Everpure, Lumentum, and Vertiv Ride Real Data Center Spending

Google News surfaced three AI infrastructure stocks tied to data center demand: Everpure, Lumentum, and Vertiv. The screen promises something more substantial than another speculative AI basket. Each company sells infrastructure that operators need after buying processors.

However, a place in the same Google News story does not make these businesses interchangeable. Everpure manages and stores data. Lumentum supplies optical components that move it. Vertiv delivers the power and cooling that keep dense computing systems operational.

That distinction creates the central tension. Data center spending is real, but revenue reaches each supplier through different projects, customers, and deployment schedules. Official results support the demand story while exposing risks that a three-stock screen cannot resolve.

Investors also need to separate the original publisher’s screening methodology from primary evidence. Simply Wall St describes the companies as businesses converting AI demand into financial results. Its stock screen is a useful starting point, not a substitute for filings.

The stronger conclusion is narrower. Everpure, Lumentum, and Vertiv have measurable exposure to expanding data centers. Whether that exposure produces durable returns depends on margins, customer concentration, delivery capacity, and the pace of AI deployments.

What the Google News Stock Screen Actually Identified

The three selections represent separate bottlenecks inside an AI data center, not three versions of the same investment.

Everpure, formerly Pure Storage, sells flash storage and data management systems. Its products hold training data, model checkpoints, enterprise records, and other information that computing clusters must retrieve quickly.

The company changed its corporate name in February 2026. An SEC filing confirms that Pure Storage became Everpure while initially retaining its existing ticker.

That rebrand matters because it signals a broader ambition. Everpure wants investors and customers to view it as a data management platform, rather than only an array manufacturer.

Lumentum occupies another layer. It makes lasers, optical components, and modules that convert electrical signals into light for high-speed connections.

Modern AI clusters distribute work across thousands of accelerators. Those processors lose value when network links cannot move data with sufficient speed, reliability, or efficiency. Optical connectivity therefore becomes more important as clusters expand.

Vertiv works closer to the facility itself. Its equipment supplies power, thermal management, racks, controls, and services for data centers and communications networks.

High-density computing creates a direct physical constraint. More computation in each rack increases electrical demand and heat output. Operators must redesign power distribution and cooling before deploying many newer systems.

These positions form a useful sequence:

• Everpure supplies the storage and management layer around valuable data.

• Lumentum supplies optical technology that carries data between computing systems.

• Vertiv supplies power and cooling systems that keep those machines available.

The Google News result therefore captured a meaningful theme. AI spending is expanding beyond processor purchases into supporting equipment and software.

Yet the screen also compresses important differences. Storage can serve conventional enterprise workloads alongside AI projects. Optical demand depends heavily on networking architectures and component cycles. Power and cooling orders follow construction schedules that can span several quarters.

The companies also recognize revenue differently. A storage subscription does not behave like an optical component order. Neither resembles a large thermal-management project with installation requirements.

That makes the screen most useful as a supply-chain map. It is less useful as a conclusion about valuation, risk, or future performance.

Investors should also notice the date difference. The source article appeared before some newer quarterly results. Those updates provide better evidence than the market-cap snapshots and forecasts embedded in the original screen.

The next question is therefore not whether data center spending exists. It is whether the reported demand is reaching each company’s financial statements.

Everpure Is Selling the Data Layer, Not the GPUs

Everpure offers the broadest enterprise exposure of the three, but its AI narrative remains mixed with conventional storage demand.

Everpure reported fiscal 2026 revenue of approximately $3.7 billion, representing 16 percent growth. It also recorded its first quarter with revenue above $1 billion during that fiscal year.

Those figures establish a substantial operating business. They do not establish how much revenue came directly from generative AI deployments.

The company’s annual filing describes an integrated storage and data management platform spanning on-premises systems, public clouds, hybrid environments, and edge locations.

This range gives Everpure more than one route into infrastructure budgets. An enterprise might buy its systems for databases, cyber recovery, virtual machines, analytics, or AI development.

That diversity can stabilize demand when a single workload slows. It also makes Everpure’s specific AI exposure harder to measure from consolidated revenue.

Its products address a genuine problem. AI systems need accessible, governed data before accelerators can perform useful work. Training and inference also generate checkpoints, embeddings, logs, and outputs that organizations must store.

Fast storage can reduce periods when expensive processors wait for information. Data management software can also help teams discover, classify, protect, and move information across environments.

