Go.AI says it raised $85M to put auditable AI inside banks' walls
Updata Partners led Go.AI's round to expand on-premises AI hardware and software; its Chicago founders met as accounting professor and student.
By RuntimeWire Staff · Published
Primary source: Business Wire
Why it matters
Go.AI's $85M round backs a hardware-and-software route to private AI. According to [Go.AI's Series A announcement](https://go.ai/company/press/series-a-funding?ref=runtimewire), it has more than 200 customers across financial services, healthcare, aerospace and defense, and manufacturing, and plans to expand beyond regulated industries into a broader set of compliance-minded organizations.
David Moscatelli and Lisa Gillespie said their Chicago company, Go.AI, raised an $85 million Series A to expand AI systems that run inside customers' own facilities. The September 22nd announcement named Updata Partners as lead investor, with existing backers GFT Ventures and LAUNCH participating. Go.AI says the financing brings its total funding to $90 million.
The founders' working relationship began in an accounting classroom. Gillespie taught Moscatelli at Loyola University Chicago, he wrote when discussing an earlier fundraise. He credited her insistence on rigor and accountability with shaping Go.AI's approach. Gillespie, who spent more than 20 years teaching and practicing accounting, has described her interest in tracing a process from end to end and asking, "What could go wrong here?" That question has a buyer at a bank trying to put AI to work without losing control of its records or its audit trail.
Selling the box and the controls
Go.AI's Go1 appliance, launched on March 24th under the Go Abacus name, packages computing hardware with Go.OS. Go.AI says the system runs models and indexes information within a customer's environment. The Go.OS product page describes local inference by default with optional cloud routing for selected tasks; Go.AI's banking page describes fully on-premises deployments with no cloud fallback or external API calls. Go.AI's pitch pairs local inference with access controls and records of AI activity that an institution can inspect. The appliance is offered on a fixed-fee basis rather than a per-token meter, according to Go.AI and Updata general partner Carter Griffin. Those are consequential purchasing terms for an institution weighing an external API against building and maintaining its own AI stack.
The original customer problem was narrower than the new name suggests. Moscatelli's account of the rebrand says he and Gillespie founded Abacus Analytics LLP in 2018 to bring AI to regulated industries, then became Go Abacus Corporation in 2022. A November 2025 funding announcement described Go Abacus as founded in 2022. The later history identifies that year as the corporate transition. Go Abacus announced a $5 million seed round that November, led by GFT Ventures with BankTech Ventures and LAUNCH participating. Go.AI adopted its current name on September 1st, three weeks before the Series A announcement.
Go.AI is also widening its sales pitch. The Series A announcement says Go.AI will hire engineers, develop Go.OS and its hardware line, and pursue organizations beyond its initial regulated-industry focus. Go.AI says it has more than 50 team members and more than 200 customers, including deployments in financial services, healthcare, aerospace and defense, and manufacturing. Those customer and staffing counts come from Go.AI; the announcement does not break the customer total down by industry or product.
The growth figures need a baseline
Go.AI says annual recurring revenue grew more than eightfold year over year and that it remains profitable. Neither figure in the announcement comes with an ARR amount or a definition of profitability. Go.AI also says customer deployments handle more than 12.5 million queries a day, while its homepage displays 8 million questions a day. The pages do not supply matching measurement dates or definitions, so the two usage figures cannot establish a growth rate. Go.AI did not announce a valuation, leaving the price paid for an ownership stake outside the public terms of the round.
Go.AI is pursuing a market with other ways to keep enterprise AI under customer control. Armor's Sovereign AI emphasizes a governed work platform; Datasaur's Forge puts engineers inside organizations to build private AI systems that customers own after the engagement. Their approaches differ from Go.AI's packaged appliance, though all address the same procurement concern: where sensitive information travels and who can account for an AI system's actions.
For Moscatelli and Gillespie, the $85 million round funds a demanding next step. Go.AI has to support physical hardware, software and customer deployments while selling beyond its initial regulated-industry focus. The founders have built their case around making AI inspectable by the people responsible for it. The expansion will test whether that design remains as useful across a broader set of institutions.