Skip Navigation Links

Gorilla Technology: Allying Taiwan and Japan industry leaders to expand global edge AI markets

Judy Lin, DIGITIMES, TaipeiThursday 27 January 2022
Gorilla Technology founder and CEO Dr Spincer Koh

Gorilla Technology Group, a Taiwan-based edge AI company, has revealed plans to invite leading technology application brand companies from Taiwan and Japan to form a "Smart City Diverse Industry Alliance" ecosystem, with the goal of jointly expanding the global markets for edge AI analytics solutions and services.

Gorilla Technology Group just announced a SPAC merger in December 2021 and plans to list on the Nasdaq under the symbol "GRRR" by the end of Q1 2022. Spincer Koh, the founder and CEO of the company, stated in an exclusive interview with DIGITIMES Asia that he will identify and invite leading companies in various critical industries as strategic partners. Gorilla will assist them to design and develop new products with embedded Gorilla AI core modules inside that will be offered in global markets with their own brand names.

"We would like to invite industry-leading companies to collaborate on expanding global edge AI markets together, using our AI software to integrate the resources and competitive advantages of Taiwanese and Japanese companies accumulated over the last few decades to achieve an eco-growth model in the Smart City Diverse Industry Alliance," said Koh.

After the IPO, Gorilla will launch a strategic partner collaboration platform. Certified system integrators or hardware partners will embed the Gorilla AI core technology into their own products and turn them into an appliance that can be utilized in a variety of scenarios. Gorilla has developed dozens of use cases in different verticals thus far.

Companies that are leaders in various industrial verticals such as robotics, industrial PCs, semiconductor chip manufacturing, and so on, are the potential partners that Gorilla would like to invite for the Smart City Diverse Industry Alliance, because not only will their products receive publicity from the IPO of Gorilla Technology, but Gorilla's edge AI technologies will also gain recognition by the customers of partner companies. "Our partners have much more prestige in their respective domains than we do. But we are an enabler of AI core technology. When we complement each other to form an edge AI ecosystem and grow, we will be able to build a huge market of Smart City edge AI applications," said Koh. He has set the company's growth goals for the next 5 years, including sales CAGR at 52%, EBITDA at 45%, and gross margins to reach 56% by 2026.

Meeting commercial and security demands with integrated solutions

Gorilla's exclusive and patented edge AI analytic models will be embedded in various equipment and devices which can be used as edge computing applications. Its AI algorithms will help achieve the optimal performance of applications used in various vertical scenarios.

However, this is only one of the three business models of Gorilla Technology. Its second business model involves industrial PC and server brand-name companies to produce several models of AI appliances that are offered with Gorilla's own brand name in the market. The partners collaborating with Gorilla under this business model include Intel, Dell EMC; Japan's Softbank, and its largest electronic distributor Macnica; telcos such as PLDT of the Philippines, Telstra of Australia, AIS of Thailand, and INTI of Indonesia; as well as communications equipment manufacturer Lanner and embedded platform system producer Neousys Technology.

Koh said with a sense of pride that Gorilla has launched 6-7 machines of various types and specifications on the Intel platform — from the low-end Core i3 to the most powerful Xeon CPUs. Customers can simply plug and play the appliance to conduct 3 to 20 channels of high-definition real-time video analytics, according to Koh. Gorilla AI supports machines on Intel, NVIDIA and ARM platforms beginning in 2021, and will soon launch machines that support Qualcomm and Google TPU platforms later in 2022. "We will have formally established Gorilla as a platform-independent edge AI company by then," said Koh.

Koh stressed that the AI appliances were created to show the world that AI technologies can be used simply by "plug and play." "AI is not just about algorithms. Our products are the combination of hardware and software that can optimize edge-AI performance and cybersecurity protection to deliver the best cost-performance efficiency to customers while minimizing data transmission latency." Koh explained, "We solve the high-stakes challenge customers need to react on the spot and that massive data needs to be collected, computed and responded to simultaneously."

The Software as a Service (SaaS) at the back end completed the last element of the golden triangle of Gorilla's business model. In today's edge-computing era, in which massive heterogeneous data may contain sensitive, private, and confidential information that needs to be processed in real-time on-site, many companies have been reluctant to upload sensitive data on public cloud platforms due to security and cost concerns. "Edge AI is not just about resolving the algorithms," said Koh. Gorilla deals with this complicated issue by screening data with privacy concerns on the edge device itself using security enforcement. The data is then encrypted and transmitted to the customers' private cloud platforms, thus resolving any data security issues. For video content that is too big to upload, they process the video into data on the edge side by dividing them into very small data packets and transmitting them to the cloud for further analysis.

