The ETF Review: This AI Power Fund Is an Industrial Contractor in Disguise

The Utility Disguise

You probably think you bought a power plant. The marketing pitches a world where artificial intelligence demands endless electricity, so you buy the fund to own the electrons. You picture regulated utilities, massive nuclear generators, and steady dividend checks funding your retirement. That picture is completely wrong. Buying this fund expecting a sleepy utility basket is like ordering a salad and getting a bucket of fried chicken. You did not buy the electric grid. You bought an industrial construction boom. The Defiance AI & Power Infrastructure ETF does not make its money by selling kilowatt-hours to server farms. It sells the heavy metal that has to sit in the dirt before a server farm can turn on. It owns the transformers, the cooling pipes, the high-voltage switchgear, and the construction crews who dig the trenches. If Nvidia sells the sports car and the cloud giants buy the gas, this fund owns the company paving the highway. That is a viable business model. It is also an aggressive cyclical bet. If the highway project gets delayed by six months, the paving company gets crushed long before anyone turns off their phone. Treating it as a defensive energy allocation is a dangerous mistake. It behaves like an industrial capex trade on leverage, and it is far more concentrated than the label suggests.

What You Actually Own

Crack open the wrapper and you find eighty-eight holdings. Eighty-eight. That sounds like a comfortable cushion of diversification. In practice, it is steered by a tiny cabal of industrial champions. The top ten names account for roughly 57% of your capital. The top three alone swallow more than a quarter of the entire portfolio. If Quanta Services, Eaton, and GE Vernova have a bad morning, your whole investment takes a beating. Quanta leads at over 9% of the fund, followed by Eaton at 8.5%, GE Vernova near 8.4%, and cooling specialist Vertiv at 6.3%. Now look at the actual sector breakdown. Power grid equipment eats up 44% of the fund. Engineering and construction takes another 12%. Add those two figures together and roughly 56% of your money sits in pure industrial manufacturing and contracting. Electric utilities make up just under 12% of the portfolio. This is an industrial capital expenditure fund wearing a power-station costume. The geographic footprint holds a quirk. Nearly 11% of the fund sits in Ireland. Do not assume you are financing Dublin tech incubators. That number is legal scaffolding. Companies like Eaton and Trane Technologies keep their corporate tax domiciles in Ireland while running heavy industrial operations everywhere else. Over the past twelve months, the fund delivered a modest price decline of roughly 4%. That tame headline number hides a violent tug-of-war underneath. A tiny handful of industrial winners carried the fund. Bloom Energy gained roughly a third, GE Vernova climbed roughly 38%, and Quanta gained roughly 30%. Vertiv added roughly 3%. Combined, those four contributed roughly eight percentage points of performance between them. The rest worked against them. The nuclear half of the thesis hit a wall: Constellation Energy fell more than 25%, while uranium miner Cameco dropped roughly 23%. Nvidia slid roughly 15%. Those laggards chipped away at the leaders. Worse, the long tail of smaller holdings outside the top ten gave back roughly eight percentage points. When the equipment complex peaked in late June, the floor gave way: Vertiv lost roughly 55% from its high while Bloom Energy plunged roughly 55% and GE Vernova sank roughly 38%. The fund suffered a maximum peak-to-trough drawdown of roughly 38%, logging an annualized volatility of 41%. If you wanted a calm utility yield to balance your equity book, you bought the wrong ticker.

