Artificial intelligence did not suddenly appear with ChatGPT. It evolved over four decades through advances in computer graphics, game development, machine vision, and parallel computing. Graphics processors, originally built to draw pixels, gradually became programmable computing engines capable of training neural networks and running AI models. JPR followed that evolution from the beginning, covering graphics hardware, GPU computing, machine learning, and AI processors as each technology emerged. Looking back, the progression seems remarkably logical, even if it rarely felt that way at the time.
I might not exist but for you.
People often ask when JPR began covering artificial intelligence. The answer depends on what they mean by AI.
Today’s AI refers to large language models, generative AI, autonomous agents, and foundation models. Forty years ago, AI meant something entirely different. Researchers talked about expert systems, neural networks, fuzzy logic, machine vision, and pattern recognition. Game developers used the term to describe scripted routines that controlled non-player characters, while engineers applied AI techniques to CAD, robotics, and image analysis. JPR and its predecessor, JPA, covered many of those technologies years before anyone imagined conversational AI.
When Jon Peddie Associates opened its doors in 1985, the graphics industry stood at the beginning of its transition from fixed-function hardware to programmable computing. Graphics workstations powered CAD, scientific visualization, and digital content creation. Researchers already experimented with neural networks and knowledge-based systems, although computing power limited what they could accomplish. AI remained an enabling technology rather than a market of its own.
The first AI many consumers encountered appeared in games. During the mid-1990s, developers programmed enemy behavior with lookup tables, decision trees, and finite-state machines. Characters reacted to player actions, navigated environments, and coordinated attacks through carefully designed logic rather than learning. In 1994, Matrox demonstrated Sentõ, a 3D game that showcased intelligent camera control and game behavior on its Impression Plus graphics board. Those techniques now seem simple, yet they established the principle that software could simulate intelligent behavior.
Figure 1. NPCs got smarter, and deadlier. (Source: JPR)
The next milestone arrived before GPUs entered the picture. In the late 1990s, AnimaTek introduced Jennifer, an interactive three-dimensional virtual spokesperson developed for e-commerce applications. Jennifer greeted visitors, answered questions, and guided customers through virtual auto shows. The project drew on Barbara Hayes-Roth’s pioneering work at Stanford on intelligent software agents, demonstrating that believable digital personalities could support commercial applications years before today’s AI assistants.
A few years later, Ananova carried that vision even further. Introduced in 2000 and widely recognized in 2001, the animated news presenter read stories on demand through speech synthesis and a carefully designed digital personality. Her creators trained the visual appearance by studying thousands of human faces to produce a friendly, approachable virtual presenter. Looking back from 2026, Ananova anticipated many characteristics of today’s AI presenters, even though she lacked large language models and modern speech generation. She represented an important step toward AI-driven digital humans.
The real turning point for AI came from an unlikely source: cats.
In 2007, Fei-Fei Li launched ImageNet at Princeton University, with the conviction that computer vision needed dramatically larger datasets rather than incremental improvements to algorithms. Working with Christiane Fellbaum, co-creator of WordNet, the team organized millions of images into a structured hierarchy. Thousands of cat photographs across numerous breeds became part of the training data that taught computers to recognize visual concepts. By 2009, ImageNet had become the benchmark that reshaped computer vision research.
Figure 2. Trying to find cats. (Source: JPR)
Google Brain expanded that idea in 2012. Andrew Ng, Jeff Dean, and their colleagues trained a deep neural network on 10 million randomly selected YouTube images. No one instructed the network to find cats. It learned that concept by itself because cat faces appeared frequently throughout the training data. The experiment demonstrated that sufficiently large neural networks could discover meaningful visual features without explicit programming. Those famous cats captured the imagination of researchers and the public alike, symbolizing a new era of data-driven learning.
Those breakthroughs depended on another technology that originated in graphics.
Graphics processing units never started life as AI processors. Their journey began with programmable shaders around 2001, which allowed developers to use graphics hardware for scientific and engineering workloads. Researchers soon explored general-purpose GPU computing, or GPGPU, for image processing, simulations, and numerical analysis.
Everything changed in 2006 when Nvidia introduced CUDA. For the first time, developers could program GPUs directly in C without disguising computations as graphics operations. CUDA lowered the barrier to parallel computing and attracted researchers working on machine learning, neural networks, and scientific computing. The GPU had become a programmable parallel processor.
Figure 3. Nvidia GTX 580 started as a game board and became an AI processing pioneer. (Source: EVGA)
Another milestone followed in 2012. Alex Krizhevsky, Ilya Sutskever, and Geoffrey Hinton trained AlexNet on two Nvidia GTX 580 GPUs. Their convolutional neural network dramatically improved ImageNet classification accuracy and convinced the research community that GPU-accelerated deep learning represented the future of AI. From that point forward, GPUs became the preferred platform for training neural networks.
JPR followed each stage of that evolution. During the early years, our coverage focused on graphics processors, workstations, visualization, and CAD. As GPUs became programmable, we expanded into GPU computing, CUDA, and heterogeneous processing. Machine learning, computer vision, autonomous vehicles, and edge inference naturally followed. In 2014, we published our first article devoted specifically to AI, asking whether machine learning would become a job creator, a job killer, or humanity’s next indispensable tool. That same year, IBM introduced TrueNorth, its neuromorphic processor inspired by biological neural networks.
Our first dedicated AI market research arrived in 2017 with the Video Processor Unit Quarterly report. VPUs accelerated computer vision, computational photography, and real-time inference before those capabilities migrated into the image signal processors now found in smartphones and autonomous vehicles. In many respects, VPUs represented the first dedicated AI processors.
Today JPR’s AI research spans AI processors, AI PCs, NPUs, physical AI, photonic processors, and quarterly market tracking covering more than 150 companies and hundreds of products. The subject matter has changed, yet the underlying story remains remarkably consistent. Graphics processors evolved into programmable processors. Programmable processors evolved into AI accelerators. AI accelerators now power the infrastructure behind generative AI, robotics, autonomous systems, and scientific discovery.
Graphics did not disappear. It became the computational foundation of artificial intelligence.
Table 1. Timeline of the evolution.
Looking back over four decades, AI did not replace graphics; it grew from graphics. Every stage of the journey built on advances in programmable hardware, software, algorithms, and data. The same GPUs that once rendered polygons now train foundation models and power large language models. JPR’s research followed that progression because accelerated computing remained the common thread connecting graphics, high-performance computing, machine learning, and artificial intelligence.
What do we think?
JPR’s AI coverage reflects continuity rather than reinvention. Graphics, GPU computing, machine learning, and AI represent successive stages of the same technological evolution. That perspective helps explain why graphics companies now lead AI infrastructure and why many of today’s AI breakthroughs originated in technologies first developed for visualization, simulation, and interactive computing.
Inflection signal
The evolution from graphics to AI marks more than a technology transition; it represents an inflection point in computing. Programmable graphics processors created the hardware foundation for deep learning, while large datasets and neural networks unlocked new applications. Today’s AI infrastructure extends that trajectory into every sector of computing. Understanding this history clarifies why GPUs, NPUs, and specialized AI processors now define the industry’s direction and why future innovations will continue to emerge from advances in accelerated computing.
Take a look at our AI library, where, among things, we keep track of the 151 companies offering 292 AI processors. JPR puts the “I” in AI (and because you will ask—Intelligence, market intelligence).
LIKE IT? THINK YOUR FRIENDS AND ASSOCIATES MIGHT? PLEASE SEND IT TO THEM WITH OUR BEST WISHES.