저자: 미상 · 외국인
요약 보기
AI 병목이 칩에서 전력·에너지·데이터센터 인프라로 이동했다는 주장. 에너지 부족이 AI 확장의 최대 제약이며 데이터센터 건설이 미국 AI 경쟁의 핵심이라고 봄. GE Vernova(2026 2분기 매출 111억달러, 수주잔고 1700억달러 이상, 연간 매출 가이던스 455~465억달러 상향), Vertiv(매출 32.7억달러 +24%, 조정 EPS 1.52달러, 2026 EPS 가이던스 6.65~6.75달러), IREN(AI 계약 28억달러, ARR 목표 40억달러 이상 상향)을 유망 종목으로 제시. 다만 주가 변동성과 리스크를 언급함.
다듬은 글
Memory is the biggest constraint we see, and equilibrium will likely not come until 2028-2029. Then can we argue about the shares or what exactly they reflect? Is this cyclical? Has this cyclicality changed the situation ? Energy is by far the biggest constraint. And I believe that the main obstacle is the need to build data centers in the US to implement all of this.
There is no doubt here. I think the biggest risk in US-China relations is data centers. Because there is no debate here: we are winning in the field of chips, models, hyperscalers, and intellectual property. China is winning in robotics and energy. But if data centers aren't built here, it will be like a hospital without surgeons or without the technology needed to operate.
So, that remains the main conclusion. Wall Street's top AI optimist just made a big announcement that most investors are still ignoring. Dan Ives, the Wedbush analyst who has been a vocal advocate of artificial intelligence for years, no longer talks primarily about chips. He talks about power, energy, and the physical infrastructure that actually supports the AI revolution.
According to Ives, energy shortages are currently the biggest limitation to all AI development. Today, more data centers are being built than are currently operating. The power grid simply can't keep up. He even launched a specialized ETF focused on companies solving this very problem. While everyone else is chasing the same semiconductor giants, the real money is quietly flowing to the companies that generate electricity, cool servers, and build the power grids that AI cannot survive without.
Today, I'll break down three stocks that are right on the path of this bottleneck, with their latest 2026 financials , orders, forecasts, and specific reasons why these companies could define the next phase of the AI market. Promotion number one: GE Vernova, the foundation of energy and electrification.
GE Vernova is a specialized energy and electrical engineering company, spun off from General Electric, which is at the heart of the energy bottleneck that Ives constantly talks about. In the second quarter of 2026, the company generated revenue of $11.1 billion, up 22% year-over-year and above expectations.
Orders increased dramatically. One of the most recent quarters showed an 88% increase in orders to over $24 billion with an orders- to-shipments ratio of over 2x. The total order book has ballooned to over $170 billion, providing exceptional long-term visibility. Management raised its annual revenue forecast for 2026 to a range of $45.5 billion to $46.5 billion.
The energy and electrification segments continue to show rapid growth driven by demand from data centers and power grids. Even though the wind energy business remains a deterrent. What makes GE Vernova attractive is its position in gas turbines, network equipment, electrification solutions, and the broader power generation stack that data centers critically need.
When hyperscalers sign 20-year power purchase agreements or require dedicated generating capacity, GE Vernova's technology and production scale often become a topic of discussion. The company is no longer a sleepy industrial giant. It has become a key factor in building infrastructure for artificial intelligence.
Shares have been volatile as investors analyze mixed results in wind power, but order momentum and elevated forecasts suggest accelerating demand related to focused on the physical AI base, this is one of the cleanest ways to capitalize on the " bottleneck" I'm talking about. Promotion number two, Vertiv—a provider of cooling and power management solutions.
If energy is the bottleneck, then cooling and power management within the data center are immediate points of containment. This is exactly the area where Vertiv works. A few days ago, Vertiv released its second quarter 2026 results, which demonstrated both the company's strength and high market expectations.
Net sales reached $3.274 billion, up 24% year-on-year . Adjusted diluted earnings per share (EPS) rose 60% to $1.52, beating estimates. Adjusted operating profit increased 51% and adjusted operating margin increased 410 basis points to 22.6%; Operating cash flow reached $1.1 billion , and adjusted free cash flow increased to $925 million —more than triple the year- ago quarter .
Management raised its forecasts for the full year 2026 across all metrics. Net sales are now expected to be in the range of $13.8 billion to $ 14.2 billion, with organic growth of about 31% on average, and adjusted diluted earnings per share are forecast at $6.65 to $6.75, representing approximately 60% growth.
CEO Giordano Albertazzi pointed directly to the increasing complexity of implementing AI as a key driver of demand. Liquid cooling, high-density power systems, and integrated thermal management are no longer just desirable options. These are the requirements for the next generation of AI clusters. Vertiv has spent years positioning itself as a premier supplier of these mission-critical systems.
The recent revenue miss from the highest market expectations triggered a sharp sell-off, but underlying growth, margin expansion and upgraded guidance remain strong. In a world where every additional megawatt of AI computing requires sophisticated cooling and power distribution, Vertiv is one of the cleanest investments in this need.
Action number three, IREN—an investment in AI power based on renewable energy sources. For higher growth and more specialized impact on the same topic, IREN transformed from a bitcoin miner to an operator of cloud storage and data centers for AI powered by renewable energy. Next, I highlighted companies in the “neo-cloud” category as part of broader infrastructure opportunities.
In July 2026, the company announced new multi-year contracts with leading AI developers worth $2.8 billion . This news prompted management to raise its annual recurring revenue target for 2026 to more than $4 billion. Since approximately 85% are already contracted, and customer prepayments cover a significant portion of the costs of GPUs.
Iron plans to provide 480 megawatts of power for AI this year and scale to 1.2 gigawatts by 2027. The company has significant cash on its balance sheet and has secured significant GPU financing to support the expansion. Previous quarterly results still reflected some of the chaos of the transition period.
Revenue and profitability have been volatile as Bitcoin mining capacity is redirected to more expensive AI tasks, but the trend is clear. Iron secures long-term AI contracts, fixes renewable energy tariffs, and rapidly scales up capacity. In a market where energy availability is a major constraint, operators who control cheap renewable energy and can quickly convert it into AI power have a structural advantage.
The shares remain volatile and carry higher risks than larger industrial companies . However, the pace of contracting and the increased ARR target indicate that the turn towards AI is gaining real momentum. Dan Ives has been right about investing in AI for years. When he says the " bottleneck" has shifted from chips to energy and infrastructure, investors should listen.
Energy shortage is not a distant theoretical problem. It is already determining which data centers can start operating, how quickly hyperscalers will scale, and which companies will capture the next wave of capital spending. GE Vernova provides the foundation for large-scale power generation and electrification with a huge order backlog and growing projections.
Vertiv delivers mission-critical cooling and power management systems directly into data centers with margin acceleration and strong free cash flow. Iron is a specialist player in the renewable energy market, signing multi-billion dollar AI contracts and rapidly increasing its revenue targets.
Comments 0