Sharon AI’s $8.8 Billion Backlog Faces a Revenue Test

Sharon AI’s central challenge is no longer simply lining up customers. It is turning a vast slate of commitments into working infrastructure, billable computing time and, eventually, cash from operations.

The Nasdaq-listed Australian cloud provider reported $8.8 billion in total contract value as of August 6, alongside $1.9 million in second-quarter revenue. That figure was up 412% from a year earlier, but the gap between signed commitments and sales underscores how early the build-out remains. The company says revenue should ramp from the third quarter of 2026 through 2027.

A recent customer deployment offers one early milestone. Sharon AI said a global technology customer accepted the first phase of a five-year agreement valued at about $950 million, releasing cash security held in escrow. Revenue under that contract is expected to begin in stages across the third and fourth quarters. Acceptance is progress; the coming revenue reports will show how quickly it translates into recognized sales.

The infrastructure plan is expansive. Sharon AI reported 212 megawatts of secured AI Factory capacity in its second-quarter release, and said its $4.9 billion collaboration with NVIDIA covers up to 40,000 GB300 GPUs. In a September 4 filing, it described a five-year Rafay Systems agreement to standardize how it provisions, monitors and governs compute across sites. The platform is designed to support orchestration of as many as 150,000 GPUs over the agreement’s term, a long-range target rather than installed capacity today.

That software layer may help manage a sprawling fleet, but it does not remove the execution burden: equipment and data-center capacity have to come online, run reliably and stay occupied. Sharon AI’s $1.9 billion in cash at June 30 provides substantial funding for near-term expansion, though the company’s small revenue base means investors still have little operating history with which to judge the economics at scale.

Accounting also complicates the headline loss. Sharon AI posted a $430.4 million second-quarter net loss, including $423.8 million in non-cash items, chiefly a $400.4 million fair-value loss on convertible notes linked to the share-price rise. Adjusted EBITDA was positive at $0.6 million. Neither number settles whether the business can earn attractive returns once the build-out costs hit.

On September 23, Sharon AI announced a VAST Data collaboration to offer DataEnclave, a confidential-computing service, across its platform. It adds a product pitch around data privacy. The harder proof will be deployments, recurring revenue and cash generation, not another large commitment.