SharonAI Holdings, Inc. Class A Common Stock

Management did not report quarterly revenue, EPS, or gross margin on the call. It said the company has raised approximately $2.2 billion of capital since December 2025, including a $1.6 billion oversubscribed financing in June and a $350 million convertible note in April. Secured AI factory capacity was updated to 212 megawatts, up from 132 megawatts previously, with 120 megawatts contracted and roughly $8.8 billion of total contracted value year-to-date. Forward guidance calls for first material revenue in Q4 2026, more than 64,000 GPUs deployed by mid-2027, and contracted visibility extending through 2031.

James Manning framed the quarter as validation that SharonAI has increased all three inputs needed to scale: capacity, demand and capital. He said the company is built to solve scarcity in GPU access, power, data residency and capital/talent, and argued that the NVIDIA partnership and other ecosystem relationships strengthen supply certainty and market credibility. His tone was confident and upbeat, but he repeatedly stressed that execution, delivery and financing discipline still matter.

The call contained no detailed CFO financial commentary on margin, cash flow, or capital allocation because Tim Broadfoot’s departure was announced as this would be his last 10-Q. The financial framing instead came from management’s discussion of the $2.2 billion raised since December 2025, including the $1.6 billion financing in June and the $350 million convertible note in April. Management also said debt providers view the NVIDIA agreement favorably because it provides a minimum revenue floor that can support financing of GPUs and infrastructure, and that the company is seeing strong pricing and strong IRRs supported by customer demand.

Analysts pressed on the source and timing of the additional 80 megawatts, and management said it came from a new partner solution and could see some early megawatts as soon as late this year, but definitely in Q1 next year. Questions also focused on the NVIDIA-backed capacity and how much would be resold versus held back; management said it is prioritizing resale of the 40,000-GPU program now, mostly on three- to five-year take-or-pay terms, with a small spot component. On timing risk, management said the main swing factors for revenue are hardware delivery and data center readiness, and it expects a solid Q4 as B300 and GB300 deployments ramp.