How the Sausage Is Made

Defiance does not pick these stocks. The fund runs on the MarketVector US Listed AI and Power Infrastructure Index, and Tidal Investments acts as adviser of record. The index itself only went live in January 2025, which means every flattering multi-year chart in the marketing deck is a synthetic backtest. The fund physically replicates the benchmark, buying the actual shares rather than using swap contracts, and in its first operating stretch portfolio turnover sat at zero percent. Do not let that figure fool you into thinking the portfolio is static. The index divides the universe into four rigid spending buckets. Power generation and grid equipment gets half of the entire pie. Data centers and AI infrastructure take 25%. Construction and engineering gets 12.5%, and electric utilities get the final 12.5%. Inside those buckets, strict capping rules kick in. Single stocks are capped at 8%. Data-center and infrastructure names, along with any company that generates less than half its revenue directly from the theme, face a tighter 4% cap. If the companies weighted above 5% collectively exceed half the portfolio, the index cuts them back to 4.5% in a cascade. This cap structure creates a strange mechanical reality. Cap-weighted indexes let their best performers compound indefinitely. This index does the opposite. Every quarter, it mechanically trims its biggest winners to keep them under the ceiling. Quanta Services and GE Vernova keep getting cut back to that ceiling each quarter. The selection barrier in the industrial buckets is surprisingly forgiving. Grid equipment and construction get no quality or valuation ranking at all. Meet the theme definitions and you are in. The only real guards at the gate are size and trading volume. New entrants need a $300 million market cap and $1 million in daily liquidity. The rules for utilities and AI hardware are much tighter. Utilities only take the top ten companies by float market cap, while hardware takes the top seven. This creates an uneven portfolio: hand-picked mega-cap tech giants paired with unvetted, small-cap electrical contractors. The methodology is also a moving target. The index guide has been revised four times in 2026. The managers added fast-track rules for newly listed initial public offerings and expanded definitions to include speculative neoclouds and GPU leasing firms. An index less than two years old rewriting its rulebook four times suggests the designers are chasing the market in real time. Reconstitution happens every quarter in March, June, September, and December. The next rebalance is announced on December 11 and takes effect on December 18. If you own the fund, those dates dictate your returns.

The Bill

The headline price tag on this fund is an expense ratio of 0.69%. That is a unitary fee, meaning you pay the full freight with no fee waivers in place. Put that percentage into real cash terms. On a $10,000 position, the fund extracts $69 in fees in your first year. Hold that position for ten years, and assuming flat asset values, you will surrender roughly $690 directly to the fund managers. Step up to a serious $50,000 position, and you burn approximately $3,450 across a decade. That money comes straight out of your compounding engine. Compare that to the broader market. A standard, broad-based S&P 500 index fund charges roughly 0.03%. AIPO costs twenty-three times more. Even within its thematic sandbox, the fund looks expensive. The iShares U.S. Power Infrastructure ETF charges just 0.39% for a portfolio that overlaps AIPO by nearly 40%. You are paying a 77% premium over the iShares alternative for Defiance's bespoke weighting system. To its credit, the wrapper trades efficiently: the median thirty-day bid/ask spread sits at 0.04%, which costs you a negligible $4 on a $10,000 order. The share price tracked its net asset value cleanly with a flat 0.00% premium or discount on November 10. Daily volume regularly runs between $37 million and $58 million. The real hidden fee lives in the income statement. You might expect a fund stuffed with grid infrastructure and power assets to throw off healthy cash flow, yet the trailing dividend yield is a microscopic 0.01%. Even worse, the 30-day SEC yield is negative 0.12%. That measure nets the portfolio's dividends against its costs, so a 0.69% fee eats income the holdings barely produce. You are paying up to own a collection of capital-hungry industrial businesses that return almost nothing.

The Scoreboard

On paper, the fund does exactly what it promises. It tracks its underlying index. The one-year performance sits within 0.7 percentage points of the benchmark, which aligns neatly with the expense ratio. The machinery works as designed. Then you look at the competition. Over the twelve months ending in November 2026, the fund slipped roughly 4% on a price basis. Its cheaper rival, the iShares U.S. Power Infrastructure fund, slipped roughly 5%. First Trust Smart Grid managed a gain of roughly 2%. Meanwhile, a dirt-cheap S&P 500 index fund gained roughly 10%. AIPO edged out its closest direct rival by less than a point, then charged 0.30% more for the privilege. A coin flip. Meanwhile, the fund trailed the basic market benchmark by more than fourteen percentage points. You endured a bone-rattling 38% drawdown and 41% volatility just to dramatically underperform a regular market index fund. Since its launch in July 2025 through October 2026, the fund has returned 23.35% cumulatively, which annualizes to nearly 18%. Those numbers look great on a brochure, but remember the timeline. The fund launched right as the market discovered the AI power shortage narrative. Today, that excitement is fully priced into the shares. The fund trades at roughly 27 times trailing earnings. The underlying index shows a price-to-earnings ratio of nearly 35 if you exclude the companies losing money. The S&P 500 trades between 23 and 24 times earnings. Look at forward multiples and the disparity becomes even stranger. The fund trades at roughly 22 times forward earnings against 18 for the S&P 500. Yet consensus estimates project next-year earnings growth for the fund at 28.7%, compared to 27.2% for the broader market. You are paying a 20% premium on valuation to capture an extra 1.5 percentage points of earnings growth. And once you stop ignoring the unprofitable companies in the basket, the index's aggregate trailing multiple drops below zero. A substantial slice of this fund does not make money. You are paying top dollar for speculative turnarounds alongside the industrial titans.