Everpure has connected its portfolio to NVIDIA reference designs and expanded its AI-oriented offerings. According to the company, these integrations simplify the assembly of storage for large computing clusters.

Those claims deserve careful wording. A reference architecture can reduce deployment friction, but it does not guarantee customer adoption or a particular level of financial performance.

Everpure’s first-quarter fiscal 2027 results provided further evidence of continued growth. The company also completed its transition from Pure Storage and adopted the new ticker associated with Everpure.

The risk lies partly in the economics of hardware. Flash systems require processors, memory, and storage components whose costs can change quickly. Higher component costs can pressure margins or force difficult negotiations with customers.

Enterprise storage also remains competitive. Dell Technologies, Hewlett Packard Enterprise, IBM, NetApp, and cloud providers offer overlapping approaches to storing and managing data.

AI can increase the total volume of stored information without directing every dollar toward Everpure. Customers can use traditional arrays, object storage, cloud services, or specialized platforms.

The rebrand introduces another execution question. Expanding from storage into broader data management requires new software, sales capabilities, integrations, and customer trust.

That expansion can raise the value of Everpure’s installed base. It can also create complexity if customers continue to view the company mainly as a hardware supplier.

For investors following the Google News theme, Everpure is the data-layer selection. Its reported growth is real, but the precise contribution from AI remains less visible than the headline suggests.

The clearest confirmation would come from sustained hyperscale revenue, expanding subscriptions, and disclosures that quantify AI-related deployments. Without those signals, enterprise storage growth and AI demand remain intertwined.

Lumentum Shows How AI Spending Reaches Optical Networks

Lumentum provides the most direct evidence that larger AI clusters are increasing demand for high-speed optical connectivity.

Lumentum reported fiscal third-quarter 2026 revenue of $808.4 million. That represented 90 percent year-over-year growth and set a company record.

Its GAAP gross margin reached 44.2 percent, while its GAAP operating margin reached 21.6 percent. Management attributed the quarter’s performance to strong demand and better operating leverage.

The company’s quarterly results provide stronger support for the AI infrastructure thesis than a stock screener alone. Revenue growth appeared alongside substantial margin expansion.

Lumentum sells several products into cloud and networking markets. These include externally modulated lasers, pump lasers, narrow-linewidth lasers, optical modules, and components for data center interconnects.

The technical mechanism is straightforward. As clusters add processors, they create more communication between servers, racks, and facilities. Copper connections become less practical over longer distances and at higher data rates.

Optics convert information into light that can travel efficiently through fiber. That makes optical components essential for scaling the network around dense accelerator systems.

Lumentum’s opportunity also extends beyond connections inside one building. Data center interconnect products link separate facilities that operators want to manage as a larger computing environment.

NVIDIA’s newer networking plans reinforce this direction. Its Vera Rubin platform includes co-packaged optics, which place optical functions closer to switching silicon.

According to NVIDIA, its photonics platform targets greater power efficiency and availability than networks using conventional pluggable transceivers.

Those are NVIDIA’s performance claims, not independent test results. Still, the architectural direction matters for suppliers across the optical ecosystem.

Co-packaged optics can expand the market for lasers and related components. It can also change which products capture the most value.

That creates a central risk for Lumentum. A company can benefit from rising optical demand while facing rapid changes in product design, integration, and customer requirements.

Customer concentration adds another concern. Large cloud companies and networking vendors purchase significant volumes, giving major buyers negotiating leverage.

Optical components also have a history of cyclical demand. Customers sometimes order aggressively during capacity shortages, then reduce purchases when inventories become excessive.

Lumentum must expand manufacturing without assuming that every surge continues indefinitely. Too little capacity can leave revenue behind, while too much capacity can burden margins during a correction.

Competition remains active. Coherent, Broadcom, Fabrinet, Marvell, and specialized photonics suppliers participate in overlapping parts of the networking chain.

The company’s recent growth suggests that demand has moved beyond presentations and pilot projects. Customers are ordering components in volumes large enough to alter reported revenue.

However, the durability of that growth depends on deployment schedules. Delayed data centers, changing network designs, or inventory adjustments can move demand between quarters.

Lumentum therefore represents the connectivity layer of the Google News screen. It offers strong financial evidence, alongside significant exposure to component cycles and a concentrated customer base.

