Accelevation, a US specialist in the physical infrastructure of AI data centers, filed for a Nasdaq IPO on September 2 under the ticker ACCV. The deal epitomizes the new wave of AI-linked listings, fueled by record backlogs and accelerating hyperscale capex.
🔑 Key Takeaways
- Accelevation filed for a Nasdaq IPO under ticker ACCV
- 2025 revenue: $448M, up +148% YoY; net income doubled to $21.8M
- $1.1B backlog as of June 30, 2026 — roughly 2.5x annual revenue
- Vertically integrated model: power distribution, liquid cooling, modular platforms
- Listed peers: Vertiv, nVent, Eaton, Schneider Electric, Legrand
A specialist in the bones of data centers
Accelevation makes neither chips, nor servers, nor software. The company sits upstream of the value chain, designing, manufacturing, and installing the interior infrastructure of data centers — everything that fills a building before servers are plugged in. High-density power distribution, liquid cooling, modular platforms: each is a critical building block for bringing a site online fast.
Its key differentiator is a fully integrated model. While most competitors outsource parts of the process, Accelevation controls everything from raw metal to on-site installation. This vertical integration shortens lead times and appeals to hyperscalers (the largest cloud operators) and colocation providers under pressure to build faster to meet AI workload demand.
Numbers that capture the market mania
The financials disclosed in the S-1 (the IPO registration document filed with the SEC, the US regulator) reflect the magnitude of the trend. Accelevation posted $448M in revenue in 2025, up from $181M a year earlier — a +148% jump. Net income more than doubled to $21.8M. The backlog reached $1.1B as of June 30, 2026, equivalent to roughly 2.5x annual revenue.
| Metric | 2024 | 2025 | Change |
|---|---|---|---|
| Revenue | $181M | $448M | +148% |
| Net income | ~$10M | $21.8M | >x2 |
| Backlog | n/a | $1.1B | ~2.5x revenue |
This multi-year visibility, uncommon in a cyclical industry, is a major selling point for investors. It reflects framework agreements with hyperscalers and colocation operators whose AI infrastructure spending now drives most of the growth. The lineup of Morgan Stanley, JPMorgan, Goldman Sachs, Barclays, and Bank of America as lead underwriters reinforces the signal sent to the market.
A $2.4T boom — and its industrial frictions
Accelevation’s IPO lands at the heart of a powerful trend. According to Bridgewater, US tech giants’ capex could rise from $410B in 2024 to more than $650B in 2026, mainly in AI. Bloomberg estimates the four largest players have already committed roughly $2.4 trillion in future AI infrastructure spending.
The planet hosts more than 11,600 data centers, with nearly 40% in the United States. In France, Emmanuel Macron announced €109B in private investment at the AI Action Summit to position the country as a data center hub. Yet rapid growth runs into industrial bottlenecks.
« The transformer shortage is becoming one of the most critical friction points in the AI boom. »
Several Bloomberg analyses, 2026
In the US, lead times for some high-capacity transformers now exceed 160 weeks, compared to roughly one year in the early 2020s. According to multiple Bloomberg estimates cited by dcmag.fr, between 30% and 50% of US data center projects slated for 2026 could be delayed or cancelled due to a lack of transformers, switchgear, and batteries needed to connect sites to the grid. The International Energy Agency (IEA) pegs data center electricity consumption at 415 TWh in 2024 — about 1.5% of global demand — and projects a doubling to 945 TWh by 2030. The Bank of America Institute also forecasts a near-doubling of data center water and electricity use over the same horizon.
A sector still far from AI pure plays
Accelevation is not, and does not claim to be, a pure-play AI company. It sells physical infrastructure and power, not models or GPUs (graphics processing units used for AI compute). That nuance applies to the entire sector: its listed peers — Vertiv, nVent, Eaton, and on the European side Legrand and Schneider Electric — are no more pure plays. They all ride the same wave with different business mixes.
It is precisely this indirect AI exposure that makes valuing equipment makers tricky. An AlixPartners survey reported by Bloomberg shows that 68% of the 400-plus executives polled in the data center industry expect a rise in distressed situations over the next 12 to 18 months. The Bank for International Settlements (BIS) has also warned against the heavy use of debt and circular financing arrangements between hyperscalers and AI developers.
Toward a closely watched listing
For Accelevation, the listing represents a key milestone: funding capacity expansion, reinforcing its backlog, and capitalizing on the current market window. The choice of five top-tier lead underwriters sends a strong signal to institutional investors and validates market appetite for AI-linked equities.
No date or price range has been disclosed yet. The ACCV ticker will be available to European investors through brokers with access to US markets. The central question remains: can the IPO market absorb this new wave of AI-linked listings without triggering a sector correction? The answer will hinge on interest rates, the cost of critical equipment (transformers, switchgear, batteries), and the actual pace of hyperscaler deployments.
Sources
- Cryptoast — Data centers IA: this little-known supplier goes public
- Zonebourse — Data centers: will AI blow the electrical meter?
- Multinationales — Profits and losses behind the datacenter boom in France
- Boursorama — AI datacenters seek their solution in power semiconductors
- DCmag — Risk of speculative bubble on AI data centers
- Le Monde — AI: how data center deployment creates a resource shock
This article is published for informational and educational purposes only. It does not constitute investment advice. Do your own research (DYOR) before making any decision.