What You Are Really Betting On

When you buy this fund, you are not betting that software developers will create brilliant new AI agents. You are placing a specific mechanical wager on physics and industrial supply chains. The thesis says that computing power has outrun the physical power grid. Hyperscale tech companies can buy as many graphics processing units as their balance sheets allow, but those chips are useless without electricity. Building new substations, installing high-voltage switchgear, and securing industrial cooling systems takes years of permits and heavy construction. Under this thesis, Quanta and Eaton hold real pricing power. They own the factory capacity and the crews. Microsoft and Amazon have no choice but to keep writing high-margin orders. Wall Street projections put data-center power demand at 165% higher before the decade ends. Analysts estimate that enterprise technology will require over 70 gigawatts of dedicated generation capacity in the next few years. Because utilities are only a fraction of the portfolio, you sidestep much of the rate sensitivity that plagues income funds, and you are buying purchase orders and swelling equipment backlogs instead. It is the pick-and-shovel trade in its purest, most literal form.

The Bear Case

The skeptic looks at this portfolio and sees an overcrowded cyclical trade waiting for execution failure. You are paying a steep 0.69% fee for an industrial basket that completely failed to beat the market over the past year. You took on 41% volatility, sat through a stomach-churning 38% drawdown, and earned no dividend yield whatsoever. The diversification is largely an illusion: over half the fund sits in an identical trade tied to commercial construction spending. If a single hyperscaler announces a temporary pause in capital expenditure during an earnings call, Quanta, Eaton, GE Vernova, and Vertiv will all plunge simultaneously. The power generation thesis has already started to crack. Constellation Energy and Cameco both suffered severe declines over the past year, proving that nuclear permits and uranium contracts do not move in a straight line. The utility holdings remain vulnerable to state-level regulatory backlash as local governments push back against tech giants offloading grid upgrade expenses onto everyday households. Worse, the fund's index rules systematically force it to buy speculative assets at the peak of the hype cycle. The August methodology rewrite added volatile neoclouds and GPU-rental operators right before they imploded. Recent additions like Fluence, Encore, and VNET suffered catastrophic drawdowns between 65% and 80% year to date. A new weekly fast-track rule now shuffles freshly minted initial public offerings into the fund at weights up to 5% within days of their debut. Meanwhile, the valuation leaves zero margin for error. The portfolio trades at nearly 35 times trailing earnings, excluding loss-makers. Count the unprofitable firms and the aggregate multiple turns negative. All of that buys a premium over the market for barely faster growth. The macro climate is darkening, with recent political shifts raising doubts about federal energy buildout timelines. If tech capital spending flatlines for two consecutive quarters, the bears will be proven entirely right.

How to Use It

This fund has no business in the core of your portfolio. It is far too volatile, carries too much industrial concentration, and yields nothing. If you use it to replace a traditional utility allocation, you are swapping stable, defensive income for an aggressive capital expenditure cycle. Instead, the fund functions exclusively as a tactical satellite holding. It belongs in a high-risk bucket for investors who already own broad-market equities and want a dedicated view on hardware supply constraints. Sizing matters. A position larger than three to five percent of an overall portfolio exposes you to unnecessary drawdown risk. If you own it, accept the ride. You need the stomach for thirty percent declines and a belief that rules-based rebalancing beats an active manager's discretion. If that sounds unappealing, skip it. If you want lower fees, buy the iShares power infrastructure alternative at nearly half the cost, and if you want steady dividends, buy a standard utility fund. If you simply want broad exposure to artificial intelligence, stick with a cheap market-wide index fund.

The Verdict

AIPO gives you clean, targeted access to the physical bottlenecks stalling AI, at a premium price for a portfolio that swings like a volatile industrial contractor. Before committing capital, keep your eyes on the calendar. MarketVector announces its quarterly index reconstitution on December 11, with implementation set for December 18. Watch closely to see whether the benchmark continues to funnel cash into speculative, newly public infrastructure firms. Keep an equal watch on January capital spending updates from the major cloud providers to confirm whether equipment backlogs are translating into sustained corporate earnings.

Analysis, not advice. We may own these names and we may sell them at any time.