Wall Street spent the last three years debating whether artificial intelligence would ever show up as real revenue — not just in NVIDIA’s data center numbers, but across the entire software and infrastructure stack that enterprise AI actually runs on.
This week, we get the answers.
With US markets closed Monday for Memorial Day, the first trading day back is Tuesday May 26 — and from that moment through Thursday’s close, a parade of AI-adjacent earnings reports will land that collectively form the most complete picture yet of whether the AI investment cycle is translating into dollars. Zscaler kicks it off Tuesday night. Wednesday brings a triple header: Salesforce, Marvell, and Snowflake all report after the close in a single session. Dell opens Thursday morning with a server backlog that would have seemed like science fiction five years ago. And on Friday, the government drops both the revised Q1 GDP estimate and the Fed’s preferred inflation gauge — a macro one-two punch that could reshape rate expectations heading into June.
If you want to understand where markets are going for the rest of 2026, this is the week to pay attention.
The Backdrop: Futures Up, Oil Down, One Big Geopolitical Wild Card
Before we get to earnings, the context matters.
Heading into the holiday weekend, the market got a jolt: US and Iranian officials reportedly agreed to broad principles of a deal that would reopen the Strait of Hormuz — the waterway that handles roughly 20% of global oil and LNG supplies and has been partially blockaded since the conflict began in February. Brent crude dropped 5.2% on the news to around $98 per barrel. West Texas Intermediate fell to about $92.
For consumers, that’s a potential gas price reprieve after months of energy-driven inflation. For the Fed, it complicates the picture — energy was a significant driver of the Q1 PCE surge to 4.5%, well above the 2% target, and a Hormuz reopening could meaningfully soften that number going forward. We’ll get the first look at April PCE on Friday.
Futures markets liked what they heard. Nasdaq 100 futures were up roughly 0.5%, S&P futures up 0.3% as of Sunday — a geopolitical relief rally heading into what was already going to be a packed week.
Tuesday, May 26: Zscaler Sets the Tone
Zscaler (ZS) — After Close | Revenue est. ~$835M (+23% YoY)
The first earnings print of the week comes from a company most retail investors don’t follow closely enough: Zscaler, the cloud-native cybersecurity platform built on a Zero Trust architecture.
What Zscaler does, in plain terms: it sits between your employees and the internet, inspecting every connection in real time rather than trusting anything on your network by default. As enterprises have moved to the cloud and distributed workforces, this model has become foundational — and as AI workloads add new attack surfaces, the demand for AI-native security is accelerating.
Wall Street is pricing in a big move. The options market has implied a ±11.3% swing on the earnings print, which means traders are bracing for volatility in either direction. Analysts are looking for revenue of approximately $835 million, representing 23% year-over-year growth, alongside guidance that reflects strong large-deal momentum and pipeline conversion.
The narrative driving Zscaler beyond the headline number is its AI Security positioning — protecting enterprise environments specifically from AI-enabled threats and governing how employees use AI tools. If Zscaler raises guidance meaningfully Tuesday night, it signals that enterprise IT budgets are holding up and security is getting a premium allocation.
Why it matters for the week: Zscaler is the canary. A strong print suggests enterprises are still spending aggressively on infrastructure — which is exactly the environment Salesforce, Snowflake, and MongoDB need to have good weeks too.
Wednesday, May 27: The Triple Header That Answers Everything
Three companies. One evening. One question.
Salesforce (CRM) — After Close | Revenue est. ~$11.06B (+12% YoY)
Salesforce has spent the better part of two years building toward a single bet: Agentforce, its AI agent platform that automates tasks like customer support, sales follow-up, and data entry directly within the CRM system. The pitch is that AI agents do the work employees used to do manually — and Salesforce charges per agent, creating a new revenue stream on top of its traditional seat-based subscriptions.
Wednesday’s Q1 FY2027 results will be the most comprehensive look yet at whether that bet is paying off.
