On September 10, 2026, a Nasdaq shell company called Aperture AC agreed to buy Atlantic HPC Group, a bitcoin miner with 16 employees, for $150 million in stock. Atlantic's own numbers for the fiscal year that ended June 30: revenue of $28.6 million and a net loss of $10.1 million.

The deal also carries an earnout. If Atlantic signs one lease for 5 megawatts of data center space, its owners get another 3,000,000 shares. At the $10 deal price, that is $30 million for a signature.

What Aperture is buying

Atlantic was founded in 2024 and is based in Irvine, California. It mines bitcoin at four sites: Oklahoma City (20 MW), Blytheville, Arkansas (10 MW), Rector, Arkansas (12 MW) and Alledonia, Ohio (9 MW). The investor presentation filed with the SEC on September 11 counts 51 MW "in operation." A footnote adds that the 9 MW in Ohio is offline for a hardware upgrade, so the capacity actually mining today is 42 MW.

In fiscal 2026 the fleet ran at 1.9 exahash per second and mined 303 bitcoin. Revenue was $28.6 million, which includes some colocation income from hosting other people's machines. Depreciation was $13.7 million, so 48 cents of every revenue dollar went to writing down mining rigs. After that and a handful of small items, the net loss came to $10,052,000.

The press release leads with a different figure: "FY2026A EBITDA of $4.4 million." That is adjusted EBITDA, and the appendix shows the path to it. Start from the loss, add back the $13.7 million of depreciation to reach $3.7 million, then add $550,000 of travel expenses, $180,000 of realized losses on selling crypto and a few smaller lines. Adding travel back into EBITDA is unusual. Most companies treat plane tickets as a cost of doing business.

The forecast the company published

Atlantic's projected revenue by source, fiscal years ending June 30 ($ millions)

Mining revenue: 28.6, 24.7, 22.8, 13.7. AI infrastructure revenue: 0, 0, 13.2, 26.4. Adjusted EBITDA: 4.4, (3.9), 6.5, 17.1. Source: Aperture AC Form 8-K, September 11, 2026, Exhibit 99.2, page 22. FY2026 is unaudited; later years are management estimates.

Fiscal 2027, the year the deal is meant to close, is a down year in Atlantic's own model. Revenue falls to $24.7 million, all of it from mining. Adjusted EBITDA swings from positive $4.4 million to negative $3.9 million. The AI revenue line reads zero.

One detail in that forecast is hard to square. Atlantic projects mining 332 bitcoin in fiscal 2027, 29 more coins than in fiscal 2026, and earning $3.9 million less revenue from them. The deck does not state the bitcoin price it assumes. Either the model expects a lower price, or the colocation income goes away, or both. The proxy statement, when it is filed, should say which.

AI revenue is projected to start in fiscal 2028 at $13.2 million and reach $26.4 million in fiscal 2029. By then mining revenue is modeled at $13.7 million and bitcoin mined drops to 166, with the halving expected around April 2028 sitting in the middle of that path. A company that is 100% mining today is modeled as two-thirds AI landlord in three years.

What "AI infrastructure" means here

The AI business is a planned campus at the Alledonia site, backed by 34.8 MW of utility-approved capacity. Phase I is 5 MW. Phases II and III, at 10 MW and 20 MW, include the 9 MW currently allocated to mining. The campus grows partly by switching off miners.

Atlantic's role would be to build a powered shell. Under the leasing structure in the presentation, the tenant supplies the liquid cooling, the racks, the GPUs, servers, storage, networking and cabling. Atlantic charges monthly rent on contracted power and passes electricity through at cost, over a seven-year term. That is a real estate business with a substation attached, and it is a reasonable one. It is also a business Atlantic has never run. The filing states the AI/HPC unit "has not generated material revenue to date," and the only customer described is a non-binding letter of intent for Phase I.

The earnout math

That letter of intent is where the 3,000,000 earnout shares come in. The condition is a binding, arm's-length lease for the 5 MW Phase I, with a tenant that is not an affiliate of Atlantic, whose obligations are not funded or guaranteed by Atlantic or its stockholders, for a non-cancelable term of at least seven years.

