Infrastructure & Construction Tech

ACCV — IPO pricing analysis (2026-10-01)

Published

Not financial or investment advice, and not a solicitation to buy or sell any security or to use our service. This is AI-assisted analysis of public SEC filings, provided for informational purposes only.

Figures are drawn from those filings and checked against them; any interpretation, conclusion or forecast is our analytical opinion rather than a statement of fact. We do not intend to defame any person or entity.

It is generated by an automated pipeline with no human review and will likely contain factual errors, interpretive errors, and errors of omission that misrepresent the business or the offering — none of them intentional. Do not rely on it as the sole basis for any investment decision. Always do your own research.

Accelevation priced 10% below its own range floor — the operating business wasn't the problem

Accelevation Holdings sold 30,000,000 Class A shares at $18.00 on October 1, 2026 — 10% below the $20.00 floor of the range it had marketed. The stock opened at $17.55, another 2.5% below even the cut price, suggesting the clearing level was found rather than imposed. The operating business — power infrastructure for hyperscaler data centers, with a 3.5x book-to-bill in Q2 2026 and a backlog that grew 165% in the first six months of 2026 — is not what the book was discounting. The structure is.

FY2024FY2025H1 2025H1 2026
Revenue$181.4M$447.8M$158.6M$437.5M
Gross profit$58.9M$144.7M$48.3M$116.2M
Adj. EBITDA$29.6M$90.9M$26.8M$68.4M
Net income$9.4M$21.7M$(8.7)M$19.3M
Backlog$63.2M$419.3M$347.2M$1,111.3M

Revenue grew 147% from 2024 to 2025, then H1 2026 revenue came in 176% above the same period a year earlier. Adjusted EBITDA in H1 2026 rose 155% year-over-year. Three hyperscalers engage the company at full-platform design level, and modular solutions — a product line that didn't exist before September 2025 — booked roughly $700M in its first nine months. Allocators who passed on this order book at $22 were not passing on the demand signal.

What the company actually builds

Accelevation makes the electrical guts of the data center white space: branch circuit whips, remote power panels, power distribution units up to 1,500 kVA, containment hardware, and since late 2025 a factory-built modular platform called SkyBridge. The pitch to hyperscalers is schedule. The company estimates modular delivery can cut field installation to as few as 8 days and lift field productivity by up to approximately 84% versus traditional field-built methods — material when the company also estimates construction delays cost an operator roughly $1M per megawatt per month.

Founded in 2017 with revenues of under $3M in 2021, Accelevation now runs roughly 1.1 million square feet of manufacturing across southwest Ohio (625,000 sq ft), Memphis (225,000 sq ft), Houston, Mississippi (220,000 sq ft) and Richmond, Virginia (57,000 sq ft). Field crews number more than 800 as of June 2026, up from roughly 140 at year-end 2024. Average revenue per customer went from $700K in 2023 to $3.9M in 2025.

One note on margins worth tracking: gross margin in H1 2026 was 26.6%, down from 32.5% in the predecessor's 2024 period and 32.3% in FY2025. The mix is shifting toward installed systems and field labor, which carry less margin than fabricated product. SG&A ran $54.6M in H1 2026 versus $28.6M a year earlier, with related-party expenses inside that line moving from $500K to $6.7M. These trends deserve a watch but they don't explain the -10.0% below-range print.

The cap table is the structure being sold

The sequence in the six months before pricing tells the story.

On June 25, 2026, the credit agreement was amended to add $346M of incremental term loans plus $10M of delayed-draw commitments, with proceeds directed to a distribution to certain members and transaction expenses. Total equity collapsed — from $307.2M at December 31, 2025 to $31.3M at June 30, 2026, nearly wiped out in a single quarter. Term loan outstanding at June 30 stood at $651.5M at a weighted-average rate of 8.772%, with nothing drawn on the revolver. In July 2026 another $22.1M was advanced as distributions to Series P unit holders including executive officers.

Then the IPO. Two-thirds — 66.7% — of the 30,000,000 shares were secondary: 20,000,000 sold by Olympus-affiliated stockholders for $341.1M before expenses. The greenshoe of 4,500,000 shares is also entirely from the selling stockholders. The company's own net proceeds — $170.6M — go directly and entirely to repaying part of the term loan, leaving $479.9M of pro forma debt outstanding. Pro forma cash on hand after all of it: $19.3M.

