Vertiv

Q4 FY2025 organic orders surged +252% YoY and book-to-bill hit ~2.9×, bringing backlog to $15.0B — up 109% in one year. Your own FY2025 10-K says enterprise datacenter square footage 'has generally been flat for the past three years.' Who is actually ordering, and what are they building for?

Both facts are primary from the same filings: Q4 organic orders +252% YoY, book-to-bill ~2.9×, backlog $15.0B (+109%) from the Q4 FY25 earnings release (2026-02-11) and FY2025 10-K Item 1. Enterprise DC square footage flat for three years is verbatim FY2025 10-K Item 1. The 10-K itself resolves the apparent contradiction: growth is hyperscale, colocation, and neocloud — the sheet estimates ~85% of FY2025 revenue is data-center end-market, with hyperscale and colocation called out as Q4 order drivers. MIT NANDA (Aug 2025) shows ~95% of enterprise GenAI pilots with zero measurable P&L impact — the enterprise segment that is flat. The hyperscalers ordering VRT are investing ahead of that ROI, not because it has arrived.

From Q1 2026 onward, management stopped reporting quarterly orders, order forecasts, and backlog — citing 'lumpy' large orders creating 'excessive volatility not representative of sustained performance.' This happened right as backlog peaked at $15B. Why pull real-time demand visibility at the top of the cycle?

The discontinuation of quarterly orders and backlog reporting is confirmed from the Q4 FY25 earnings call transcript (med conf). The rationale — lumpiness creating misleading quarter-to-quarter volatility — is management's stated reason. The timing — quarterly reporting stopped with Q1 2026, the same quarter backlog peaked at $15B — is noted in Indicator 4 as a "mild opacity flag" that "reduces real-time demand visibility." The sheet scores Indicator 4 GREEN-AMBER (~32–42), not red, because the balance sheet is clean (~0.5× net leverage, $2.6B liquidity, $2,114M OCF FY2025). The opacity concern is informational: investors lose the quarterly forward signal precisely when the 10-K's own Item 1A language — "We may not realize all of the sales expected from our backlog" — becomes most material.

The FY2025 10-K cites neocloud customers like CoreWeave and Nebius as 'growing rapidly.' CoreWeave is a GPU-collateralized borrower per the NVDA sheet. If neocloud customers cannot service their debt obligations, what is the exposure in Vertiv's $15B backlog?

The neocloud citation is primary: "growing rapidly" is verbatim FY2025 10-K Item 1. Revenue mix between neocloud, hyperscaler, and colocation is NOT SOURCED — Vertiv does not break out revenue by customer type in the 10-K. The 10-K is explicit on backlog risk: orders "subject to cancellation or rescheduling" (Item 1A). The sheet's honest position is that the $15B backlog includes neocloud exposure of unknown size, and that neocloud concentration "amplifies AI-capex cyclicality." The link to CoreWeave's GPU-collateralized financing structure is from the NVDA sheet Indicator 4, not from VRT's own filings — the sheet presents it as analytical synthesis, not a sourced VRT filing fact. The bear case stops at: neocloud is a real and growing customer segment, cancellation risk is explicit, and the dollar exposure within the $15B is NOT SOURCED.