WhiteFiber, Inc.

WhiteFiber builds and operates high-performance-computing data centers and rents GPU cloud compute, selling the infrastructure that AI training and inference run on rather than AI models themselves.

Ticker

WYFI

Company

WhiteFiber, Inc.

Exchange

Nasdaq

Listing

listed

Sector

AI Infrastructure / Neocloud (HPC data centers + GPU cloud)

On X, $WYFI is framed as a picks-and-shovels neocloud play: instead of betting on which AI model wins, you own the data-center and GPU-cloud capacity every model needs. The bull case leans on a long-dated contracted backlog underwriting revenue that most newly public compute names lack.

• WhiteFiber is listed on Nasdaq under the ticker WYFI and files reports with the U.S. SEC.S1
• WhiteFiber operates two lines of business: HPC data-center colocation and a GPU cloud-services business, providing AI compute infrastructure rather than AI models or applications.S2
• For fiscal year 2025 (ended December 31, 2025) WhiteFiber reported revenue of about $79.2 million and a net loss of about $24.7 million.S2
• As of December 31, 2025 WhiteFiber held about $114.4 million in cash and cash equivalents (plus about $3.9 million of restricted cash), with no funded debt outstanding.S2
• WhiteFiber has signed a 10-year colocation agreement with Nscale at its NC-1 site representing approximately $865 million of total contracted revenue, with billing expected to begin in the second quarter of 2026.S2

• narrativeBacklog-as-moat framing: a 10-year, roughly $865 million contracted deal reads as revenue visibility most early-stage compute names do not have. Bulls treat it as de-risked revenue; bears note it is concentrated in one customer and not yet billing.
• narrativeNeocloud beta: as a recently IPO'd, single-theme AI-infrastructure name, it tends to trade as a high-beta proxy for AI-capex sentiment rather than on current earnings.

• ▸WhiteFiber is a newly public, capital-intensive company that ran a net loss in 2025 and leans on a small number of large customers and contracts plus continued AI-infrastructure demand. Build-out execution, customer concentration, financing and competition all apply. Confirm specifics against the 10-K.S2

Colocation

— Renting space, power and cooling in a data center so a customer can run their own servers there.

GPU cloud

— Renting access to graphics-processor (GPU) compute over the internet — the hardware AI training and inference actually run on.

Neocloud

— A newer kind of cloud provider focused on renting out AI/GPU compute, as opposed to the broad general-purpose clouds like AWS, Azure or Google Cloud.

Contracted backlog

— Revenue a company has signed contracts for but has not yet earned or billed.