The sponsor levered the company to pay itself in June, sold two-thirds of the IPO in October, and the public's $170.6M retired a fraction of the debt the business is left carrying.

Layered on top is a Tax Receivable Agreement paying 85% of realized tax savings — estimated at $589.8M of a projected $693.9M total, paid out over roughly 15 years — to TRA rights holders, not to Class A shareholders. The basis step-up that generates the shield accrues almost entirely to Olympus.

Governance is consistent with the economics. Olympus controls approximately 85% of voting power post-IPO (83% if the greenshoe is fully exercised). Public Class A buyers collectively hold roughly 13%. The company takes Nasdaq controlled-company exemptions from independent-board, compensation-committee and nominating-committee requirements. A Director Nomination Agreement lets Olympus name the entire board while it holds 40% of its original stake. Pro forma net tangible book value per share is negative — buyers at $18.00 absorb dilution of $19.67 per share.

The company also disclosed material weaknesses in internal controls for both FY2024 and FY2025, spanning entity-level governance, insufficient accounting personnel, absent formal risk assessment, undocumented accounting policies and IT general controls. Auditor attestation is not required until the second Form 10-K under the company's emerging growth company status.

Growth here doesn't self-fund — and the financing is not yet committed

Adjusted EBITDA of $68.4M in H1 2026 sat alongside operating cash flow of negative $4.2M and free cash flow of negative $15.8M. FY2025 was the same shape: $90.9M of adjusted EBITDA against negative $7.2M of operating cash flow and negative $16.4M of free cash flow. The gap is structural — this is a working-capital-heavy manufacturer buying steel, copper and switchgear ahead of revenue recognition, carrying $368.7M of surety bonds against its project obligations. Every quarter of 3.5x book-to-bill is a quarter that needs outside financing.

The prospectus anticipates replacing the existing facility with a new $250M term loan and a revolver with up to $225M available at closing. Both are explicitly disclosed as not yet committed as of the filing date.

The strongest case against this read

The honest counter: the window may have closed on the entire September 2026 IPO cohort, not on Accelevation specifically. Per Reuters, Renaissance Capital noted that AI infrastructure remained the dominant listing theme but that "market conditions have worsened," and at least one other infrastructure listing postponed that same week. If every deal repriced simultaneously, attributing Accelevation's discount to its TRA and June recap is overfitting.

That argument has real weight. It runs into two problems, though. A softening tape typically produces a low-end print on a deal, not a 10%-through-the-floor print with full-size and 66.7% secondary. And the most natural sector comparable — Forgent Power Solutions, which listed in February 2026 on the same AI-electrical thesis and priced at $27 inside its marketed range — argues the theme itself was not broken as recently as eight months ago. Something changed between those two prints, and the cap-table differences between the two deals are the clearest explanation on the evidence available.

The second steelman is more durable: if the $1,111.3M backlog converts on schedule, the cash-consumption dynamic reverses, and at that point the TRA only costs the company when it's generating real taxable income. The leverage looks more manageable against a business compounding at these rates. The live operational risk is concentration — two customers accounted for 61.2% of direct revenue in FY2025 — not the capital structure.

What to watch

  • Q3 2026 backlog: a print below $1,111.3M or book-to-bill below 1.0x would mean demand was always part of the discount, not just the structure.
  • New Credit Facilities closing terms: the expected $250M term loan and revolver are not committed. If they close materially wider than the current 8.772% weighted-average rate, the leverage picture worsens and the interest coverage math changes.
  • Customer concentration: two customers at 61.2% of FY2025 direct revenue means the loss or deferral of one hyperscaler program is a backlog event.
  • Lockup expiry: 180 days from pricing puts Olympus's remaining stake available in late March 2027. Per Bloomberg, the employee award program is structured in batches tied to Olympus selldowns — a retention plan that contractually anticipates sustained sponsor supply.

Accelevation appears to be a genuinely good industrial business wrapped in a structure built for the seller. The operating thesis and the equity thesis are not the same thesis here.

Sources

Regulatory filings

  • SEC filings

News & analysis

Market data

  • Markets data: Massive

Some figures are computed from the source material above.