Dan IVES Wedbush AI Power & Infrastructure ETF (IVEP)

Artificial intelligence is shifting from a laboratory curiosity to a defining economic force, and that shift is reshaping capital allocation at every level. Power consumption is exploding as companies train and run machine learning models at scale. Semiconductor design is becoming the domain of specialists. Data centers are transforming from real estate into the essential machinery of computation itself. The companies that own, build, and operate this infrastructure — the picks and shovels in the AI gold rush — are the subject of the Wedbush AI Power & Infrastructure ETF.

IVEP is a thematic, concentrated fund. It does not attempt to capture “AI broadly” — a definition that has become so loose as to be nearly meaningless — but rather focuses deliberately on the physical and computational substrates upon which AI depends. The fund holds semiconductor companies designing chips for AI workloads (NVIDIA’s GPUs, AMD’s chips, Broadcom’s interconnect silicon), the companies manufacturing and assembling those chips, the data-center owners and operators that house the clusters running large language models and image-generation systems, and the software and service providers that optimize, manage, and power AI infrastructure.

This lens on infrastructure reflects how the AI economy is actually being built. The largest economic rents are flowing not equally to all software vendors but concentrating among the companies controlling the silicon, the power, and the processing architecture. A data-center-as-a-service provider or a semiconductor designer has different economics, competitive moats, and cash flows than a software company that licenses AI models. IVEP’s bet is that the infrastructure layer will be where durability and value accumulate.

The fund is explicitly thematic and concentrated. It holds perhaps 30 to 50 stocks globally, not the 500-plus you might find in a broad-market index. That concentration makes it more volatile — when sentiment around AI infrastructure sours, the entire fund can move sharply. But it also means the fund is not diluted by companies that merely use AI internally, which would bloat the eligible universe to include almost any tech or financial service company. The narrow focus is the point.

Power and heat are physical constraints on AI scaling. Training a large language model or a vision-model infrastructure consumes enormous electricity — a single data center can have power requirements measured in tens or hundreds of megawatts. That demand is reshaping utility economics and driving investment in power generation, transmission, and cooling infrastructure. IVEP captures some of that theme through data-center owners and through companies providing the physical infrastructure (power distribution, cooling, backup generation) that AI clusters depend on. As AI workloads grow, the companies providing that infrastructure become more central to the overall system.

The semiconductor angle is where the fiercest competition and the richest margins often sit. Custom chips designed specifically for AI workloads have elbowed aside general-purpose processors in the data-center world. That specialization creates lock-in: a software framework or data pipeline built around a specific chip architecture is costly to migrate away from. The manufacturers and designers of those chips — and the companies that support their creation and deployment — accumulate economic value.

One structural risk is that thematic funds can feel expensive just when they become most popular. When the media and retail investors are most enthusiastic about AI, valuations have often already priced in years of growth. That creates a collision: the fund can underperform sharply if sentiment shifts, even if the underlying technology and infrastructure continue advancing. The reverse is true too — buy it at peak pessimism and you may be positioned for strong returns.

The fund’s concentration is both a feature and a risk. The concentrated approach makes it efficient at tracking the specific AI infrastructure narrative, but it also means the fund can swing sharply on company-specific news. If a major chip designer disappoints or a data-center operator faces regulatory backlash, the entire fund’s price can move decisively.

For investors researching IVEP, the prospectus and fact sheet will detail the current holdings and the exact criteria for inclusion. The fund’s performance over different market cycles — particularly comparing it to tech-heavy indices like the Nasdaq-100 and to pure semiconductor or data-center indices — reveals whether the AI infrastructure thesis is delivering value or if the fund is riding a temporary narrative wave. Understanding the composition of holdings also matters: a fund tilted more toward semiconductor makers will have different volatility and cycle dynamics than one focused primarily on data-center operators.

IVEP is not a core holding for a retirement portfolio. It is a satellite position, an expression of conviction that AI infrastructure will remain a high-growth, high-investment area. It works best for investors with higher risk tolerance, a longer time horizon, and a willingness to see sharp drawdowns as the price of exposure to a theme that could deliver outsized returns over a decade.