Analysts expect revenue of approximately $11.06 billion, growth of around 12-13% year-over-year. Current Remaining Performance Obligation — the backlog of contracts already signed — is expected to grow around 14%, which signals forward momentum even if the current quarter’s top line is modest. Multiple analysts maintain Buy ratings with an average price target around $268, and the company issued confident guidance of $11.03–$11.08 billion entering the quarter.
But the number Wall Street is really hunting for isn’t revenue — it’s Agentforce adoption metrics. How many enterprise customers have actually deployed AI agents? What’s the attach rate on new contracts? Is Agentforce pulling in new logos or just cross-selling to existing customers? Those are the questions that will determine whether CRM trades up or down Wednesday night.
The broader significance: Salesforce is the clearest test of whether AI is generating incremental enterprise software revenue, or whether it’s mostly a marketing story that hasn’t hit the income statement yet.
Marvell Technology (MRVL) — After Close | Revenue est. ~$2.40B
This is the one most retail investors don’t have on their radar — and it might be the most revealing print of the week.
Marvell doesn’t make the GPUs. It doesn’t run the hyperscaler data centers. What Marvell does is design the custom AI chips — called ASICs — that companies like Amazon, Google, Microsoft, and Meta build into their own AI infrastructure instead of buying NVIDIA’s off-the-shelf products. When a hyperscaler wants a chip purpose-built for their specific AI workload rather than a general-purpose GPU, Marvell is one of the two companies they call (the other being Broadcom).
The scale of this business is becoming clear. Marvell recently disclosed more than 50 custom AI silicon design wins across over 10 hyperscaler customers. In March, NVIDIA — not exactly a company that needs partners — announced a $2 billion investment in Marvell and a strategic integration via NVLink Fusion, connecting Marvell’s custom ASICs into NVIDIA’s AI factory ecosystem. That’s a validation worth paying attention to.
Management guided Q1 FY2027 revenue to approximately $2.40 billion (plus or minus 5%), with non-GAAP earnings per share of around $0.79. High-volume production has already begun with two major hyperscale AI customers this year.
Marvell’s report answers a different question than Salesforce does. It’s not about whether enterprises are buying AI software — it’s about whether the hyperscalers are still accelerating their AI buildouts at the infrastructure layer. If Marvell beats and raises, the AI capex supercycle is intact. If it misses or guides cautiously, that matters for the whole stack above it.
Snowflake (SNOW) — After Close | Revenue est. ~$1.32B (+27% YoY)
Snowflake is the data layer. Before an enterprise can train an AI model or run an AI agent, it needs to store, manage, and query enormous amounts of structured and unstructured data — and Snowflake has built the dominant cloud platform for doing exactly that.
The investment thesis is straightforward: more AI adoption means more data generated and queried, which means more consumption revenue for Snowflake. The company charges based on how much compute customers actually use (rather than seat licenses), so its revenue tracks closely to actual AI workload activity.
Analysts are looking for revenue of approximately $1.32 billion, up 27% year-over-year — an acceleration from prior quarters that would signal AI workloads are genuinely driving incremental consumption. Under CEO Sridhar Ramaswamy, who took over in 2024, Snowflake has been pivoting aggressively toward AI-native capabilities, including a suite of tools that let enterprises build and run AI applications directly on top of their Snowflake data.
Snowflake is also one of the cleaner barometers of enterprise cloud spending broadly — not just AI. A strong print here confirms that the enterprise cloud cycle is healthy heading into the second half of 2026.
Thursday, May 28: Dell Makes It Physical
Dell Technologies (DELL) — Before Open | Revenue est. ~$35B (+51% YoY)
After three days of software and silicon, Dell brings it back to earth — literally. Dell builds and ships the physical servers that house the GPUs, the custom ASICs, and the memory that AI models run on. And right now, it cannot build them fast enough.
Dell enters Thursday’s report with an AI server backlog of approximately $43 billion — meaning customers have already ordered and paid deposits on $43 billion worth of servers that haven’t been delivered yet. The company has guided its full fiscal year 2027 AI server revenue toward $50 billion. For context, that would represent more AI server revenue in a single year than Dell’s entire company generated in revenue just a few years ago.
