Payloop reported 10 investors and $2M left to sell; its filing names Adam Struck and Moonshots Capital's Kelly Perdew as directors.
By Ryan Merket · Published
Primary source: U.S. Securities and Exchange Commission
Why it matters #
AI turns software delivery into a variable-cost business. Payloop is betting finance and product teams will pay to trace model costs to each customer, workflow and agent before pricing erodes gross margin.
John Kim, co-founder and CEO of Payloop, has sold $4,533,992 in equity for an Austin-based software business built to show AI developers whether their customers, agents and features are profitable.
Payloop disclosed the sale in a Form D filed with the Securities and Exchange Commission on August 24th. The offering has a target of $6,533,992, leaving $2 million unsold. Payloop reported that the first sale occurred on August 10th and that 10 investors had participated.
The filing does not identify those investors or disclose Payloop's valuation. It names Kim as an executive officer and director, alongside directors Adam B. Struck and Kelly Perdew. Their board positions do not establish that Struck Capital or Perdew's Moonshots Capital participated in the offering.
Payloop reported no commissions or finder fees and said none of the offering's proceeds were designated for payments to the executives and directors named in the filing. Payloop elected not to disclose a revenue range. The securities were sold under Rule 506(b), which allows a private issuer to raise an unlimited amount from accredited investors and a limited number of sophisticated non-accredited investors without a public registration.
Kim is taking quote-to-cash into the AI bill
Kim's wager on AI unit economics follows a career spent around the machinery companies use to price, sell and bill for software. In an October 2025 account of Payloop's formation, he wrote that he entered technology during the dot-com era, selling colocation, managed servers and private-network infrastructure in the Bay Area.
He later built and led teams at BigMachines, which Oracle acquired, and SteelBrick, which Salesforce acquired and turned into part of its configure-price-quote business. Kim also worked at product-visualization software provider Threekit. That history matters to Payloop's pitch: Kim is treating AI costs as a pricing and billing problem, rather than another model-development problem.
Kim said Payloop emerged from conversations with Struck Studio, the venture-building operation connected to Struck Capital. Its research across portfolio companies and prospective investments had found that businesses lacked cost visibility at the customer, agent and action level. Kim and Struck Studio began building Payloop around the idea that those figures could guide pricing, product design and customer-level profitability.
Struck, Payloop's director, founded Struck Capital after working as an M&A and private-equity attorney at Kirkland & Ellis. He previously co-founded Long Island Brand Beverages, which was acquired by Cullen Investments and taken public on Nasdaq, according to his firm biography.
Perdew brings a separate operator-investor network to Payloop's board. Moonshots Capital identifies him as its co-founder and managing general partner. He graduated from West Point, served as an Army intelligence officer and later earned JD and MBA degrees from UCLA. Perdew also won the second season of NBC's "The Apprentice" in 2004, a biographical detail that tends to overshadow a longer record of operating and investing in early-stage technology businesses.
Metering AI at the customer level
Payloop's product uses a software development kit to wrap calls to large language models and collect token usage, model selection, timestamps and cost data. Payloop says customers can then attribute spending to individual tasks, workflows, agents, features and accounts.
That attribution is the core distinction Payloop is trying to establish. LLM observability products commonly show developers what a model did, how long it took and how many tokens it consumed. Payloop is positioning its platform closer to a finance and pricing layer, connecting those events to gross margin and letting users simulate token-based, cost-plus, subscription and outcome-based pricing.
An AWS Marketplace listing offers Payloop under a 12-month contract priced at $5,000 for up to 1 billion processed tokens. The listing says Payloop can compare models by cost, latency and performance, giving developers a way to choose cheaper models for specific workflows instead of applying one model across an entire product.
Payloop also markets Sentinel, a security component that it says blocks prompt-injection attempts, jailbreaks and off-topic requests. The commercial logic is straightforward: a malicious or irrelevant request is also a variable expense when every prompt, tool call and response consumes paid compute.
Payloop's savings claims have increased as its marketing has developed. In Kim's 2025 announcement of Payloop, he said the product could reduce costs by up to 50%. The AWS listing now says reductions can reach 65%. Both figures are Payloop's assertions, and the Form D provides no operating metrics that would substantiate them.
The margin problem grows with every agent action
Traditional subscription software gave operators a relatively stable infrastructure cost against recurring revenue. AI applications can produce a new model charge every time a customer sends a request, an agent retries a task or a workflow calls another tool. The cost of serving two customers on the same plan can diverge sharply based on how they use the product.
That creates pressure on the metrics used to run software businesses. Aggregate cloud spending can show that costs are rising without identifying which customers or features caused the increase. Token dashboards can quantify consumption without showing whether the revenue attached to that usage covers the bill. Payloop is trying to occupy the control point between those two systems.
The remaining question is distribution. Developers already have tracing, evaluation and observability products embedded in their AI stacks, while cloud providers and model platforms have their own usage dashboards. Payloop has to persuade product and finance organizations that customer-level margin intelligence deserves another integration and another software budget.
The $4.5 million already sold gives Kim capital to pursue that argument while Payloop seeks the remaining $2 million in its offering. His advantage is a career spent on software monetization during earlier shifts in enterprise infrastructure. Payloop's challenge is converting that experience into a system companies will treat as a financial source of truth, rather than another dashboard beside the model bill.