Auddia Highlights LT350 AI Infrastructure Opportunity

Auddia Inc. issued a GlobeNewswire release on May 18, 2026, highlighting market pricing disclosed in a SharonAI SEC filing. According to Auddia's release, SharonAI disclosed a $1.25 billion GPU cloud contract spanning 60 months, which Auddia reports implies about $30,488 in annual revenue per high-performance GPU. The release also reiterates prior company disclosures that LT350's REIT partnership controls roughly 4,000,000 sq ft of parking-lot airspace and could support up to 960,000 GPUs if fully deployed. The release contains a direct quote from Jeff Thramann, CEO of Auddia: "We are not providing forecasts or forward revenue projections outside of what is disclosed in our recent S-4 financing," said Jeff Thramann, CEO of Auddia and Founder of LT350.

Industry-pattern observations: public, contract-level pricing datapoints such as a large multi-year GPU cloud agreement provide a concrete revenue-per-GPU benchmark practitioners use to compare economics across deployment models. For infrastructure teams and financial modelers, a $30,488/GPU-year figure can be compared against known cost inputs: GPU acquisition or lease costs, data-center power and cooling, colocated power delivery, and utilization rates. Distributed or edge-adjacent designs that reuse parking-lot airspace and canopy modules trade off site accessibility, power distribution complexity, and thermal management versus proximity to regulated customers.

Editorial analysis: REIT-backed or real-estate-centric approaches to hosting AI compute have surfaced as a capital-efficient route to scale physical footprint, but they introduce operational variables that materially affect realized revenue per GPU. These variables include time-to-deploy, local permitting and grid interconnection, battery and resilience architecture (the release mentions a 2:1 GPU-to-battery cartridge ratio and 2,000-square-foot modules in prior disclosures), and achievable utilization. Public benchmark contracts inform investor expectations, but realized cash flows will depend on conversion from prospective rooftop/parking airspace into fully commissioned, high-utilization clusters.

• •SharonAI's full SEC filing for granular contract details, including disclosed GPU count, model families, and pricing schedule, which determine the robustness of the $30,488 per-GPU-year derivation.
• •Auddia's S-4 and subsequent filings for any quantified revenue guidance or technical deployment timelines tied to LT350.
• •Proof-of-concept or pilot deployments that disclose power density, GPU model types, and sustained utilization, which will validate or compress benchmarked revenue assumptions.
• •Local permitting and grid interconnection outcomes at pilot REIT properties, which affect build-out pace and unit economics.

Editorial analysis: For practitioners, the immediate value of the story is the addition of a public pricing datapoint that can be used alongside cloud list, reserved, and colo pricing to stress-test total-cost-of-ownership and revenue models for distributed GPU infrastructure. This does not, by itself, establish realized revenue for LT350; observed deployment metrics and contract wins will determine whether a theoretical footprint translates to operating scale.