Why No GeForce RTX 4070? It’s Coming, Nvidia Says
The GeForce RTX 4070 was conspicuously absent when Nvidia announced the GPU yesterday. However, the company plans to release one as well as a lower-end RTX 4000 model in the near future after it starts ramping up production.
“We’re not ready for a one-off launch,” Nvidia CEO Jen-Hsun Huang said in a question-and-answer with reporters on Wednesday. “What we have ready is the 4090 and 4080. But over time, we will bring other low-end products to the market.”
The announcement also suggests that the RTX 4000 GPU will eventually reach a more consumer-friendly price point. Right now, the most affordable entry in the lineup is the 12GB GeForce RTX 4080, available in November, starting at $899.
The other two, the 16GB RTX 4080 and RTX 4090, which start at $1,199 and $1,599 respectively, are out of reach for consumers looking for a mid-range PC graphics card.
Why Nvidia is ignoring the low-to-mid-end segment for now is “simple” and “not that complicated,” Jensen said. “We usually start at the high end because that’s where the enthusiasts want to update first. We found the 4080, 4090 to be a good place to start. We’ll move the stack down as quickly as possible,” he said.
Still, Nvidia has made the eye-popping decision to sell two very different RTX 4080 models, despite their identical names. That’s because the 12GB model not only has less memory, but also contains only 7,680 CUDA cores. At the same time, the CUDA core count for the 16GB model increased significantly to 9,728. The two models also use different GPU chips, powered by the company’s Ada Lovelace architecture.
As a result, some critic(opens in new window) The purported 12GB RTX 4080 is actually a 4070 model in disguise, but with a $400 price increase over the original two-year-old RTX 3070, which starts at $499.
However, Nvidia sees things differently. In a separate Q&A with reporters, company executives said the RTX 4080 12GB model is a “high-performance GPU” that can outperform the older RTX 3080 12GB card by three times.
Our Editors’ Picks
“We really think it’s worthy of an 80-level product. That’s the real reason it’s named that way,” said one executive. In other words, Nvidia reserved the RTX 4070 for low-power GPUs.
Nvidia’s older RTX 3000 series is also facing an oversupply situation. GPU demand has dropped to the point where retailers are slashing prices to help sell existing inventory. So Nvidia may be waiting for its existing supply to clear before releasing the RTX 4070 and other low-end models. Otherwise any remaining stock of RTX 3000 cards may not yield.
Nvidia’s CEO added: “I hope that by the time frame of Q4, sometime in Q4, the (sales) channel will normalize, which will make room for a great launch of Ada.”
Get our best stories!
register What’s new now Get our top stories delivered to your inbox every morning.