129 Megawatts at a Dead-Miner Price, and the Insiders Are Loading - $BGDE

A failed Bitcoin miner, an aligned board buying with both hands, and 129 megawatts the market is pricing at a liquidation.

Big Digital Energy trades around 40 million dollars. It controls roughly 129 megawatts of energized, grid-connected power inside PJM, the most supply-starved power market in the United States. The people who just seized control of it have bought more than 6 million dollars of stock on the open market, paying up from 4 dollars to over 8, and have not sold a single share. The entire new board bought alongside them.

That is the trade in one paragraph. The market is still pricing a failed Bitcoin miner. The insiders are pricing the power. We are siding with the insiders, and we have started buying.

This is Part 1. It lays out what the company is, how it got here, what the asset is worth, and the one event that closes the gap. Parts to follow will track the catalysts in real time.

How we got here

You cannot understand the opportunity without the wreckage that created it.

The company used to be Mawson Infrastructure, ticker MIGI, and it was one of the worst-run names in the public Bitcoin-mining sector. While direct peers compounded more than 600 percent on the shift to AI infrastructure, Mawson destroyed roughly 97 percent of its value. Management churned. Creditors filed an involuntary bankruptcy petition. The stock was staring at a Nasdaq delisting.

Then an investor group called Endeavor started buying. Through late 2025 it accumulated stock and issued shareholder letters arguing the obvious: the infrastructure was worth far more than the equity, and the problem was leadership, not assets. By April 2026 the campaign was over. Endeavor had won board control, the prior CEO was gone, five new directors were seated, the company was renamed Big Digital Energy, and the company set itself on a path back into Nasdaq compliance. The board has since retired the takeover poison pill, now that the new owners are the ones in control. The new team pointed the company at the highest-value use of its core asset: low-cost, energized power for AI and high-performance compute.

The stock has already moved off the lows, from a roughly 16 million dollar market cap to roughly 40 million. We think that is the first leg of a much larger re-rate.

What you are actually buying

Strip away the ticker change and the narrative, and BGDE is a power-first infrastructure operator with a real, verifiable footprint. From the company's own SEC filings, not a pitch deck:

The flagship is Midland, one of the largest sites in PJM operated by a public miner, with roughly 38,810 machine positions and carbon-free power supplied under a fixed-price agreement with Energy Harbor, owner of the adjacent Beaver Valley nuclear station. Bellefonte is a smaller complementary site, and Corning is an additional development parcel that carries optionality the market ignores. A second 250-megawatt substation already sits on the site, currently offline, next to an option on 200 adjacent acres. That is development upside, not operating capacity, and none of it is counted in the 129-megawatt figure.

This is not a shell. The infrastructure is real enough that the company runs production data-center management tooling and live energy metering across its sites. You are buying a genuine operator, not a story.

The power economics

Here is the part the market is not modeling. BGDE buys power under a fixed-price contract at a blended cost in the low 40s of dollars per megawatt-hour, roughly 4 cents per kilowatt-hour, while the 2025 PJM all-in wholesale price ran closer to 79 dollars per megawatt-hour. That spread is the entire business model: control cheap, contracted power in a market where power is repricing violently upward.

It monetizes that position two ways today. First, hosting and colocation, renting capacity and racks to customers. Second, and more interesting, energy management: getting paid by the grid to curtail and provide demand response when PJM is short. That second line is small relative to what it becomes if the company converts capacity to AI, but it is real, it is growing, and it is the proof that scarce power is a cash-generating asset, not just a cost center.

The business already generates real revenue

The bear reflex on any ex-miner is "no revenue." That is wrong here. In 2025 the company produced roughly 38 to 40 million dollars of revenue across two real lines: digital colocation at about 26 million, and energy management at 11.8 million, up 56 percent year over year.

That energy-management line is the important one. It is how a power-rich operator turns grid scarcity into cash. As power gets scarcer, that line gets more valuable, and it is already the second-largest revenue source in the company. And the platform has shown it can monetize the assets themselves: it previously sold its Sandersville, Georgia site to CleanSpark for up to 42.5 million dollars, more than the company's entire market cap going into this takeover.

The macro: power is the bottleneck, and BGDE sits on it

The AI build-out is not constrained by chips or capital. It is constrained by energized megawatts with grid interconnection already in place, the thing that takes years to permit and build. Nowhere is that tighter than PJM, and the price proves it.

PJM capacity cleared at 28.92 dollars per megawatt-day for 2024 to 2025. It then cleared 269.92, then 329.17, and the most recent auction hit 333.44, the regulatory cap, the first time the entire grid fell short of its reliability requirement. Data centers drove the majority of that move, and PJM is warning of a shortfall of up to 60 gigawatts over the decade. BGDE controls a scarce, grid-connected position in the exact market where everyone now needs power and nobody can get it fast.

The tell: the people who know the most are buying the most

Cheap is not a catalyst. Aligned owners buying their own stock is. And the insider record here is as one-directional as we have ever underwritten.

From late November 2025 through June 2026, the controlling group bought stock almost continuously. They did not buy the bottom and stop. They paid 4.07, then 4.84, then 5.37, then 6.01, then 8.38, and on one print 12.00. They paid up into their own thesis. Then, in June, the entire newly seated board followed with personal cash in the high 6 to mid 7 range. Total open-market insider buying runs past 6 million dollars. Real insider selling is zero, into a float of under 6 million shares. That is accumulation by the most-informed buyers in the building.

The prize: AI conversion

This is where the asymmetry lives. The same megawatt earns very different money depending on what runs on it. Bitcoin and legacy hosting generate roughly 20 to 45 million dollars of revenue per 100 megawatts at mining-grade margins. The publicly disclosed AI and HPC colocation contracts that former miners have signed imply 140 to 225 million dollars per 100 megawatts at 40 to 65 percent site-level margins.

