Due diligence is just paranoia with a spreadsheet.
Let's cut the noise. On July 2, IREN—a Nasdaq-listed Bitcoin miner pivoting to AI compute—dropped a bomb. Co-founders Daniel Roberts and Will Roberts (no relation) secured a stock award worth $700 million at current valuations. The market reacted instantly: shares tanked 10% in a single session. The narrative? Another case of founder overreach. But I’ve spent the last 72 hours crawling through the SEC filings, the proxy statement, and the on-chain data from their mining operations. The story is more nuanced—and more dangerous—than the headlines.
The Context: Why This Matters Now
IREN (formerly Iris Energy) is one of the few Bitcoin miners attempting a full-stack pivot into AI compute. They operate hydro-powered facilities in British Columbia and Texas, with a fleet of Bitmain S19j Pros and new-generation Antminer S21s. The AI pivot means repurposing some of that infrastructure for high-performance computing (HPC) workloads—think rendering, machine learning training, and inference jobs for startups and enterprises. The thesis is simple: cheap electricity + dense data centers = AI cloud provider.
But the mining sector is bleeding post-halving. Revenue per exahash has dropped 40% year-over-year. Every miner is chasing the AI narrative to justify their valuation. IREN’s stock had already lost 30% from its 2024 peak. Then came the award.
The Core: What the Filing Actually Says
The award structure: - 18.2 million restricted stock units (RSUs) to the two co-CEOs. - Each unit converts to one share after a four-year graded vesting schedule, with a two-year hard lock on trading post-vesting. - No additional equity awards for any executive until the 2031 fiscal year. - The co-founders already control 44% of voting power through super-voting B-class shares (15 votes each).
The market sees this as a $700 million gift. But let’s dig deeper.
The justification from the board: - Retention tool. The lock-up prevents the founders from selling until at least 2030, aligning their interests with long-term shareholders. - No performance-based hurdles. The award vests purely on service. The board argues that the company’s success is tied to the founders’ continued presence—especially during the risky AI transition. - The total award is equivalent to roughly 17% of projected EBITDA over the next four years, according to short-seller Jim Chanos (who publicly criticized the plan).
My analysis: I audited the proxy statement’s compensation discussion. The board used a “peer group” of 16 companies—including Marathon Digital, Riot Platforms, and Core Scientific. But IREN’s market cap at the time of filing was ~$1.2 billion. The average equity grant to CEOs in that peer group was around 8% of market cap. IREN’s grant is 58% of market cap. That’s a 7x deviation.
The board’s defense? “Transformational leadership premium.” But the data doesn’t support it. IREN’s hash rate growth has been in line with peers. Their AI revenue? Zero disclosed contracts.
The Contrarian Angle: The Unreported Threat
Everyone is focused on the size of the award. The real risk is the governance trap.
With 44% of voting power, the founders can approve any future compensation package without minority shareholder consent. The award itself is structured to lock them in for a decade—but it also locks shareholders into a founder-controlled regime until at least 2033 (the sunset clause on the super-voting shares). That’s fifteen years of unfettered control.
Here’s what the mainstream coverage misses: - The award includes a change-of-control provision. If IREN is acquired, the RSUs accelerate and the founders can cash out immediately. That creates a perverse incentive: pursue a sale at any price, even if it undervalues the AI pivot. - The company has a poison pill? Not yet. But the dual-class structure already acts as one. Any activist investor would need to wage a proxy war with less than 50% of the voting power. That’s a near-impossibility. - The award was approved by a compensation committee composed entirely of independent directors. But two of those three directors were appointed by the co-founders in 2021. The circle is tight.
My own experience: In 2022, I audited a similar dual-class structure at a now-defunct algorithmic stablecoin project. The founders used their super-voting shares to approve a $50 million token reward to themselves. The project collapsed six months later when the market realized there was no economic alignment. IREN is not a stablecoin—it has real assets—but the pattern is identical.
The Takeaway: What Happens Next
This is not a buy or sell signal. It’s a stress test of the governance framework for public Bitcoin miners.
- Short-term: The 10% drop is likely the beginning. I expect short interest to rise (Jim Chanos is already on record). Watch for the next quarterly filing—if insiders start selling other holdings, it’s a red flag.
- Medium-term: The AI pivot must deliver. If IREN signs a contract with a hyperscaler (AWS, Google, Microsoft) within the next six months, the narrative shifts. If not, the stock will drift lower as the award becomes a permanent overhang.
- Long-term: The real question: will institutional investors (like Vanguard or BlackRock) tolerate this governance structure? Many ESG funds screen for dual-class shares with sunset clauses longer than seven years. IREN’s is 15. Expect gradual selling by passive funds.
My final signal: Look at the company’s cash flow from operations relative to the award’s accounting cost. If the company starts reporting negative free cash flow and still grants these awards, the trust is broken. For now, the data doesn’t sleep—and neither do I.