$DMGI.V / $DMGGF — why the Christina Lake tenant is probably a real one, and why Canada is the point

Not confirmed. Tenant is still under NDA. This is the case for *why it looks like a hyperscaler / frontier-lab / sovereign-grade offtaker rather than a random GPU reseller.*

**Why it is probably one of the big names**

* Single tenant. Investment-grade backstop. 12-year initial term with renewals that can run \~27 years. That is not a 24-month neocloud flip.
* Tenant has a ROFR on the site. That is how a Google / Microsoft / Meta / Amazon / large neocloud or frontier lab behaves when they want the campus, not a cage.
* Management has said inbound interest picked up once the industry knew they were converting a large, already-operating site. People with land and a power quote are everywhere. People who have run 75 MW for years are not.
* Pricing was described as current AI colo market rates, and they have not been in a rush to rip up the LOI even as headline $/kW has moved higher. That usually means the original counterparty was already in the right zip code.
* Fiber ask is serious: two 100-gig routes, working on a third, tenant willing to pay for diversity. That is training / inference campus behavior, not a 2 MW inference closet.
* Could still be a strong IG neocloud instead of GOOGL/MSFT/META/AMZN. The credit quality and term structure are the tell, not the logo. The logo is the multiple.

**Canadian sovereignty angle**

* Ottawa’s 2026 AI strategy is explicit: sovereign AI starts with sovereign infrastructure. If Canada does not build it, it rents it.
* Canada has \~337 MW of AI data center capacity today and something like 5.5 GW of AI compute demand by 2030. The gap is the whole market.
* Meta already committed \~C$13B to Alberta. The US is running \~$75B annualized data center construction. Japan earmarked $60B. Google just put $15B into Finland. This is a national-power contest, not a miner pivot fad.
* Christina Lake is a Canadian-owned, already-energized BC campus. Zoning for data warehouse / AI is done. That is the scarce product Ottawa says it wants: domestic megawatts under Canadian control.

**How close they actually are**

* LOI has been live since June 1, 2026. Commercial terms have been described as largely agreed. Remaining fight is the fat stuff: delivery dates, SLA penalties, who pays if the hall is ready and chips are late.
* Site is already permitted for AI / data warehouse. Not waiting on a rezoning war. Building permit filed to take the building from \~30k to \~60k sq ft.
* Power is not theoretical: 75 MW available (15 MW firm / 60 MW curtailable). Written utility approval for the last 10 MW. Application in for another 150 MW firm.
* Existing electrical layout is a head start on roughly the first 30 MW of critical load. Phased at \~10 MW IT per hall, matching the transformer bays they already have.
* Fiber, GCs, mechanical, and vendors are being spent *ahead* of the definitive. First halls targeted Q1/Q2 2027. Full 50 MW in 2027 called realistic. They will not give a signing date. “You’ll know when we know.”
* Risk that still matters: LOI is not a lease. Dilution if they raise equity for a second site. Curtailable power / firming. Execution on 2027. Liquidity.

**Size it can go to**

* Base lease: 50 MW critical IT.
* Management has floated upsizing to 60 MW IT.
* At \~1.20–1.25 PUE, 50 MW IT is \~60–63 MW gross against a 75 MW campus. 60 MW IT uses almost the whole site.
* Same campus has a 150 MW firm-power application behind it. That is the “this is a platform, not a one-hall story” case.
* Other options, not in the LOI: Malahat Nation MOU (\~15 MW), Boardman, Oregon (3.75 MW + expansion, deposit still down), plus other Canadian sites in BC / Alberta / Ontario. Those are extras. Christina Lake is the company-maker.

**Market cap it could reach (scenarios, not a target)**

* Today: roughly C$0.50–0.60 and \~C$100–120M market cap on \~208M shares. US ticker $DMGGF.
* Signing a generic IG lease: the stock stops trading like a fading miner. Illustrative re-rate talk in the community has been C$3–5 on announcement, C$4–5 on a completed 60 MW residual. Call that high-hundreds of millions to \~C$1B.
* Signing a named hyperscaler / frontier offtaker: tighter project debt, better LTV, higher multiple. Same 60 MW residual talk has been C$5–8, or \~C$1.0–1.7B. Platform case with a second site stacked on top is how people get to C$8–11.
* Aggressive cash-flow case some holders use: 60 MW IT, mid-teens $/kW-month, high colo margins, 2027 run-rate EBITDA that can exceed today’s entire market cap. Put a 15–20x infrastructure multiple on that and you are talking multi-billion. That only works if the lease signs, phases deliver, and they fund with debt instead of a constant ATM.

**The actual bet**
$DMGI.V / $DMGGF is a \~C$100M Canadian miner with an already-built 75 MW campus and an unsigned 50–60 MW AI lease.

If the tenant is generic IG, it still re-rates.

If the tenant is one of the names building Finland, Alberta, and $75B of US shells, the multiple is the whole story.

the logo is the last thing they will tell you. the contract shape is already doing the talking