$DGXX: A Mispricing Framework
I bought
$DGXX because I believe the market is effectively ascribing approximately zero value to the signed, executed $1.1B Master Services Agreement with Cerebras, who IPO’d yesterday at $70B.
$DGXX trades at ~$530M market cap against a disclosed, executed $1.1B 10-year Master Services Agreement with
$CBRS, a counterparty that simultaneously IPO’d at $70B, raising $5.5B in primary capital. The contract was never anonymized; Cerebras was named in the 8-K at signing. The market is not discounting counterparty quality. It is discounting execution.
The execution discount rests on two legitimate concerns: (i) construction financing and (ii) delivery timeline. Both were directly addressed on the Q1 call. CEO Amar stated explicitly: “We already signed a term sheet with a lender in order to grow our business through smart debt financing” and outlined a 70/30 loan-to-cash structure. The Cerebras contract is what makes that debt bankable. A 10-year MSA with a $70B counterparty is precisely the collateral against which construction lenders underwrite. DGXX does not need to tap equity markets to build Columbiana. It needs a construction lender, and it already has a signed term sheet.
This matters structurally. A signed 10-year colocation agreement with an investment-grade-equivalent counterparty is precisely the bankable instrument project lenders underwrite against. DGXX does not need equity markets to build Columbiana. It needs a construction lender and it has a term sheet.
The residual risk is pure execution: can a company that was a Bitcoin miner 12 months ago deliver a 40MW Tier-III AI data center by December 2026? This is where Hans Vestberg becomes analytically significant, not as a reputational signal, but as operational infrastructure. He oversaw Verizon’s nationwide 5G physical build and ran Ericsson’s global network deployment across 180 countries. He currently sits on the BlackRock board. The construction risk and the capital network risk are the same person’s problem.
The valuation dislocation becomes precise when you layer in NAV and live revenue. Hard asset NAV as of March 31: $71.4M cash, $13.6M in digital assets, $26M net fixed assets, and 210MW of grid-connected power infrastructure owned outright including substations, conservatively worth $105M at $500/kW. Total hard NAV: ~$216M. Against that, the company has zero long-term debt. Even before you apply any multiple to GPU revenue, the hard asset NAV alone is $216M against zero debt, and the GPU revenue stream is live and growing, with 7MW of bare-metal B200/B300 capacity expected operational by year-end at prevailing AI compute rates of ~$300/kW/month. Hard NAV plus GPU revenue stream approaches or exceeds the current $530M market cap in its entirety.
The implication is not subtle: the market is ascribing approximately zero value to the signed, executed $1.1B Master Services Agreement with a counterparty that IPO’d yesterday at $70B. That is not a marginal mispricing. It is a categorical failure to integrate simultaneously available public information, the contract, the balance sheet, the financing term sheet, and the Cerebras IPO, into a coherent valuation.
$DGXX