For three decades, NVIDIA remained largely unknown to the general public unless you were a PC gamer. Today, the artificial intelligence boom has propelled the chipmaker to become the world's most valuable company, yet its original fan base—gamers—feels increasingly sidelined.
"Gaming is no longer the core growth engine for the company, though it unquestionably once was," stated Stacy Rasgon, an analyst at Bernstein Research.
NVIDIA popularized the graphics processing unit (GPU), the type of chip responsible for high frame rates and efficient rendering that enables premium gaming experiences. When NVIDIA launched its first GPU, the GeForce 256, in 1999, the company had undergone massive layoffs and was near bankruptcy due to the product's development costs. It was gamers worldwide who eagerly purchased the new processor, rescuing NVIDIA from the brink.
Now, with soaring demand for AI, nearly all of NVIDIA's revenue comes from AI-related products rather than gaming hardware. Simultaneously, the production of AI chips is exacerbating a shortage of memory resources, forcing NVIDIA to make difficult prioritization decisions.
Given the current undersupply of memory, it is inevitable that NVIDIA would prioritize manufacturing higher-margin data center AI chips, such as those in the Hopper and Blackwell architectures. Over the past three years, NVIDIA's compute and networking business segment has achieved an average operating margin of 69%, compared to just 40% for its consumer-facing gaming GPU division.
Greg Miller, co-founder and host of the renowned gaming podcast *Kinda Funny Games Daily*, expressed his disappointment: "I understand chasing higher profits, but it's genuinely heartbreaking. Gamers built NVIDIA; they should remember their roots."
If analyst predictions hold, 2026 could mark the first time in thirty years that NVIDIA does not update its consumer GeForce graphics card lineup. NVIDIA has stated that the gaming community remains vital to the company and that the brand continues to innovate, develop, and release new technologies for gaming. The currently available GeForce RTX 50 series cards were announced at CES in January 2025.
With both CES and the GTC developer conference having concluded this year, many gamers worry there will be no new generation of cards announced in 2024, though NVIDIA has historically sometimes unveiled new hardware in September. Despite the company's strategic pivot, some gamers believe a slower upgrade cycle could benefit consumer budgets.
Tim Gettys, Miller's partner, commented, "The hardware refresh rate is too fast for the average person to keep up with; there's no need to upgrade annually. Slowing the cycle and waiting for a truly worthwhile new generation is actually more consumer-friendly."
NVIDIA's path to AI dominance began two decades ago. In 2006, the company introduced the CUDA software platform, allowing developers to use GPUs for general-purpose computing beyond just graphics. The year 2012 is often seen as the start of the modern AI explosion, when the AlexNet neural network, powered by NVIDIA GPUs and CUDA, decisively won an image recognition competition, showcasing profound deep learning capabilities.
NVIDIA has not completely halted gaming GPU production, but its $7 billion acquisition of high-performance computing chipmaker Mellanox Technologies in 2020 signaled a formal shift in focus toward dedicated AI chips. Since then, NVIDIA has consistently iterated on high-end AI processors and launched complete AI computing systems, including the new Vera Rubin platform.
NVIDIA does not publicly disclose AI chip prices, but analysts indicate a single Blackwell AI chip can cost up to $40,000, while Futurum Group estimates an entire Vera Rubin system could reach $4 million. In contrast, NVIDIA's consumer RTX 50 series gaming cards are priced between $299 and $1,999.
During the cryptocurrency booms of 2018 and 2021, GPUs were essential for mining, with prices on e-commerce platforms soaring to three times their MSRP. After the crypto mining downturn in 2022, prices normalized somewhat, though the flagship RTX 5090 still sells online at nearly double its official price. Strong ongoing demand for previous-generation products also reduces NVIDIA's incentive to release new gaming GPUs.
The core issue behind NVIDIA's deprioritization of gaming is a severe memory supply shortage. Industry sources indicate that due to a critical shortage of general GPU memory, NVIDIA plans to reduce production of its latest gaming graphics cards by up to 40%.
