Vertiv sits at the unavoidable bottleneck of AI

Before any server rack can power on — before any Nvidia GPU can crunch a single token — someone has to build the physical infrastructure to support it. Power management systems, uninterruptible power supplies, thermal management, precision cooling, and liquid-cooling distribution units must all be in place first. Vertiv designs, manufactures, and services exactly that equipment, and roughly 75% of its revenue comes from data center customers.

This is not a software bet or a chip bet. It is a bet on the physical layer of computing itself, which is non-negotiable. Every new AI infrastructure build needs Vertiv’s products before anything else can happen.

Micron’s HBM memory chips are also essential to AI training, but they sit one layer up in the stack, where they face intense competition and price cyclicality. Vertiv’s position is structurally stickier. Once a company’s power and cooling architecture is embedded in a data center, replacing it is enormously expensive and disruptive.

The $15 billion backlog creates unmatched revenue visibility

Vertiv’s project backlog more than doubled in 2025 to over $15 billion. That figure represents approximately 12 to 18 months of forward revenue and is composed of large, complex infrastructure projects with substantial advance payments — $1.8 billion in deferred revenue, up 71% year over year. These are not soft order intentions. Customers have committed capital because they have no choice: they cannot build without Vertiv’s systems.

Micron has no equivalent revenue shield. Memory is priced on spot and contract markets that shift every quarter. When the memory cycle turns — and it always does — Micron’s revenue can fall 40–50% in a matter of quarters, as it did in 2022 and 2023. Vertiv’s backlog insulates it from that kind of sudden shock.

Secular tailwinds that don’t depend on memory prices

Vertiv’s growth does not require memory prices to hold up. It requires that hyperscalers keep building data centers, and the evidence for that is overwhelming. Alphabet, Amazon, Meta, and Microsoft have collectively guided to more than $650 billion in capital expenditure for 2026 alone. That spending must flow through power and cooling infrastructure before it can reach servers, chips, or memory.

What makes this particularly durable is the shift to liquid cooling. As AI workloads drive higher rack power densities — with modern AI racks requiring 50 to 100 kilowatts versus the 5 to 15 kilowatts of a standard rack — air cooling simply cannot keep pace. Liquid cooling is being adopted as the primary thermal solution for the most power-intensive compute rows, and Vertiv is the largest global player in that transition.

Micron is a cyclical asset approaching peak earnings

The single most important thing to understand about Micron is that it is a cyclical commodity business. Its history is one of violent boom-bust cycles. Gross margins collapsed below zero in 2023. Then they roared back as supply tightened and HBM demand exploded. Now, analysts tracking the memory market are already discussing peak-cycle risk: capacity expansions planned for 2027 and 2028 by Samsung, SK Hynix, and Micron itself could trigger a supply glut that crushes pricing.

MU’s earnings and margins are likely to peak in 2026 or early 2027, with a probable downturn thereafter as the memory cycle turns. Samsung’s advances in HBM4 and its history of aggressive pricing add competitive pressure. Buying Micron today, after a 700%-plus run from its 2023 lows, means buying at or near peak earnings — historically the worst entry point in a cyclical. Vertiv does not carry this risk profile.

Margin expansion has a clear, structural path

Vertiv’s adjusted operating margins expanded 430 basis points year over year to 20.8% in Q1 2026, and management is guiding for margins to reach 25% by 2029. This is not a one-time event or a lucky pricing environment. It reflects the structural shift in Vertiv’s mix toward larger, more complex, higher-margin projects — incremental margins on these projects exceed 30%.

Micron’s margin expansion is real but fragile. It is almost entirely dependent on HBM mix and the current pricing environment. Gross margins of 75–81% sound extraordinary, but they reflect peak-cycle pricing on a commodity input. The moment pricing softens — which it will, eventually — margins compress dramatically, as they did in prior cycles.

Strategic partnerships with Nvidia create a technological moat

Vertiv’s collaboration with Nvidia gives it early visibility into the thermal and power specifications of future GPU generations before those chips are publicly announced. This means Vertiv is co-designing the cooling and power distribution architectures that will support Nvidia’s next-generation AI accelerators, not scrambling to retrofit existing products after the fact.

No equivalent relationship exists for Micron. HBM memory is competitive across three major vendors, and while Micron has made significant progress, SK Hynix remains the dominant HBM supplier to Nvidia. Micron is fighting for share in a market where its largest customer has a pre-existing, deeply integrated relationship with a rival.

Revenue growth is accelerating, not decelerating

Vertiv posted 30% organic revenue growth in Q1 2026 — on top of already elevated 2025 comparables — and raised full-year guidance to $13.25–$13.75 billion in revenue with 51% earnings growth. Full-year 2025 revenue was $10.23 billion, up nearly 28% from $8.01 billion in 2024. Earnings grew 169% in 2025.

Micron’s revenue growth is extraordinary right now — Q3 2026 guidance implies $33.5 billion. But analysts forecast revenue growth slowing to 22% annually over the next three years as the cycle matures. The base effect becomes unfavorable as 2026 numbers embed the peak pricing environment. Vertiv’s growth is structural; Micron’s is cyclical.

S&P 500 inclusion signals institutional durability

In March 2026, Vertiv was added to the S&P 500, cementing its status as a large-cap industrial infrastructure company, not just an AI trade. This matters for two reasons. First, it forces passive index funds to own VRT, creating a permanent structural buyer. Second, it signals that the investment community views Vertiv as a durable, multi-decade business — not a flash-in-the-pan beneficiary of a single cycle.

Micron’s cyclical nature means institutional investors treat it as a trading position, not a core holding — adding and reducing exposure based on where they think the memory cycle is heading. That dynamic creates more volatility and less support during downturns.

The services business creates a recurring revenue floor

A portion of Vertiv’s revenue that gets insufficient attention is its lifecycle services segment — preventative maintenance, acceptance testing, fluid management, remote monitoring, spare parts, and performance assessments. These services create a predictable, recurring revenue stream tied to the installed base of Vertiv’s equipment worldwide. As the installed base grows with each new data center build, the services business grows automatically without requiring new customer acquisition.

Micron sells commodity products in one-time transactions at prices set by the market. When chip prices fall, every dollar of Micron’s revenue shrinks in real time. Vertiv’s three-layer earnings structure — project revenue, backlog, and services — is substantially more resilient than Micron’s two-variable business of volume × price.

Valuation is justified by growth quality, not speculation

VRT trades at roughly 53 times forward earnings, which sounds rich until you examine the quality of those earnings: 30% organic revenue growth, 51% earnings growth guidance, 25% margin targets, and a $15 billion backlog. The PEG ratio — price-to-earnings divided by growth rate — sits around 1.07, which is actually cheaper than many S&P 500 peers on a growth-adjusted basis.

Micron trades at just 8–12 times forward earnings, which looks cheap until you remember that forward earnings estimates embed peak-cycle assumptions. When the cycle turns, those earnings can collapse 60–80% in two to three quarters, and the stock follows. A company trading at 10x peak earnings that reverts to trough earnings is not cheap — it is an earnings trap. Vertiv’s premium multiple is defensible. Micron’s “cheap” valuation is an artifact of where we are in the memory cycle, not a signal of fundamental value.