xis.ai Joins the NVIDIA Inception Program for AI Visual Inspection

xis.ai, an AI visual inspection company, has joined the NVIDIA Inception Program. Every inspection system we deploy already runs on NVIDIA hardware, so working more closely with NVIDIA has a direct effect on how our models perform on our customers' production lines.

For manufacturers, that matters because manual inspection can't hold attention at line speed, and whatever it misses reaches the customer.

This article explains what the NVIDIA Inception Program is, how xis.ai inspects products on the line, why NVIDIA hardware is central to that, and what it means for your production.

What is the NVIDIA Inception Program?

The NVIDIA Inception Program is a free program that supports startups building products with AI and other advanced technologies. Members can access resources across NVIDIA’s technology ecosystem, including:

• Developer tools: NVIDIA SDKs, model libraries and developer platforms.

• Training: Self-paced technical courses and discounted expert-led workshops.

• Hardware and software offers: Preferred pricing on selected NVIDIA products and discounts from NVIDIA’s partners.

• Development and growth support: Cloud credits, investor connections, networking opportunities and go-to-market resources.

For xis.ai, joining the program means closer access to the tools and training behind the hardware our inspection systems already use.

What Does xis.ai Do? AI Visual Quality Inspection Ecosystem for Manufacturing

xis.ai gives manufacturers a complete, end-to-end AI visual quality ecosystem, from building and training inspection models to running them on the line and turning the results into production insights. The software runs on an industrial PC at the production line, checks every unit in real time and records a result for each one.

It does more than find defects. Each inspection is set up around your own quality parameters, such as color, dimensions, shape, position and surface finish, so every unit is checked against the specifications that matter for your product.

It covers a wide range of inspection tasks:

• Defect and anomaly detection

• Color, dimension and measurement checks against your specifications

• Label and code verification

• Seal and packaging checks

• Object detection and classification

This is where xis.ai differs from traditional rule-based machine vision inspection. Fixed rules and tolerances can reject perfectly good products when no two units look exactly alike. xis.ai’s models learn from images of your own product, including its natural batch-to-batch variation, and judge each unit against the parameters you define.

The platform serves a wide range of manufacturing industries, including:

• Consumer goods

• Food and bakery

• Pharmaceutical

• Automotive

• Electronics

• Metal and machining

• Logistics and packaging

• General manufacturing

How the xis.ai Platform Works

The platform has three parts that cover the full quality process, end to end:

• xis.dev: Builds, trains and retrains AI models. It is a no-code environment, so your quality and process engineers can annotate images and train models without a data science background, and retrain them when the product changes.

• xis.rt: Runs the AI models at each inspection point on the production line. It runs state-of-the-art models at 150 frames per second and signals reject, alarm or line halt to your PLC and MES over OPC-UA and MQTT.

• xis.dashboard: Brings inspection results, images and data from every line, plant and country into one view, hosted on your own servers. It turns that data into actionable insights, such as defect trends, recurring failure patterns and line-by-line performance, so your teams can act on problems earlier, reduce scrap and optimize production.

How xis.ai Uses NVIDIA Technology for AI Visual Inspection

In computer vision, a model is only as fast and as accurate as the hardware it runs on. That is why NVIDIA technology is at the core of every xis.ai deployment.

NVIDIA's Role in AI and Computer Vision

NVIDIA GPUs are widely used to run modern AI models. NVIDIA also builds the software that helps those models run efficiently. Its CUDA-X libraries give developers ready-made, GPU-accelerated tools, including libraries that speed up the loading, pre-processing and post-processing of images for vision AI.

NVIDIA TensorRT then optimizes trained models for deployment on NVIDIA GPUs, for low-latency, high-throughput inference.

On a high-speed production line, processing speed decides whether every unit can be inspected. The more frames per second a model can process, the more units it can check without slowing the line down.

The Hardware Behind Each Inspection Point

Each xis.ai inspection point has three components:

• Industrial Camera: Standard industrial cameras and interfaces, with nothing proprietary.

• Controlled Lighting: Specified as part of the design, because lighting is where most vision systems fail on a live line.

• Industrial PC (IPC) with an NVIDIA GPU: A compact computer built for the factory floor and placed at the line.

Using an industrial PC with GPU acceleration lets xis.rt run AI models in real time, right at the line. Running on NVIDIA GPUs with NVIDIA's accelerated libraries, our models deliver higher frame rates and higher accuracy than they do on other devices and GPUs.

Why NVIDIA is the Right Platform for xis.ai

• End-to-end Integration: NVIDIA develops both the GPU and the software stack that runs AI on it, so models and hardware are designed to work together.

• Readily Available Hardware: Industrial PCs with NVIDIA GPUs are widely available, so customers can source them easily. xis.ai helps select and specify the right one, and offers a partner program for industrial PC and edge hardware manufacturers to validate their configurations.

• Works with Existing Equipment: In most cases, xis.ai can be retrofitted onto the cameras, lighting and industrial PCs already on your line.

• Ongoing Engineering Through Inception: Access to NVIDIA's latest SDKs, training and developer tools helps our engineers keep xis.ai's models fast and accurate on the hardware our customers use.

What Does the xis.ai and NVIDIA Partnership Mean for Manufacturers?

For quality and operations managers, the partnership shows up in results on the line:

• 100% inspection, not sampling: xis.rt brings automated quality inspection to the line at 150 frames per second, so every unit is inspected at line speed, 10x faster than manual inspection.

• High detection accuracy: xis.ai's AI defect detection delivers 99.9% accuracy, and every decision is saved with its image and measured value for traceability.

• Standard hardware: Each inspection point uses a standard industrial camera, controlled lighting, and an industrial PC with an NVIDIA GPU, with no proprietary hardware.

• Deployment without downtime: Installation requires zero line stoppage, and each inspection point can be added independently.

• On-site data processing: Inspection runs fully offline at the line. There is no cloud service, and no images leave your network.

• More control for your engineering team: When a product or recipe changes, your engineers retrain the model in xis.dev without waiting for a service visit.

See How xis.ai Performs on Your Product

xis.ai's AI visual inspection checks every unit at line speed, so defects are stopped at the line instead of being found by your customers. The fastest way to see what it can do on your production line is a free feasibility check on your own product.

Share 20 images of the defect you need to detect. Our engineers will train a model on your product and send you a clear assessment of whether it meets your requirements, at no cost and with no commitment. If it does, the next step is a pilot on one production line, running alongside your current process.

Start a feasibility check

Frequently Asked Questions

What is the NVIDIA Inception Program?

The NVIDIA Inception Program is a free startup program that supports companies developing AI and emerging technologies through access to NVIDIA’s developer tools, technical training, hardware offers, and business resources.

What is an industrial PC (IPC)?

It is a compact computer built for the factory floor. At each xis.ai inspection point, it holds the NVIDIA GPU that runs the inspection.

Can xis.ai use the cameras and industrial PCs we already have?

Usually, yes. xis.ai works with standard industrial cameras, lighting, and industrial PCs, and we check compatibility before you buy anything.

Can one xis.ai system check several products and defects at once?

Yes. One xis.rt instance can check several product variants and defect types at the same time, and each check gives its own result for every unit.

What happens if xis.ai makes a wrong call?

Every decision is saved with its image and measured value, so your team can see exactly what happened. They can then adjust the settings in xis.dev based on those records.