CodeBoxx Technology, which has offices in St. Petersburg, Florida, and Quebec City, is now taking orders for CrewKit Forge, an AI software factory that companies own outright and install in their own buildings. The self-contained appliance, a one-meter cube (about 3.3 feet per side), runs the equivalent workload of up to 20 AI-equipped software development professionals in parallel on local AI models, so routine code work never leaves the premises. CodeBoxx unveiled the CrewKit Forge appliance on October 22 at the Titan 100 Tampa Bay awards gala at the Yuengling Center, with some 500 guests. The same evening, founder and CEO Nicolas Genest was honored as one of the region's top 100 CEOs and C-level executives for the second year in a row.

Titan 100 recognizes Tampa Bay's top executives each year for their leadership, vision and passion. The back-to-back honor highlights an entrepreneur who builds and deploys technology teams at the service of entrepreneurs and business leaders.

A software factory the company owns

The CodeBoxx appliance, which runs the CrewKit Forge operating system, combines compute, storage and networking, and seamlessly orchestrates inference across local AI models, private open-weight models and frontier models. A true autonomous data center, it integrates its own control plane in a single one-meter cube of black aluminum and smoked glass. It can sit on a rolling Power Core Table that carries its own power conditioning, UPS and batteries, so it can be placed in a boardroom, an executive office or an open-plan office rather than a data center. Five touch screens (three 32-inch panels and a two-screen control console) show company leaders in real time where the work happens and how it performs. Its agents run the equivalent of up to 20 governed human workloads in parallel across the core disciplines of software delivery: digital product, system architecture, design, testing, cybersecurity, monitoring, compliance and documentation.

The reference configuration pairs a 96-core processor with inference gateways powered by NVIDIA, 1 TB of RAM and 12.8 TB of Gen5 NVMe storage. Specifications are subject to constant improvement and change as units ship.

The CodeBoxx Forge is local-first. Repository work, builds, code scaffolding, tests, code review and monitoring run on the box. When a job needs frontier-scale reasoning or a very long context, CrewKit escalates it to its private AI Cloud through a managed inference gateway and logs the decision; no code or data leaves the building without a recorded decision. People stay in charge of ideation, requirements, defining success, judgment and approach at every iteration. For organizations with confidentiality or compliance requirements, CrewKit Forge offers a fixed, owned asset in place of a per-token cloud AI bill.

"AI shouldn't replace human action. It should multiply it. With Forge, companies stop renting their intelligence by the token. They own their software factory, they govern it, and they stay in full control of what it produces," said Nicolas Genest, Founder and CEO of CodeBoxx.

Ready for the Quantum Computing era

CrewKit Forge is built for the quantum computing era. An optional quantum gateway lets enterprise software developers develop, simulate and rehearse their quantum algorithms on the appliance before running them on actual qubits. Teams can build quantum skills and prepare their code today, then move to quantum hardware when their use cases are ready.

Built on results already delivered

CodeBoxx built CrewKit Forge inside its own software delivery business, where it performed a remarkable pivot to applied artificial intelligence. Across seven client platforms, CodeBoxx estimates the factory delivered about $14.9 million in savings compared with a traditional 2022 development team, and in four benchmark studies it averaged 5.2 times faster delivery at 78% lower cost. Actual figures come from billed amounts and delivery records; the traditional-team comparisons are CodeBoxx estimates. Methodology is published at codeboxx.ai/crewkit-forge20/dive-deeper.