This is not theoretical. Former miners have signed enormous, long-dated AI contracts that re-rated their entire equities. These are the precedents the market will use to price BGDE the day it signs its first deal.

Be clear-eyed: BGDE has not signed an AI offtake yet. That is the un-hit catalyst, and it is exactly what the new team was installed to deliver. But that is also why you are getting 129 megawatts at a price that assumes it never happens.

What is $BGDE worth?

We value this on the asset, because that is what a re-rate prices. Today BGDE trades around 0.3 to 0.5 million dollars per megawatt. The miners that have signed AI contracts trade at 11 to 30 million per contracted megawatt. That gap is the opportunity, and the risk. On the company's own platform math the existing assets carry a stated net asset value near 75 million dollars, so the stock changes hands at roughly half of that book value before any AI conversion is priced in.

The variable is dollars per megawatt, and the swing factor is whether the company converts capacity into contracted AI compute.

• Base case, roughly 18 dollars. The miner-plus-demand-response business stabilizes, the market gives partial credit for 129 megawatts of PJM power, and the stock re-rates toward 1.5 to 2 million per megawatt. Roughly 2.5 times from here, and that is the case where the AI thesis never even lands.

• Bull case, roughly 55 dollars. The team signs its first real AI or HPC offtake, even 30 to 50 megawatts, and the contracted capacity reprices toward 5 to 7 million per megawatt, still a steep discount to peers for a first contract. Roughly 7 to 8 times, after the shares issued to fund the conversion.

• Moonshot, 150 dollars and up. Multiple sites convert to executed contracts and the platform re-rates toward the peer band of 11 to 30 million per megawatt. Twenty times and up. This is the home run, and it requires the team to execute against the clock, but it is the deck's own platform math, not a fantasy.

Against a price around 7 dollars, the base case is a winner and the bull case is a multi-bagger. The targets are aggressive on purpose. They are also anchored to the asset and to peer multiples, not to hope.

The team

The thesis rests on whether this group can do the one thing the old regime never did: convert power into contracted compute. The control group, operating as Endeavor and SixThirty.AI, are the chairman, chief executive, and operating chief, and they collectively own about 29 percent of the company. Their backgrounds run across energy, AI and HPC infrastructure, real estate, and operations, and one built and sold a cybersecurity company. The most important fact about them is not a resume line. It is that they have bought more than 6 million dollars of stock with their own money and put it under multi-year performance milestones. Their outcome is the public shareholder's outcome.

What are the catalysts?

This is an event-driven setup. The things that re-rate it, in order of importance:

1. The first executed AI or HPC offtake with a named compute partner. This is the single re-rating event. It converts a few-hundred-thousand-per-megawatt valuation into a contracted multiple overnight.

2. A new long-dated power contract replacing the December 2026 agreement at a still-competitive rate, defusing the cost-reset clock.

3. Refinancing or terming-out the near-term loans, removing the going-concern overhang.

4. A clean, independently priced combination of the controlled development pipeline into the public company, turning the platform story into public-company megawatts.

5. Litigation resolution at de minimis cost, clearing the last legacy overhangs.

What are the risks?

We are bullish, not blind. At any real position size you hold these in view.

• The cheap-power agreement reprices at the end of 2026. Midland's fixed-price contract runs through December 2026; beyond it, power costs move toward the PJM market. A new long-dated contract is the catalyst we are watching, and its absence is the clock on the cost advantage.

• Funding and dilution. The balance sheet carries a going-concern note and roughly 26.5 million dollars of near-term loans, and the funding answer is mostly debt rather than fresh stock. The executive chairman's own entity has extended the company a secured revolving credit line of up to 40 million dollars, while the public at-the-market equity program is capped at just 9.6 million, which holds near-term dilution down. Two items inside that stack are worth watching: a defaulted legacy promissory note with about 11.1 million dollars outstanding that now sits in active litigation, and a Nasdaq condition that requires at least 5 million dollars of stockholders' equity in every quarter through the middle of 2027. The value per share still has to outrun the share count, and so far the insiders are buying faster than the company is issuing.

• No AI contract is signed yet. The thesis lives or dies on the first real offtake. A prior 2024 attempt did not convert. This is a bet that this team gets it done.

• Related-party structure. Part of the growth pipeline and the flagship colocation deal involve entities affiliated with the same principals who control the company, reviewed under a stated independent-committee framework with the conflicted principals recused. The alignment here is structural, not cosmetic. The affiliate's upside on the colocation deal is paid in warrants struck at 20 dollars, nearly three times the current price, so those insiders only monetize that piece by driving the stock to 20 and above. It still deserves ongoing scrutiny, and the offset is that the same principals hold common stock alongside outside shareholders.

None of these is disqualifying. All of them are why a 129-megawatt PJM asset is still available at a fraction of replacement cost, and why we treat this as a high-conviction, defined-risk position rather than a sure thing.

Is $BGDE a buy? The bottom line

Big Digital Energy is a real 129-megawatt power operator in the most power-starved grid in America, with a demand-response revenue line growing 56 percent, nuclear-adjacent power, and an activist team that put more than 6 million dollars of its own money in from 4 to 8 dollars with zero selling. It trades at a few hundred thousand dollars per megawatt while contracted peers trade in the tens of millions. The one thing it has not done, sign an AI offtake, is the one thing that closes that gap, and it is the catalyst we are positioned for.

We are buyers. This is the kind of asymmetry the portfolio exists to find.

Disclosure

Shawarma Capital holds a position in BGDE and intends to add to it. This is research and opinion, not investment advice. Microcap stocks are volatile and can go to zero. Do your own work and size accordingly.