kind of run means a lot of good news is already priced in, and the stock actually fell more than 6% during Wednesday's regular session before clawing back the losses and popping higher after hours on the earnings beat.

Rising component costs are also squeezing margins even as sales surge — Dell has had to raise prices to offset those pressures, and investors should watch whether that trend eats into profitability in future quarters. Supply constraints on AI hardware remain a real risk to execution, not just a talking point, and any slowdown in enterprise capex could hit a stock trading on such lofty multiples harder than most.

For investors looking at an entry point, Dell's next earnings check comes with third-quarter guidance of $6.50 in EPS on roughly $49 billion in sales — a bar management just set for itself. Watch whether AI server orders keep growing at triple-digit rates, and keep an eye on the installed base of roughly 1.2 million traditional servers still due for a refresh cycle, since that's a demand tailwind that doesn't depend on the AI trade holding up. Dell's AI customer base has also grown past 6,500 accounts spanning neocloud, sovereign, and enterprise buyers — a diversification signal worth tracking against any single-customer concentration risk.

Bottom line: Dell isn't just riding the AI wave anymore — it's compounding it across both AI and traditional hardware, and the backlog gives visibility that few other AI-adjacent names can match right now.