A shipping container is an unromantic place to meet artificial intelligence. It has identification numbers, locks, seals and sometimes a dent that somebody will have to explain. At Britain’s Port of Tyne, Gorilla Technology’s video analytics help turn those details into information an operator can use. Cameras supply the pictures. Software searches for damage. A routine inspection becomes a record, an alert and, potentially, less time spent beside moving machinery.
That is the accessible version of Gorilla’s business: giving existing video a useful second job. The larger version now involves financing and deploying the machines on which AI runs. A company with roots in Taiwan’s video analytics industry and a London headquarters is moving into Asian data centers and GPU services. Its customers increasingly need both intelligence and the physical capacity to produce it.
The transition makes Gorilla an unusually revealing company to watch. Its products occupy the everyday world of police investigations, network monitoring and industrial safety. Its newer ambitions occupy the expensive world of equipment procurement and infrastructure finance. Between the two sits a deceptively simple question: how much must a company spend before its next contract becomes a working service?
The cameras already have a day job
The Port of Tyne project offers a concrete answer to what Gorilla actually does. Working with BT, the company supplied video analytics for container inspection and tracking. Gorilla’s case study describes cameras capturing views around containers at inspection points, with software classifying damage and flagging possible problems. The system integrates with the port’s Milestone video management software.
The original bottleneck was manual work. People had to inspect containers, check locks and seals, and keep consistent records while operating near heavy equipment. The case study describes plans to send confirmed damage reports and images into the terminal operating system in later phases. That distinction matters: an alert on a screen and an entry in the port’s operating records are separate steps.
In February 2025, Gorilla announced an expansion with BT and the port focused on safety in hazardous areas, including protective equipment compliance and unsafe behavior near machinery. The commercial logic is straightforward. A useful installation creates an opportunity to add another task using the same operational setting. The lesson for an enterprise buyer is to begin with a specific inspection that consumes time or exposes people to risk.
The year Gorilla chose a smaller business
Gorilla’s founder is Spincer Koh; the company was incorporated in 2001. Its public-market chapter began with a SPAC transaction in July 2022, bringing the Nasdaq ticker GRRR. The name supplies the growl. The accounts supply a less theatrical story about which customers are worth keeping.
In its 2022 results, Gorilla said it had discontinued unprofitable or marginal customer relationships while emphasizing security convergence. Annual revenue fell from about $42.2 million to $22.4 million. Video IoT revenue dropped particularly sharply. Management described the contraction as part of a deliberate commercial reset, alongside changes in leadership and geographic focus.
This was a problem with the economics of the customer book, according to management’s account. The response was to change what the company sold, where it sold it and which work it accepted. The public listing helped provide capital and visibility for that shift. Readers can copy the discipline of examining customer profitability; they cannot assume that removing weak accounts will automatically produce larger replacements.
Gorilla also bought capability. In January 2023 it acquired SeeQuestor’s technology assets, adding live and retrospective video intelligence. A subsequent financial disclosure described $6 million of fixed consideration and a further $3 million of contingent share consideration tied to performance. It was a purchase of intellectual property, a useful distinction from casually describing the transaction as buying an entire operating company.
Four products, two clocks
Gorilla’s video portfolio addresses two different moments: something happening now, and something an investigator needs to reconstruct later. IVAR, the Intelligent Video Analytics Recorder, processes video for people, faces, behavior, vehicles and license plates. EVMS, its Edge Video Management System, provides monitoring groups, geographical alerts, searches and dashboards that help operators use those detections.
Post Event works on the investigative clock. It helps teams import footage, search across cameras and manage a case. iCCTV serves live policing workflows, analyzing camera feeds for people, vehicles and objects. A police team facing hours of footage and a security operator watching a live incident have related needs, but different deadlines and different jobs to finish.
Local, or edge, processing is central to the proposition. Running analysis near the source can reduce the need to move all raw footage elsewhere before extracting information. Buyers still need to match the installation to their cameras, network and operational requirements. The useful question is which information must arrive in time for somebody to act.
The cybersecurity portfolio extends the pitch beyond images. NetProbe, NetTrap and host-based malware detection sit alongside a Security Convergence Platform and security operations centers. Gorilla is selling the integration of information that would otherwise arrive through separate tools. That breadth suits public agencies and enterprises dealing with physical premises, communications networks and ongoing security operations.