At $10 a share the milestone pays $30 million, or $6 million per megawatt, for turning one letter of intent into a contract. The shares go to Atlantic's existing owners, not to the company. For comparison, the presentation's own estimate for a shell retrofit is $3 million to $3.6 million per megawatt. The reward for signing the lease is larger than the cost of building the space it covers.

Two further tranches of 1,500,000 shares each vest if the stock averages $12.50 and then $15.00 over any three consecutive calendar months. Those depend on the market. The lease tranche depends on one counterparty, and the deck already calls that counterparty the "expected first contracted AI infrastructure customer."

The term that matters more than the headline. The lease earnout tells you two things. Management is confident the letter of intent converts, or they would not have priced it into the deal. And the price of that conversion, $30 million in shares to insiders, was negotiated before public shareholders get a vote. I think this clause deserves more attention than the $150 million figure.

Where the $324 million comes from

Aperture raised $102 million in its IPO on May 22, 2026, selling 10,200,000 units at $10 each. That cash sits in a trust account. The combined company's pro forma equity value of $324 million assumes every public holder keeps their shares at the vote. The pieces, from the sources and uses table on page 24 of the deck:

  • Atlantic's owners: 15,000,000 new shares, $150 million, 46% of the company.
  • Public shareholders: 10,200,000 shares, $102 million, plus rights that convert to about 2,550,000 more shares at closing. Each IPO unit came with one-quarter of a right.
  • Sponsor: 4,200,000 shares and rights, $42 million at the deal price, 13%. Aperture Sponsor LLC paid $25,000 for its 3,828,082 founder shares in September 2025, about $0.008 per share, and later bought 311,000 private placement units at $10.
  • Underwriters: 450,000 representative shares, issued at one cent apiece.

Cash to the balance sheet is $97 million after $5 million of fees, in the zero-redemption case. Every public shareholder can instead take back roughly $10 a share when the merger comes up for a vote. If many do, the cash shrinks and the sponsor's and Atlantic's percentages rise. Under its current charter Aperture must close a deal by May 22, 2027; the target is the first quarter of 2027.

On valuation, the deck puts enterprise value at $227 million, 7.5 times fiscal 2027 estimated revenue, against a peer range of 3.0x to 25.7x. That looks mid-pack until you notice the peers are miners with hundreds of megawatts already under contract to AI tenants. Atlantic has none under contract.

Three risk factors that read differently for a miner

The forward-looking statements section is boilerplate for most SPACs. Three lines here are specific to Atlantic, and none of them appears in the press release.

First, the company depends "on a single mining pool operator for substantially all of its mining revenue," and that operator can adjust its fee rates. Second, Atlantic has "fixed-delivery hashrate purchase and sale arrangements and the related derivative liability, including the consequences of non-delivery of bitcoin." In plain terms, it has sold future mining output for money up front; if the machines underdeliver, it owes bitcoin it has not mined. The balance sheet carried a $500,000 derivative liability for these contracts at June 30. Third, Atlantic "holds all mined digital assets in self-custody without a third-party custodian," which for a 16-person company means the coins are as safe as whoever holds the keys.

What to read when the S-4 arrives

  • The bitcoin price assumption. The forecast has more coins and less revenue in fiscal 2027. The registration statement should state the price behind that.
  • The hashrate contracts. How much output is pre-sold, to whom, and what happens on a shortfall.
  • The letter of intent. Who the Phase I tenant is and whether any affiliate relationship exists. The earnout excludes affiliates for a reason.
  • The redemption result. The $97 million cash figure only holds if nobody takes their $10 back.

What to watch between now and the vote

Atlantic CFO Benson Liu said in the release that a public listing would "support our ability to scale rapidly through increased financial flexibility." The flexibility in question is $97 million of other people's IPO money, and the scaling is a 5 MW building that has one prospective tenant on a non-binding letter.

The single event that changes this story is a Form 8-K announcing a binding seven-year lease for Phase I before the shareholder vote. That would make the earnout a formality and the AI revenue line credible. If no such filing appears, Aperture's public holders will be voting on a mining company priced as a data center developer, in a year that company itself projects will lose money. The September 11 Form 8-K and its two exhibits contain every number cited here.