HOUSTON, June 4, 2026 — ERock is coming to market at exactly the moment public investors are looking for ways to underwrite the physical bottlenecks behind AI growth. The company launched its roadshow on June 1 for 27,906,977 shares at $20 to $23 each, a deal that points to roughly $600 million of gross proceeds at the midpoint before any greenshoe. It has also applied to list on the NYSE as EROC.
That timing is the real hook. ERock, formerly Enchanted Rock, is pitching natural-gas-fueled onsite power as a bridge, backup and dispatchable solution for data centers and other large-load customers that cannot wait for utilities to solve interconnection queues and grid constraints. Its own marketing has leaned hard into that demand window, arguing that onsite generation can help operators move despite delayed utility capacity, while a January hire for the data-center business underscored management’s push into hyperscale and AI-related workloads. Those are not abstract slides: ERock said in its original prospectus that it had more than 2,000 deployed units across about 400 sites, roughly 1,000 MW of installed base and approximately $1.3 billion of contracted power-system-sales backlog as of March 31.
The backlog is what keeps this from reading like a pure thematic float. In the June 1 amendment, ERock said it had started in-house assembly at its Titan facility in 2025, expects Hyperion to come online in the second half of 2026, and expects to work through backlog over about three years. A separate February company release adds more operating color: ERock signed a lease for a 407,302-square-foot Cypress, Texas facility that it described as an assembly, production and testing hub for its natural-gas generators. IPOGrid reads that as one of the more tangible capacity-expansion stories in the current new-issue calendar. This is not a pre-revenue power concept. It is a manufacturer-operator trying to industrialize faster.
The underwriting bench matches that ambition. Morgan Stanley and J.P. Morgan are leading, with Barclays, BofA Securities, Evercore ISI, Guggenheim Securities, Wolfe | Nomura Alliance and BNP Paribas on the book. For a company still in the red, that syndicate matters. It suggests ERock has found an audience among sponsors and institutions that want exposure to power-constrained data-center buildouts, utility flexibility and replacement of diesel-heavy resiliency systems with something cleaner and more scalable.
The financial picture is good enough to support a real debate, but not clean enough to coast on the AI-power narrative. In the latest prospectus, ERock reported $31.7 million of revenue for the quarter ended March 31, up from $24.1 million a year earlier, alongside a net loss of $17.2 million versus $15.9 million. For full-year 2025, revenue rose to $183.1 million from $128.5 million, while adjusted EBITDA improved to a loss of $22.6 million from a $34.9 million loss. The balance sheet also shows why the deal is arriving now: as of March 31, ERock had $300.5 million of cash, $564.1 million of assets and $645.5 million of liabilities. Our interpretation is that investors are being asked to fund a scale phase before the income statement has proved it can absorb that buildout cleanly.
The bigger reviewer concern is structural, not thematic. ERock is not selling a straightforward C-corp IPO. The June 1 filing lays out an UP-C structure with Class A stock, non-economic Class B voting stock, exchangeable units at the operating company, and a tax receivable agreement that could require substantial future cash payments to pre-IPO holders. The company also says a portion of IPO proceeds will be used to buy Class A units, buy Class B units from certain pre-IPO owners including Energy Impact Fund (FT-D) LP and some current and former insiders, pay blocker-merger cash consideration, and then send the remaining proceeds toward debt repayment and general corporate purposes. At the midpoint, the filing says public buyers would own only 10.21% of ER Holdings’ economic interests and 12.74% of ERock’s voting power, assuming no greenshoe exercise.
That does not kill the deal. It does frame it. ERock looks like a timely, scaled power-infrastructure issuer with real demand tailwinds, a meaningful backlog and a bank group that gives the offering more weight than the average industrial-energy IPO. But the public-market ask is also clear: buy into the AI power-shortage trade, accept a layered sponsor-and-insider-friendly structure, and trust that manufacturing expansion plus backlog conversion can outrun dilution, complexity and execution risk. The bullish case is easy to see. The discipline is in remembering that the public float is getting only a slice of the economics while taking on the full debate over whether this business can turn a hot market window into durable public-company returns.