There are credible alternatives. BriefCam offers video search, real-time alerts and analytical insights. Genetec’s Security Center unifies functions including surveillance, access control and license plate recognition. Gorilla’s distinctive pitch is the combination of video intelligence, cyber capabilities, edge deployment and project integration. Breadth is a purchasing argument, not evidence that every component beats a specialist’s product.
A software pitch meets a financing bill
The commercial relationships help explain how Gorilla reaches customers. BT brings telecommunications and integration capabilities to the port work. In Colombia, partner Protactics supported an announced intelligence-center project in Casanare using Post Event and iCCTV. SINTRONES provides an in-vehicle computing partnership. These are routes into specific operating environments, where a software vendor needs more than an impressive demonstration.
Its infrastructure work takes that dependence on partners further. Gorilla’s July 2026 update described Indian programmes with Yotta involving an initial 5,120 GPUs and a second planned deployment of 20,736. In Indonesia, its NeutraDC Batam programme supports a signed five-year AI compute contract that Gorilla valued at approximately $2.5 billion in expected revenue. These are company-described programmes and future values, spread across deployment schedules.
Gorilla confirmed that two bond financings had raised $232 million in aggregate gross proceeds. That corporate capital supports project contributions, capacity commitments and equipment purchases. Separate project financing remains part of the plan. The arrangement illustrates why supplying compute is financially different from selling another software license: the provider must arrange expensive productive assets before the customer can use them.
Gross proceeds from two completed 2026 bond financings.
The earlier $105 million equity offering, completed in July 2025, shows the same requirement at a different scale. Gorilla said proceeds would support performance guarantees, bid bonds and acquisitions. Six million shares or equivalent instruments were offered at $17.50 each. Getting qualified to bid and promising to deliver can themselves tie up meaningful capital.
The uncomfortable arithmetic of growth
The company reported $101.4 million in 2025 revenue, up from about $74.7 million in 2024. That is evidence of a larger operating business. It does not turn the much bigger numbers attached to future projects into present-day sales. A multi-year contract, equipment deployed, customer acceptance and recognized revenue are different milestones.
The August 2026 update makes the tension plain. Unaudited first-half revenue reached $78.4 million, while the IFRS net loss widened to $46.9 million. Adjusted EBITDA was also negative. Management raised its full-year revenue expectation to at least $200 million, which remains guidance. Growing sales and spending ahead of infrastructure deployment can occur together.
“Infrastructure cannot be switched on like a tap.”
Jay Chandan / Chairman & CEO / August 2026
That sentence is a better guide to this phase of Gorilla than a futuristic rendering. Hardware has to arrive, work together and serve customer workloads. In its July update, the company targeted September and December 2026 for initial Indonesian deployment tranches. Those targets describe a programme under execution, not proof that every scheduled installation has been accepted.
Buy the workflow, then count the results
For a potential customer, Gorilla belongs on the list when the problem crosses systems: video that needs searching, alerts that need context, or infrastructure that needs integrating. It makes less sense to begin with a sprawling smart-city purchase when the actual requirement is one narrow function. The Port of Tyne example is useful because its starting point can be described without using the word transformation.
The practical buying approach follows from that example: define the inspection, agree what constitutes a useful detection, connect it to the operator’s response and measure the work completed. A credible pilot should test representative footage and difficult conditions, including the alerts staff decide to dismiss. These are editorial conclusions from the workflow, rather than published guarantees of Gorilla’s accuracy or return on investment.
The same discipline applies to infrastructure customers. Capacity must be available where workloads need it, with workable power, networking, financing and delivery arrangements. If those conditions are absent, demand for AI alone will not make a project function. A contract can establish intent and obligations; commissioning establishes whether customers can actually use the service.
Gorilla’s engineering history gives that effort some continuity. Technology executive Alan Luo joined as a senior software engineer in 2003. Its careers materials emphasize open communication and learning, while its current business requires coordination across lenders, hardware providers and local partners. The next test is satisfyingly concrete: installed machines, useful output and customers able to put both to work.