Key Takeaways:
Wall Street's first "Neocloud" ETF launched Aug. 6, bundling six companies that rent AI compute capacity back to the same hyperscalers that dominate QQQ.
Wall Street launched its first "Neocloud" ETF on Aug. 6, bundling six companies that rent AI compute to hyperscalers, with CoreWeave anchoring the fund on a $104 billion revenue backlog and 112 percent revenue growth. The fund landed the same day as a sister photonics and optics ETF carrying a 0.65 percent management fee, and it gives investors a direct route into the AI infrastructure rental trade that broad tech funds dilute.
"Scale began to translate into expanding operating leverage," CoreWeave CEO Michael Intrator said of Q2 results that showed adjusted EBITDA margin at 59 percent. CoreWeave reported revenue of $2.6 billion for the quarter, up 112 percent year over year, with more than $25 billion in net new customer commitments added early in Q3.
The fund's other holdings span the AI infrastructure rental chain. Nebius grew revenue 454 percent to $582 million, with a first-ever capacity auction clearing 15 percent above its highest prior Blackwell price. Applied Digital holds $36 billion in contracted leases, roughly 76 percent backed by investment-grade hyperscalers. TeraWulf signed a 20-year, roughly $19 billion lease with Anthropic for 401 megawatts at its Kentucky campus. IREN inked a five-year, $3.4 billion AI Cloud contract with NVIDIA tied to the eventual deployment of 600,000 GPUs. Lumentum, the optics arm, posted fiscal Q4 revenue up 109 percent to $1.01 billion with non-GAAP operating margin at 36.6 percent.
The fund isolates a trade that QQQ dilutes. Hyperscaler capex, which Applied Digital's Wes Cummins pegged at "nearly $700 billion" up from around $400 billion, flows directly into companies with the power, sites, and GPU allocations to absorb it. QQQ's top holdings are the hyperscalers spending on AI infrastructure, but AI revenue sits buried inside diversified businesses generating cash from unrelated products. The Neocloud basket is 100 percent AI infrastructure, contracted years out, and financed against those contracts.
The tradeoff is structural. In the last week alone, CoreWeave fell 18.63 percent and Nebius fell 21.55 percent. CoreWeave's Q2 interest expense reached $640 million versus $267 million a year earlier. TeraWulf posted a $939.92 million net loss driven mostly by non-cash warrant marks. IREN took a $140.4 million non-cash impairment, retiring mining hardware. Customer concentration is real: Nebius disclosed three customers representing 24 percent, 21 percent, and 14 percent of revenue, and TeraWulf leans on Anthropic and Google's $600 million credit backstop for Fluidstack. None of these names pay a dividend.
The broader AI compute financing picture adds another layer of risk. Nvidia and six of the largest Wall Street firms signed memorandums in August to mobilize more than $500 billion of outside capital around AI compute infrastructure, with Nvidia potentially providing residual-value support for up to 25 percent of an opportunity on a project-by-project basis. But a secondary market for used GPUs only opened in July, and it posts estimates rather than completed sales. There is no public benchmark for what a whole cluster fetches when the seller has no choice.
Whether the Neocloud trade works comes down to two pillars. First, hyperscaler capex commitments must hold — CoreWeave alone raised roughly $18 billion in the second quarter, which shows the scale of capital markets dependence. Second, capital markets need to stay open and cooperative for these structurally levered, cash-flow-negative businesses. If either pillar weakens, the Neocloud ETF will take a much bigger hit than QQQ. If both hold up, the concentration works in the investor's favor.
For investors weighing the swap, the two funds are structurally different. QQQ is a broad, diversified core holding. The Neocloud ETF is a satellite that hones in on one specific theme through levered businesses. A partial reallocation — sizing the Neocloud fund as a modest slice of AI exposure rather than replacing QQQ entirely — captures the pure-play upside without betting the whole portfolio on capital markets staying open for six companies at the same time. In taxable accounts, cost basis matters before rotating out of appreciated QQQ shares.
This article is for informational purposes only and does not constitute investment advice.
