Ionic Digital: A Real Pivot, but the Stock Is Priced for 2028 Cash Flow

I am rating Ionic Digital (IOND) a Hold after its first quarterly report as a public company, and the reason is valuation, not the business. The bitcoin-miner-turned-AI-landlord story is genuinely improving — the direct listing on July 28 was priced at a reference price of $53, and the stock now trades around $67, roughly 27% higher in under a month and a short distance below its $71.90 high. That run-up happened before the quarter even printed, which means the easy money from the pivot narrative has already been made. The question now is whether a ~$3.0 billion market cap for a company guiding to ~$190-195 million of 2026 revenue is a bargain, a fair price, or a bet on cash flow that will not actually arrive until 2027-2028. My answer: it is a bet on 2028 cash flow, and at this entry the risk/reward is no longer skewed in the buyer's favor.

The debut quarter was a good one

Ionic Digital reported revenue of $48.6 million for the second quarter, up 31% year over year, with 90% of that coming from digital infrastructure leasing and the remainder from what is left of its Bitcoin mining operation. Adjusted gross margin jumped to 93% from 40% a year earlier, and the leasing segment alone ran at a 99% adjusted gross margin. Adjusted EBITDA — earnings before interest, taxes, depreciation, and amortization, the metric the market is using to value this transition — came in at $37.6 million, or about 77% of revenue.

That is an eye-popping margin for a company that, a year ago, was still mostly a mining operation. But the GAAP loss tells a different story: Ionic posted a net loss of $35.3 million for the quarter, hit by a $28.2 million non-cash fair-value charge on its Bitcoin holdings and a $27.2 million income-tax provision. In other words, the headline profitability the bulls are celebrating is an adjusted number, while the actual reported line is still deeply red.

The catch hiding inside the EBITDA

Here is the part that matters for anyone pricing this stock: the $48.6 million of revenue is mostly not cash that hit the bank account. Roughly $43.8 million of the quarter's digital infrastructure revenue was straight-line lease recognition — an accounting convention that spreads the total value of a long-term lease evenly across the contract period, regardless of when payments actually arrive.

The cash economics tell a very different story from the income statement. Actual cash rent from the company's anchor tenant, Nscale, only began in August, after the quarter ended. Management guides to roughly $29 million of cash rent for all of 2026, ramping to about $183 million per year at full run-rate by the end of 2028. Against that, the company is guiding to capital expenditures of $45 million to $60 million this year. So in 2026, Ionic will spend more on building out data centers than it will collect in cash rent — this year's free cash flow is comfortably negative even though the guided adjusted EBITDA of $137.5 million to $142.5 million, roughly a 71% margin on revenue, looks spectacular.

This matters because of what the market is capitalizing. At roughly a $3.0 billion market cap, the stock trades near 15.5 times forward 2026 revenue and about 21 times the guided adjusted EBITDA that, on a cash basis, does not exist yet this year. On trailing twelve-month revenue of roughly $152 million — a base that still includes the dying mining business and the transition period — the price-to-sales multiple is about 19.9 times. Adjusted EBITDA here is a lease-recognition construct: it adds back depreciation on assets that are the entire point of an asset-heavy landlord business, and it counts revenue that will not become cash until 2027-2028. Paying 21 times that number today presumes the ramp not only happens, but happens on schedule.

Why the contract still deserves credit

None of this is an argument that the underlying asset is weak, and I want to be clear about that. The Nscale lease is a 10-year agreement that started at 234 megawatts and roughly $2 billion in contracted revenue, then expanded via amendment to 323 megawatts and $2.6 billion in total contracted revenue. The critical detail is that NVIDIA guarantees the first five years of rent on the initial contracted capacity — meaning a top-tier AI hardware credit stands behind a large chunk of the cash flow, not just Nscale alone. Ionic's model is to build powered, cooled data-center shells and lease them out in full rather than retaining operating control or selling equity stakes in the hardware inside, which removes GPU-depreciation and chip-obsolescence risk from the balance sheet entirely.

The balance sheet is the other reason this is not a distress story. The company ended the quarter with more than $400 million in cash, zero debt, and 2,882 Bitcoin worth roughly $169 million, for total liquidity near $600 million. The direct listing, with no underwriters, plus a June preferred stock sale at $53 per share, brought the capital in. A single-quarter net loss driven by non-cash fair-value marks on Bitcoin does not threaten a company with that much dry powder.

The growth leg is gated by regulators

The problem is what comes after the Nscale lease, because that is what the current valuation implicitly requires. The next big step is a planned 466-megawatt expansion at the Ward County, Texas site, which would bring the campus to roughly 700 megawatts and roughly triple the current anchor capacity. But the energization timeline is tied to an interconnection review by ERCOT, the Texas grid operator, and ERCOT has paused its large-load interconnection process while Texas regulators verify projects. Ionic says it has no date for the review and no certainty on the phasing of approvals. The first data center is energized and a second development site in Midland is being converted from mining to high-performance computing, but those are early-stage building blocks, not contracted cash flow.

This is where the premium gets uncomfortable. At $67, the market is not just paying for the Nscale contract — it is paying for the Ward County expansion, the Midland conversion, and a second tenant that does not exist yet, all arriving while the one catalyst that would prove the thesis (energization of the expansion) is stuck in a regulatory queue with no timeline. On trailing price-to-sales, Ionic sits in the same neighborhood as fellow AI-pivot miners — roughly 20 times for IREN, about 14 times for Core Scientific, and over 30 times for Cipher Mining — but those comparisons flatter it because peers have already begun converting to cash-generating AI hosting, while Ionic's trailing base still reflects the mining wind-down.

Risks that would break the thesis

• Concentration. One tenant (Nscale) and one anchor site (Ward County) drive essentially all the revenue. The NVIDIA guarantee de-risks the initial capacity, but the model beyond it is unproven.
• Regulatory gating. The Ward County expansion — the source of nearly all future growth — is waiting on an ERCOT review with no scheduled date.
• Recognition versus cash. Roughly 85% of this year's guided revenue is straight-line recognition. If investors ever start valuing Ionic on cash rather than adjusted EBITDA, the multiple re-rates lower, fast.
• Crypto exposure. The 2,882 Bitcoin add liquidity but also inject volatility into reported earnings, as the $28.2 million Q2 fair-value charge showed.
• Thin scaffolding. The stock has been public for three weeks with no underwriters and essentially no sell-side coverage or consensus estimates to anchor the valuation. Price discovery is still forming, which cuts both ways — rallies can run, but so can drawdowns when there is no analyst floor.

Investor takeaway

Ionic Digital has done something most bitcoin miners only talk about: it converted a stranded-power asset into a contracted, long-duration AI infrastructure lease with a top-tier credit backstopping the early cash. That is why I would not short the story. But the stock's 27% run since listing has already absorbed the good news, and at roughly 15.5 times forward revenue and about 21 times a guided adjusted EBITDA that is mostly non-cash this year, the market is paying for the 2027-2028 cash ramp in full, today.

I rate IOND a Hold at current levels. The risk/reward becomes compelling again under two conditions: a pullback into the low-to-mid $50s, which would put the forward multiple closer to where the uncertainty actually lives, or proof of the next leg — a new anchor lease, Ward County energization on a visible schedule, or cash rent tracking toward the guided $135 million for 2027. Until one of those materializes, the premium is the risk, and waiting is cheaper than paying for it.