The Complete Guide to NVIDIA Graphics Cards: H100 vs. RTX vs. GTX Differences, Essential Knowledge for Survival in the AI Era
Hello. This is OEING.
Recently, there has been a lot of buzz about NVIDIA CEO Jensen Huang’s visit to Korea. We are in an era where the very fact that he met with domestic companies and the president to discuss “supplying 260,000 GPUs” becomes massive news.
Why is that?
It’s because the entire world is talking about AI. Not just large corporations, but even individuals interested in side hustles or self-development feel a pervasive anxiety that they will fall behind if they don’t understand AI. The ‘heart’ that powers this massive AI wave is NVIDIA’s GPU.
In particular, data center GPUs like the H100 have become the ‘modern oil’—you can’t get one even if you have tens of millions of won. A single word from Jensen Huang can adjust the speed of the global AI industry, a scene that starkly reveals the reality of capitalism.
When faced with such massive issues, even if you’re not planning to buy an H100 right away, you might start to wonder: “So, what exactly is the graphics card in my computer?” or “What do I need to know to even start studying AI?”
If you’re building a new PC or planning to do AI, video editing, or gaming, you’ve probably worried, “What on earth is the difference between RTX, GTX, and H100?”
Therefore, in this post, we will systematically organize the entire lineage and roles of NVIDIA’s graphics cards.
1. What kind of company is NVIDIA?
NVIDIA is a company specializing in GPUs (Graphics Processing Units) founded in 1993. It boasts such overwhelming market share that the formula “graphics card = NVIDIA” has practically become a given.
Currently, NVIDIA’s graphics cards have moved beyond simple gaming use and have established themselves as core components in almost all advanced industries, including AI, robotics, autonomous driving, and supercomputing.
In short, NVIDIA is the company that has transformed from ‘a gamer’s equipment’ to ‘the engine of the AI world.’
2. NVIDIA Graphics Card Lineup at a Glance
| Division | Representative Series | Use | Features |
| GT Series | GT 1030, etc. | Office use, video viewing | Low-cost, low graphics performance |
| GTX Series | GTX 1060, 1070, 1080, etc. | Gaming (old generation) | The former mainstream GPU centered on gamers |
| RTX Series | RTX 20/30/40 Series | Gaming, video, 3D, AI computing | Real-time ray tracing, DLSS support |
| Quadro Series (Now RTX A Series) | RTX A5000, A6000, etc. | Professional workstation | Specialized for CAD, 3D modeling, video editing |
| Tesla / A100 / H100 Series | A100, H100, GH200, etc. | Data center, AI training | Ultra-fast computing, HBM memory, no display output |
3. GTX vs. RTX — The Most Confusing Pair
The GTX series was once the standard for gaming graphics cards. However, with the appearance of the RTX series, a complete generational shift occurred.
| Division | GTX | RTX |
| Release Period | Early to mid-2010s | 2018–Present |
| Core Technology | Pure graphics rendering | Ray Tracing + AI Deep Learning |
| DLSS (Upscaling) | Not supported | Supported (Upscales resolution using AI) |
| RT Cores / Tensor Cores | None | Yes (For AI & light processing) |
| Representative Models | GTX 1060, 1070, 1080 | RTX 3060, 4070, 4090 |
In short, you can understand it as GTX is the old-generation graphics card, and RTX is the new AI-based generation graphics card.
4. The Meaning of RTX Numbers
The number after RTX signifies “Generation + Class.”
For example, in the case of the RTX 4090:
- 40 → 4th Generation (2022–Present, Ada Lovelace architecture)
- 90 → Top-tier Class (Flagship model)
The meaning of the classes is as follows:
- 4090 / 4080: High-end (Top performance for gaming, 3D, AI)
- 4070 / 4060: Mid-to-high range (Value-focused)
- 4050 / 3050: Entry-level (Office use, light gaming)
Additional notations:
- Ti (Titanium) → A higher-performance model within the same class (e.g., 4070 Ti)
- /SUPER → An improved or mid-cycle refresh version
5. Quadro (Now RTX A Series) — Professional GPUs
The Quadro lineup is for people who need precise colors, high stability, and long run times, such as video editors, architects, designers, and 3D artists.
The name has now changed to the “NVIDIA RTX A Series,” with representative models like the RTX A4000, A5000, and A6000.
In short, it’s a GPU optimized for “money-making work” rather than gaming.
6. H Series — The Engine of the AI World
The recently sensational H100 is not a regular graphics card. It is a data center GPU for training AI models and is NVIDIA’s real core source of profit.
| Division | RTX 4090 | H100 |
| Use | Personal graphics | Data center, AI training |
| Memory | GDDR6X | HBM3 (High-Bandwidth Memory) |
| Output | Monitor connection (Yes) | None |
| Price | Approx. 3-4 million KRW | 40-100 million KRW |
| Environment | PC, Workstation | Server, Supercomputer |
These H-series GPUs are the core components of the dedicated AI training hardware used by companies like OpenAI, Google, and Meta, with thousands connected to form massive clusters.
7. A Simple Way to Categorize NVIDIA Graphics Cards
| Purpose | Recommended Series | Example |
| Office/Web Surfing | GT Series | GT 1030 |
| Gaming | RTX 40 Series | RTX 4070, 4080 |
| Video Editing / Design | RTX A Series | RTX A5000 |
| AI Learning / Research | H Series | H100, H200 |
| Industrial Robots / Servers | Jetson / Tesla | Jetson AGX, A100 |
In Summary
RTX is a gamer’s equipment, Quadro (RTX A) is a professional’s tool, and H100 is the engine of artificial intelligence.
We now live in an era where the “game” and the “AI industry” are distinguished by a single graphics card.
Going forward, NVIDIA is completely transforming from a simple graphics card company into an “AI infrastructure company.”
If you are planning to buy a new graphics card right now, don’t just compare performance numbers; choose the GPU that “fits your purpose.”
A gaming RTX is inefficient for AI training, and an H100 is completely unusable in a regular PC.
By understanding NVIDIA’s product line, you can choose the perfect GPU for your needs without unnecessary spending.
If this post was helpful, please leave a “like” and follow us.
OEING reflects on realistic side hustles and survival strategies in capitalism.
#NVIDIAGraphicsCard #RTX4090 #H100 #AIGPU #GPUComparison #OEING #ComputerParts #GamingPC #DeepLearning #JensenHuang