The most useful next signal will be whether revenue and margins remain elevated as new manufacturing capacity arrives. Sustained demand would strengthen the case that optical spending is structural, rather than a temporary shortage response.

Vertiv Turns Computing Demand Into Power and Cooling Orders

Vertiv provides the clearest proof that AI spending has reached the physical data center, where heat and electricity cannot be solved with software.

Vertiv reported second-quarter 2026 net sales of $3.274 billion. That was 24 percent higher than the same quarter in 2025.

Operating profit increased 44 percent, while adjusted operating profit rose 51 percent. The company also raised its full-year guidance across several financial measures.

Earlier evidence was equally notable. Vertiv ended 2025 with a backlog of $15 billion, up 109 percent from one year earlier.

Its fourth-quarter 2025 book-to-bill ratio was approximately 2.9. Book-to-bill compares incoming orders with recognized revenue, so a result above one indicates orders exceeded sales.

Vertiv’s second-quarter update also identified supply-chain congestion and complex, multi-phase projects as timing factors.

That disclosure captures both the opportunity and the risk. Customers are planning larger deployments, but those projects become harder to manufacture, deliver, install, and recognize as revenue.

Vertiv sells uninterruptible power supplies, switchgear, power distribution, cooling equipment, racks, controls, and associated services. These systems support nearly every stage between a utility connection and operating servers.

The shift toward rack-scale computing raises the importance of liquid cooling. Air cooling becomes less effective as equipment packs more electrical power into limited floor space.

Liquid systems move heat using coolant circulated near processors or other hot components. They can support denser installations, but they require pumps, heat exchangers, controls, plumbing, and specialized maintenance.

Vertiv participates across that chain. It has also expanded its capabilities through acquisitions and additional manufacturing capacity.

NVIDIA’s DSX reference design connects computing hardware with facility-level power, cooling, controls, and simulation. Vertiv is among the infrastructure companies contributing technology to that ecosystem.

The DSX architecture shows why Vertiv’s role is becoming more strategic. Facility systems must increasingly coordinate with computing equipment.

That does not make every backlog order equally secure. Data center projects can face utility delays, permitting problems, financing changes, construction bottlenecks, or customer redesigns.

Backlog therefore provides visibility, not certainty. Orders can move, and revenue recognition depends on delivery and contract terms.

Large customers can also exert pressure. Hyperscalers and colocation providers buy at scale, compare multiple suppliers, and expect equipment road maps aligned with future processors.

Vertiv competes with Eaton, Schneider Electric, Johnson Controls, Trane Technologies, and other specialists. Some competitors combine electrical equipment, cooling, automation, and building systems in broader portfolios.

Another uncertainty involves data center efficiency. Better cooling and power management can support more processors within a fixed electrical envelope. It can also reduce equipment needs per unit of computing output.

Demand should be judged through both capacity and intensity. More facilities help Vertiv, while denser racks can increase the value of each deployment.

Vertiv offers the strongest physical confirmation in the three-stock group. Its revenue, order growth, backlog, and guidance all show that spending has moved into real projects.

The remaining question concerns execution. Converting large orders into profitable revenue requires factories, suppliers, trained installers, and careful project management across several regions.

Real Demand Does Not Remove Valuation and Execution Risk

Operational growth validates the infrastructure theme, but it does not determine whether expectations embedded in these stocks are reasonable.

The original Google News headline encourages readers to focus on real data center spending. That is a useful correction to AI stories based only on demonstrations or distant forecasts.

Yet “real” demand is not the same as predictable demand. Infrastructure projects involve long planning cycles, concentrated buyers, financing requirements, and dependencies on utilities.

Everpure faces uncertainty about its revenue mix. Enterprise storage demand can grow while direct AI contributions remain difficult to isolate.

Lumentum faces product and inventory cycles. Optical demand can rise quickly, but changes in network architecture can shift spending between modules, components, and integrated systems.

Vertiv faces construction and delivery risk. A strong backlog can still encounter site delays, supply constraints, and changes in customer schedules.

All three companies depend on continued capital spending by cloud providers, enterprises, data center operators, and AI developers. A slowdown would not affect them at the same speed.

Optical component orders can adjust relatively quickly. Facility projects have longer cycles, although cancellations or postponements remain possible. Storage purchases can follow enterprise budget decisions that differ from hyperscale construction.