The consensus estimate for Q1 FY2027 is approximately $35 billion in total revenue, up 51% year-over-year, with non-GAAP EPS around $3.00 — itself up roughly 94% from the same quarter last year. Dell’s Infrastructure Solutions Group, which includes servers and storage, is expected to show growth of over 100% as hyperscalers and enterprises race to build out AI capacity.
If Dell beats those numbers and raises guidance, it is the single most bullish data point possible for the AI infrastructure supercycle. Unlike the software companies reporting Wednesday, Dell’s numbers are physical — they reflect actual hardware being manufactured, shipped, and installed. You can’t fake an AI server backlog.
Also Thursday: Costco (COST) and MongoDB (MDB) report after the close. Costco will be read as a consumer health indicator — if the warehouse club giant sees margin compression from tariff-related cost increases, it matters for the broader consumer discretionary picture. MongoDB’s print (~$659-664M guided) will add another data point to the enterprise AI developer spending story.
Friday, May 29: The Macro Wildcard
Two critical government data releases land Friday morning, and either one could override the week’s earnings narrative.
GDP Q1 2026 Second Estimate — The advance estimate showed the US economy contracting in Q1, combined with the PCE price index surging from 2.9% in Q4 2025 to 4.5% in Q1 2026. Core PCE — the Fed’s preferred inflation measure — jumped from 2.7% to 4.3%. Economists attributed the spike to tariff pass-through, energy price pressure from the Middle East conflict, and broader supply-side cost shocks.
Friday’s revision incorporates more complete trade and inventory data. If the inflation numbers are revised down, the stagflation narrative softens and the market gets relief. If they hold or worsen, the Fed’s path to rate cuts gets considerably harder — and a market that’s been counting on cuts later this year will have to reprice.
April PCE — Alongside the GDP revision, the Bureau of Economic Analysis releases April’s Personal Consumption Expenditures data. This is the number the Fed watches most closely. If April PCE begins to show moderation from Q1’s spike — plausibly driven by the early signs of an Iran deal reducing energy pressure — it would be a significant signal that the inflation problem may be peaking rather than accelerating.
Put simply: a week that starts with AI earnings could end with a macro decision that shapes the next six months of Fed policy. Don’t turn off the screen after Thursday.
The Question Underneath All of It
Every company reporting this week — in different ways, from different angles — is answering the same question: has three years of AI investment actually produced a business?
NVIDIA answered its part last week with $81.6 billion in quarterly revenue, up 85% year-over-year, and guidance of $91 billion for the current quarter. The infrastructure layer has clearly worked. The demand for AI compute is real and it’s enormous.
But compute is only the foundation. The question this week is whether the layers above it — the custom silicon routing AI workloads (Marvell), the security protecting AI deployments (Zscaler), the data platforms AI agents query (Snowflake), the software orchestrating AI agents in enterprise workflows (Salesforce), and the servers housing all of it (Dell) — are seeing the same validation in their revenue lines.
If the answer is yes across the board, this week could cement the AI trade as the defining investment theme of the decade. If there are cracks — a guidance cut from Snowflake, a miss from Salesforce’s Agentforce metrics, a cautious outlook from Marvell on hyperscaler design win timing — the market will notice.
Either way, we’ll know a lot more by Friday than we do today.
Follow Ledger Prime News for earnings reaction posts throughout the week. Wednesday night’s triple header (CRM, MRVL, SNOW) will have its own breakdown published after the close.
Sources:
- CNBC: NVIDIA Q1 FY2027 Earnings
- CNBC: Stock Market Next Week Outlook May 25–29
- Fortune: Oil Drops as Iran and Hormuz Reopening Nears
- Salesforce IR: Q1 FY27 Earnings Webcast
- 247 Wall St: Marvell’s AI Tailwinds Are Heating Up
- HeyGoTrade: Marvell Custom AI Silicon
- Stock Titan: Zscaler Q3 FY2026 Earnings Call
- Tickeron: Top Earnings Week Preview
- Investing.com: PCE Inflation and Q1 GDP — Fed’s Next Move
- Wichita Liberty: US GDP Q1 2026 Advance Estimate