Video footage that originally needed to be monitored by 1,000 people, can be processed into data packets under 500k each and combined into one data stream at the backend. Even for objects that are constantly changing in the front-end environment, edge data processing can record the specific people, vehicles, objects, or even behaviors that appear in specific times and locations. Gorilla completely broke down the concept of relational-based databases and replaced it with heterogeneous unstructured databases to record objects, times, and places, which is multi-dimensional.

Koh expanded on the rationale behind the heterogeneous unstructured database. It is impossible to predict when an object or person would appear in an IP camera —but what if a car drives past several cameras? When the video stream from those cameras is analyzed at the edge and the analytics data stream is transmitted to the backend for integrated data analysis, the path of the car during that period of time would be more evident and likely predictable. "Cross-edge computing" is the process of combining data from different edge devices for backend integration and data analysis within seconds.

How much can Gorilla AI improve operational performance? In an online meeting with institutional investors, Koh offered the case study of the Kaohsiung Transportation Bureau's Gorilla AI system for monitoring illegal left turns at intersections as an example. Where there used to be an average of 3 traffic violations incidents caught by the police per day, the Gorilla AI system captured an average of 305 events daily, thus increasing monitoring efficiency 100-fold.

Solving 4 critical problems of edge AI

Gorilla optimizes the data processing performance on local sites with their algorithms running on the smallest CPUs. As SaaS is based on a concept of resource sharing, solving issues at the edge side would save customers lots of money by allowing them to save the data that are necessary and important.

"We believe the hybrid cloud is the way of the future, so we proposed the idea of a distributed data center or an edge-based data center to place the backend servers as close as possible to where the customers' data are located, in order to reduce transmission latency," Koh said depending on the location of customer's cloud platform, it normally takes from 1 to 10 seconds for the data to be transmitted from edge devices to the backend.

Gorilla will also recommend suitable AI appliances to customers for deployment on different use cases. There are 6-7 different types of edge devices for example that can be used in a train station scenario, optimizing the performance and meeting the customer's demands.

"True edge AI involves the edge computing box, the backend server, cross-edge computing, and the communications between machines. The sum of all these is edge AI by my definition," said Koh, who earned his Ph.D. degree from the University of Washington in the 1980s.

Gorilla uses Cisco's methodology to turn AI into a hardware solution, enabling its customers to plug and play easily, and to solve the four major problems of edge AI — cost, latency, security, and reliability — with a distributed server network that provides Software as a Service closest to the customer's sites.

"AI is not a fairy tale. It has to be down-to-earth and easy to use in our daily lives." Koh said the end-to-end solution would provide the convenience of AI for customers by simply using plug and play, so that "they don't need to spend 20 years learning AI on their own."

Factbox of Gorilla Technology Group Inc.
Company Info details
Established in 2000
Core management Founder and CEO Dr. Spincer Koh
CTO Alan Luo
CSO Felix Song
Investors SBI Investment, Telstra Ventures, Acer, Dell EMC
Award recognition Best edge computing product, Interop Tokyo Japan 2019
2020 Intel Partner of the Year – Best Innovative Design
Post-IPO growth objectives Sales CAGR 52%, EBITDA CAGR 45%, and gross margin to reach 56% by 2026

Source: Gorilla Technology, compiled by DIGITIMES, January 2022


IT + CEServer, IPC, cloud computing, IoTSoftware, big data


AICEOCloudDeepTechEDGEedge AIGorilla TechnologyInterviewJapanTaiwan

Edge AI Video Analytics IVAR® Smart Attendance EVMS Cybersecurity Analytics NetProbe NetTrap HMD FR-MOTP SCP Whitepapers
AI Toolkit AI Toolkits by Vertical AI-Appliances Video Analytics Vision AI Mobile Vision AI Edge Vision AI AI Screen Station Cybersecurity NetTrap Server NetProbe Server EdgeGuard FR-MOTP Server AI SaaS Smart Retail Smart City & Transportation Endpoint Security
Solutions Government & Public Services Smart City & Transportation Industrial Commercial
About Us
The 800 pound Gorilla Press Room Events & Trade Shows Videos Partner Program Contact Us Blog
Service Portal Partner Program