The companies also occupy different competitive positions. Everpure must distinguish its platform from other storage systems and cloud services.

Lumentum must deliver components that meet rapidly changing specifications. Vertiv must coordinate facility equipment with processor road maps and regional construction standards.

Margins deserve equal attention. Revenue growth that requires costly capacity expansion, expedited components, or unfavorable contracts can create weaker economics than the headline implies.

Investors should distinguish reported GAAP figures from adjusted measures. Adjusted results can clarify recurring operations, but they can also exclude costs that still affect shareholders.

Company guidance deserves cautious treatment as well. Guidance reflects management’s expectations and assumptions at a specific moment. It is not a guaranteed outcome.

The Simply Wall St article also included valuation observations generated from models and analyst forecasts. Such estimates depend on growth rates, discount assumptions, terminal values, and future margins.

Small changes in those assumptions can produce large changes in calculated value. That limitation becomes more important when recent growth rates are unusually high.

The primary opponent in this story is therefore promise versus conversion. AI investment creates demand signals, while each supplier must convert those signals into delivered products, revenue, cash, and durable margins.

The evidence currently supports conversion. Everpure has expanded revenue and entered hyperscale storage. Lumentum has reported record sales and wider margins. Vertiv has converted infrastructure demand into sales, cash flow, and backlog.

Still, none of those outcomes removes the possibility of slower orders, customer concentration, component inflation, or competitive pressure.

Readers should avoid treating the three companies as a diversified substitute for the entire AI economy. Their exposure overlaps because the same large projects can drive spending across every layer.

A delayed campus could affect networking, storage, cooling, and power suppliers at different stages. Shared exposure can create correlated risk even when the products appear different.

The stock screen succeeds at identifying infrastructure beneficiaries. It cannot decide how much future growth the market already expects from each one.

Three Signals to Watch After the Google News Headline

The next phase should be judged through financial conversion, network architecture, and project delivery rather than another round of AI announcements.

The first signal is Everpure’s hyperscale and subscription performance. Investors need evidence that its data platform is gaining workloads beyond traditional enterprise storage replacements.

Watch for continued revenue growth, expanding recurring commitments, and clearer disclosures about hyperscale customers. Those signals would strengthen the claim that AI data requirements are creating an additional growth engine.

Weak subscription expansion or vague AI attribution would reduce confidence. It would suggest that conventional storage upgrades still explain more of the business than the headline implies.

The second signal is Lumentum’s ability to sustain revenue and margins as optical capacity expands. Record growth matters most if customers continue absorbing production without building excessive inventory.

Future earnings should reveal whether cloud demand remains broad across components and modules. Commentary about order visibility, manufacturing utilization, and customer concentration will be especially important.

Co-packaged optics also deserves attention. If deployments move from reference designs into volume systems, suppliers will need to show where they participate and how those products affect margins.

A sharp inventory correction would weaken the structural-growth argument. Continued demand across several product categories would strengthen it.

The third signal is Vertiv’s conversion of backlog into completed projects. Revenue growth should remain accompanied by healthy cash generation and controlled execution costs.

Watch for supply-chain congestion, project phasing, manufacturing expansion, and order cancellations. These details show whether demand is becoming operational capacity or remaining in planning documents.

The company’s progress in liquid cooling provides another practical indicator. New processor platforms require more integrated thermal designs, creating opportunities for equipment and ongoing services.

A growing backlog paired with slower conversion would raise concerns about site readiness or delivery constraints. Continued sales growth and stable margins would support the long-duration infrastructure thesis.

These signals also matter beyond investors. Developers and AI product teams depend on available computing capacity, predictable storage performance, and reliable networks.

Enterprise buyers should care because infrastructure constraints can affect cloud availability, deployment schedules, and service costs. More capacity does not automatically mean immediate access.

Knowledge workers experience the effects indirectly. Faster storage, denser computing, and larger networks can support more responsive AI services, but deployment bottlenecks can delay those improvements.

The Google News result captured a legitimate shift from model announcements toward physical delivery. Infrastructure companies are reporting measurable demand instead of only discussing potential markets.

The sharper question is now about durability. Can Everpure deepen its data-management position, can Lumentum sustain optical growth, and can Vertiv deliver its backlog efficiently?

Follow those three operating signals before treating another AI infrastructure headline as confirmation. Real spending has arrived, but execution will determine which suppliers retain its value.