원문 보기
There's more data centers under construction today than active data centers. So, when you think about, you know, for the hyperscalers, they're going to need to ultimately pay their own bill. I think it's a smart move. You know, and especially the last thing you want is data centers being built in an area, in a community, and that becomes a huge negative given electricity prices going to go up for everyone else.
So, I think this was a smart move by administration. White House did that was a step in the right direction. You know, making sure these big tech players, you know, ultimately pay their own bill. We're re's more data centers under construction today than active data centers. You need to have the right guardrails in place from electricity to safeguards to everything else for these big tech players.
Otherwise, they become bull's-eye in their back within communities. There are more data centers under construction today than active data centers on the ground. When Wedbush's Dan Ives dropped that bombshell, it wasn't just a statistic. It was a warning to every investor still stuck in the 2024 mindset.
We have officially entered the most critical phase of the AI revolution, the prove-it year of 2026. While the mainstream media is obsessed with the daily fluctuations of chip makers, a far more structural and permanent shift is happening in the physical world. We are witnessing a massive $650 billion AI-driven capital expenditure wave that is literally outstripping the capacity of the global power grid.
This has turned basic infrastructure, electricity, cooling, and physical space into the most valuable assets of the decade. The game changed forever on March 4th, 2026, when the White House orchestrated a historic policy shift, the ratepayer protection pledge. By mandating that hyperscalers like Microsoft, Google, and Amazon build, bring, or buy their own power generation, the administration essentially deregulated a trillion-dollar surge in private infrastructure.
This move removed the local political bull's-eye from Big Tech's back and cleared the bottleneck for massive self-funded utility projects. For investors, this is the ultimate green light. We are no longer just betting on software margins. We are betting on the picks and shovels that make the AI cloud physically possible.
To find the multi-baggers of this cycle, you have to look at the firms providing the high-voltage grids, the specialized thermal management, and the high-speed networking fabric. Dan Ives and the Wedbush team believe are the real beneficiaries of this unprecedented spending spree. Let's begin with the powerhouse of the physical AI layer, Vertiv Holdings, ticker VRT.
As Dan Ives notes, the AI revolution is moving into a phase where monetization must be met with physical execution, and Vertiv is the company providing the mission-critical systems that keep those high-performance chips from melting. Looking at their performance through early 2026, the data is staggering.
VRT has surged over 180% in the past year, proving it is the undisputed bottleneck for the global data center buildout. The company just capped off 2025 with a blowout fourth quarter, reporting record organic sales growth of 23% and a massive consolidated top line of $2.88 billion. But the real number that should catch any investor's eye is their order book.
In Q4 alone, Vertiv saw organic orders explode by over 250% year over year. This momentum has pushed their total backlog to a whopping $15 billion, more than double where it sat just a year ago. This isn't just a short-term spike, it's a deep multi-year reservoir of revenue visibility. What's driving this?
It comes down to Vertiv's dominant business model in thermal and power management. As AI racks move toward densities of 100 kilowatts or more, traditional air cooling is no longer physically viable. Vertiv is the clear leader in the shift to liquid cooling, recently reinforced by their strategic acquisition of Perdrix to bolster their industrial fluid management capabilities.
They aren't just selling fans, they are selling integrated 800-volt DC power architectures and prefabricated one-core modular systems that can reduce data center deployment times by up to 50%. Looking ahead, the catalysts are lining up perfectly. Management has already crushed expectations with guidance for 28% organic growth and a projected 43% jump in adjusted EPS for the full year.
Furthermore, their bring-your-own power and cooling initiative allows hyperscalers to bypass grid constraints by deploying complete self-funded infrastructure blocks, directly addressing the biggest hurdle to AI expansion. With a net leverage ratio of only 0.5 times, Vertiv has the clean balance sheet needed to keep scaling alongside the biggest names in tech.
While Vertiv cools the servers, the second stock on our list, GE Vernova, ticker GEV, is the firm that ensures those servers actually have the gigawatts required to boot up. Dan Ives has pointed out that the new White House mandate for hyperscalers to pay their own energy bills is a massive catalyst for GEV, as it effectively transforms the company from an industrial manufacturer into a single-seller infrastructure-as-a-service powerhouse.
While most of Wall Street still misprices this as a cyclical play, the underlying data suggests a much more profound transformation. As of early 2026, GE Vernova is sitting on a massive $150 billion backlog, driven by an insatiable global demand for electricity to fuel the AI revolution. The core of this bullish thesis lies in their slot reservation agreements, or SRAs.
GE Vernova currently holds 43 gigawatts of these reservations, which carry pricing premiums 10% to 20% higher than their firm backlog. Essentially, they have created a zero-cost capital flywheel by leveraging $8 billion in customer progress collections to fund their expansion. This strategy has allowed them to scale their heavy-duty gas turbine production toward a target of 24 gigawatts by 2028 without taking on traditional debt burdens.
The company's technological moat is equally impressive. Their 7HA.03 gas turbine operates at a record-breaking 64% net efficiency, pushing the absolute thermodynamic limits of power generation. But the real high-margin alpha is expected to come from their electrification segment and the GridOS substrate.
By launching GridOS as an AI-ready hardware abstraction layer for the electric grid, they are shifting from selling one-off transformers to a high-margin sticky SaaS ecosystem. This transition is being accelerated by the $5.3 billion vertical integration of Prolec GE, which secures their control over the transformer supply chain, the ultimate choke point in modernizing the grid.
Looking toward 2028, the financial potential is immense. Management is targeting a 20% adjusted EBITDA margin across its power and electrification segments, with cumulative free cash flow scaling to $24 billion. Between their leadership in the current gas-powered bridge power era and their leading position in small modular reactors for the zero-carbon future, they are effectively hedging the entire energy transition.
With a net leverage ratio below one times and an aggressive share buyback program, GEV is perfectly built to capture the massive capital outlay coming from the world's largest cloud providers. Once the power is secured and the cooling is in place, the third stock on our list, Arista Networks, ticker ANET, provides the high-speed networking fabric that allows the entire AI factory to communicate as a single organism.
While Nvidia provides the brains, Arista provides the nervous system. As we move into the second half of 2026, the company is transitioning to what its leadership calls Arista 2.0, having just hit a historic milestone of shipping over 150 million cumulative ports. Arista reported a record-breaking fiscal year 2025 with revenue growing 28.6% to surpass $9 billion.
Looking ahead, the company has significantly raised its growth targets, now projecting roughly $11.25 billion in total revenue. This optimism is fueled by an 83% surge in AI-related networking revenue alone, which is expected to hit $3.25 billion this year as hyperscalers unleash their estimated $666 billion in capital outlays.
Arista's cloud titan segment, including giants like Microsoft and Meta, now accounts for nearly half of its annual revenue. The company's primary edge remains its unified software stack, specifically the Arista extensible operating system. Unlike competitors with fragmented systems, every Arista product runs on a single binary image, which simplifies management and self-healing for massive AI clusters.
Its newest flagship, the 7800R4 series, delivers an industry-leading 460 terabits per per of system throughput and is specifically designed to minimize job completion time for large-scale AI training. Additionally, Arista is expanding into agentic AI with its autonomous virtual assistant, which uses AI agents to troubleshoot and secure networks in real time.
While surges in memory chip costs have created a challenging supply environment, Arista has strategically increased its purchase commitments to $6.8 billion to secure its supply chain. Despite these headwinds, management has maintained a robust margin outlook supported by a healthy $10.74 billion cash position and zero debt.
As the industry moves toward 800G and 1.6T speeds, Arista is perfectly positioned to capture the massive shift from traditional silos to the high-performance AI factories of the future.
Comments 0