Future Tech & AI Wonders · Jordan Lee · 27 August 2026

JPMorgan flags risk as IREN earnings test AI buildout

JPMorgan flags risk as IREN earnings test AI buildout

JPMorgan has maintained a Sell rating on IREN ahead of fiscal fourth-quarter earnings, warning that customer contracts may not keep pace with the company's massive GPU expansion and financing needs. With IREN earnings due after the Aug. 27 close, investors will watch whether contracted AI Cloud revenue can de-risk one of the market's most aggressive AI infrastructure buildouts.

Key Takeaways

Why Is JPMorgan Cautious Before IREN Earnings?

IREN, the former Bitcoin miner now racing to become a major AI infrastructure provider, reports fiscal fourth-quarter and full-year 2026 results after the market close on Aug. 27, with a 5 p.m. ET call to follow. JPMorgan analyst Richard Choe's concern is straightforward: IREN is building GPU capacity faster than demand is becoming contractually committed.

In July, IREN signed $2.8 billion of additional contracts with AI developers and lifted its targeted year-end AI Cloud annualized run-rate revenue to more than $4 billion. Roughly 85% of that targeted ARR was under contract at the time—including a landmark $9.7 billion, five-year Microsoft deal for 200 megawatts of NVIDIA GB300 GPU infrastructure. Yet capacity still needs customers, which explains the bank's caution heading into earnings.

What Should Investors Watch in the IREN Earnings Report?

Wall Street enters the release expecting revenue of about $135.6 million to $138.9 million and an adjusted loss of $0.42 to $0.46 per share—a sharp swing from positive EPS a year ago. That reflects the pivot away from Bitcoin mining: fiscal Q3 revenue fell to $144.8 million as mining revenue dropped to $111.2 million, while AI Cloud revenue reached $33.6 million, nearly doubling sequentially.

The real test is operational, not headline EPS. Investors should track contracted AI Cloud ARR, GPU deployment schedules toward a 150,000-GPU fleet, utilization rates, capital spending, and any new financing needs. IREN secured more than $9.3 billion in funding in March and later a $3.65 billion GPU financing facility for its Microsoft contract, with roughly $3.5 billion in additional capex expected for those orders.

Can IREN's AI Buildout Justify Its Premium Valuation?

IREN trades at roughly 4.23 times forward 12-month revenue, above a cited industry average of 2.59 times, despite remaining loss-making. The stock has returned about 11.8% year to date—outperforming its broader industry but lagging several AI-infrastructure peers. The premium depends on hitting the $4 billion ARR target and bringing capacity online at its 750-megawatt Childress, Texas campus.

Management ended the period with $7.6 billion in cash, including $1.7 billion in restricted cash tied to customer prepayments, and issued $1 billion in zero-coupon convertible bonds plus $200 million in equipment financing. Partnerships with Together AI, Fluidstack, and Fireworks AI add pipeline, but delays in chip deliveries, weaker utilization, or rising electricity costs could strain the timeline to deploy 140,000 active GPUs.

What Would Confirm—or Undermine—the Bull Case?

The strongest rebuttal to JPMorgan would be another increase in contracted capacity without significantly greater balance-sheet risk. Conversely, slower customer signings, deployment delays, or additional financing needs would reinforce fears that IREN is building ahead of demand. For broader context on AI infrastructure pivots, see our Future Tech & AI Wonders coverage.

As GuruFocus reported via TradingView, Thursday's results should be judged on whether IREN can de-risk its AI expansion—not on a single quarter's bottom line. Guidance on megawatt capacity, prepayment inflows, and the path toward 1,210 megawatts of power will determine whether contracted AI demand is finally turning into lasting growth.

← Open in blast feed