DRAM provides high-speed temporary data caching for GPUs, enabling parallel processing. Personal computers equipped with gaming graphics cards are the segment most severely impacted by this memory shortage. Rising memory costs directly increase GPU production expenses, which are ultimately passed on to consumers. Gartner predicts global PC prices will rise 17% this year, with PC shipments declining 10.4% year-over-year.
"Hardware costs are soaring across the board, with gaming hardware prices showing no signs of decreasing, and NVIDIA is clearly favoring another customer segment. It's frustrating," Gettys said. Gartner anticipates the entry-level home PC market will shrink by 2028, which will correspondingly contract the market for NVIDIA's entry-level gaming GPUs.
NVIDIA will undoubtedly allocate its limited memory inventory to higher-profit, higher-margin AI chips first. "The delay and slowdown in gaming GPU roadmap updates are largely due to memory allocation issues. Almost all available memory resources are prioritized for AI compute operations," Rasgon analyzed.
High-end AI chips like Blackwell and similar architectures heavily utilize High Bandwidth Memory (HBM). Producing 1GB of HBM requires four times the silicon wafer area compared to standard DRAM. This resource倾斜 directly leads to depleted traditional memory supplies for consumer electronics and civilian hardware.
NVIDIA responded that the entire GeForce series remains in normal supply, market demand is stable, and the company is working closely with suppliers to increase memory allocation as much as possible. Gettys remarked bluntly, "AI business profits are triple those of gaming, and shareholder returns are doubling. Even though gaming built NVIDIA, they will gradually abandon the gaming market."
Many in the gaming community perceive this shift as a direct insult. At the GTC keynote in March, CEO Jensen Huang did announce a significant gaming-related update, but it was largely poorly received by gamers. He revealed that a new version of its upscaling technology, DLSS 5, would launch in the fall. While DLSS traditionally uses AI to upscale lower-resolution images, boosting frame rates and allowing games to run smoothly on less powerful hardware, the new version incorporates generative AI.
Gamers expressed concern that the AI might alter the original artistic style of games. Huang demonstrated AI-enhanced, more realistic-looking characters from popular titles like *Resident Evil: Village*, *Starfield*, and *Hogwarts Legacy*. "I play games because they are works of art. I value the developers' original creative intent. The game industry is already suffering from constant layoffs and studio closures; this has created more anxiety," Miller said.
Gettys, who previously praised DLSS for lowering hardware barriers, stated, "The technology itself is impressive, but adding generative AI to alter the visuals feels like an insult to players." He worries this could lead to fully AI-generated games, which he believes is NVIDIA's ultimate goal, noting that Elon Musk's xAI studio plans to release its first AI-generated game by late 2026.
"AI directly modifying developers' original art could eventually replace the developers themselves, leading to studio closures," Gettys added. NVIDIA issued a statement asserting that "games are a creative art form for storytelling and immersive worlds," and that its RTX technologies, including ray tracing, path tracing, DLSS super-resolution, frame generation, and DLSS 5, are "tools to help developers realize their creative vision while balancing performance and visual quality."
At GTC, Huang claimed AI will "completely revolutionize computer graphics." Addressing criticism that DLSS 5 might make all games look similar, he responded firmly in a Q&A session the next day, calling such views "completely incorrect," and emphasizing that developers retain full control and can fine-tune the generative AI to match their artistic style.
Despite the strategic shift, NVIDIA remains the undisputed top choice for many PC gamers. For over a decade, NVIDIA's cloud gaming service, GeForce NOW, has operated with free and paid subscription tiers, allowing users to stream games they own from platforms like Steam, powered by GPUs in NVIDIA's data centers, independent of their local hardware.
"Companies like Sony and Microsoft are exploring cloud gaming, but NVIDIA's GeForce NOW offers a truly mature experience," Miller evaluated. Gettys was more direct: "It surpasses all competitors, enabling players with low-end devices to enjoy high-end gaming experiences. Its technical prowess is irreplaceable."
AMD is NVIDIA's primary competitor in the gaming GPU space with its Radeon lineup, but both companies face the same memory constraints. "If NVIDIA can't get enough memory, AMD can't either. Both brands have loyal followers, but among gamers, the preference is clear," Rasgon noted. Gettys added, "Within the PC gaming community, NVIDIA remains the unquestionable preferred choice."
Comments