Super Micro Announces $39B AI Server Orders, Prices Equity Offerings

Super Micro Computer disclosed on June 9, 2026 that it had received roughly $39 billion in new AI server orders from more than 20 unnamed customers in recent weeks, a figure that would represent a transformational pipeline for one of the largest AI infrastructure suppliers . To fund the component purchases needed to fulfill the backlog, the company simultaneously priced a $7 billion equity and equity-linked financing package, triggering a sharp selloff that erased more than a quarter of the stock's value in a single session .

The financing has three tranches: approximately $1.25 billion in new common stock, approximately $3.75 billion in depositary shares representing newly issued Series A mandatory convertible preferred stock (each with a $1,000 liquidation preference and a June 2029 conversion date), and an at-the-market program of up to $2 billion in common stock set to begin no earlier than Q3 2026. J.P. Morgan, Goldman Sachs, and Citigroup are lead underwriters. The net proceeds are earmarked for component procurement, with any remainder available for working capital, capital expenditures, and debt repayment.

Investors reacted with skepticism on two fronts. First, Super Micro's own prospectus disclosed that the $39 billion in orders "do not constitute firm commitments and are all subject to cancellation," making the order size a pipeline estimate rather than a binding backlog . Second, the combined offerings represent significant dilution to existing shareholders at a moment when the company's gross margin has come under pressure from component-intensive AI server configurations. SMCI shares fell 7.6% in after-hours trading on June 9, then dropped a further 19.7% intraday on June 10 to close at approximately $29.27, a one-day decline of roughly 28%.

The announcement lands against a backdrop of renewed scrutiny for SMCI. The company only recently resolved a prolonged accounting and auditor controversy that had threatened its Nasdaq listing in 2024. Bears argue the non-binding order caveat, combined with the scale of dilution, echoes prior episodes where headline metrics overstated near-term certainty. Bulls counter that the breadth of the customer base spanning 20-plus buyers suggests real demand from hyperscalers and enterprise AI adopters, and that securing component supply now positions the company ahead of competitors reliant on scarce NVDA GPU allocations.