CoreWeave, the AI-focused cloud infrastructure company, faces a critical test in early August when it reports second-quarter results. According to Bank of America, the outcome may depend less on demand for AI computing and more on a practical bottleneck: how fast new data center power actually gets switched on.
In a research note, BofA Securities highlighted "active power" — electricity capacity that is energized and ready to run chips — as the key near-term swing factor for CoreWeave. The firm estimates CoreWeave has about 1 gigawatt active today and is targeting 1.7 gigawatts by year-end, implying a significant ramp in the second half of the year.
Why Power Activation Matters
CoreWeave builds and operates data centers packed with graphics processing units (GPUs) used for training and running artificial intelligence models. These facilities consume enormous amounts of electricity, and the company's revenue is directly tied to how many of its chips are running at any given time. If new data center capacity is built but the power isn't yet connected, those chips sit idle and generate no income.
The company's capital expenditure is rising sharply to $34 billion, reflecting the massive investment needed to expand its data center footprint. But spending money on construction is only half the battle — the power must actually be turned on for that investment to start paying off.
This dynamic is not unique to CoreWeave. Across the AI infrastructure space, companies are racing to secure electricity capacity, often facing delays from utility companies, grid connection queues, and regulatory approvals. The gap between announced capacity and active power has become a key metric for investors tracking the sector.
What It Means for Investors
For everyday investors, CoreWeave's situation illustrates a broader theme in the AI boom: building the physical infrastructure to support AI is just as important as developing the software. The company's Q2 results will offer a real-world check on whether the industry can deliver on its ambitious expansion plans.
Bank of America's focus on active power suggests that investors should watch not just revenue and earnings numbers, but also operational metrics like megawatts of energized capacity. If CoreWeave can show steady progress in activating new power, it could signal that the company is on track to meet its year-end target. Delays, on the other hand, could raise questions about the pace of growth.
The broader context is that AI-related stocks have driven a significant portion of market gains this year. As noted in a recent report, chip stocks drove nearly half of S&P 500 Q2 earnings growth amid volatility. CoreWeave sits at the intersection of this trend, providing the cloud infrastructure that powers AI workloads for companies like Microsoft and others.
Looking Ahead
CoreWeave's Q2 earnings report, expected in early August, will be closely watched for updates on power activation timelines, customer demand, and the company's path to profitability. The $34 billion capex plan underscores the scale of the bet the company is making on AI's future, but the near-term focus will be on execution.
Bank of America's note also highlights the importance of the second-half ramp. If CoreWeave can accelerate power activation in the coming months, it could boost revenue and reassure investors that the massive spending is justified. If not, the company may face pressure to explain the gap between investment and returns.
For investors, the key takeaway is that in the world of AI infrastructure, turning on the lights is just as important as building the building. CoreWeave's next test is turning power into profit.
