Brendon Ray Hedrick, a former EVGA employee, has published a post looking back on the inside of the company's Nvidia GPU graphics card business, which he experienced from 2016 to 2019. In the post, "Why I Left a Future at EVGA", which we covered on September 28, 2026, he says EVGA felt it had to keep budget models that lost money in order to maintain its GPU supply allocation.
However, Hedrick worked at EVGA from February 2016 to December 2019, so he was not in a position to know firsthand about the management decisions behind the company's 2022 exit from the graphics card business. Read alongside official pricing and demand trends from the period, the picture that emerges is one of competition from Nvidia itself, the GPU supplier, and of growing difficulty in reflecting the cost of custom boards and cooling designs in product prices.
How money-losing budget models made profitability harder
Hedrick moved from technical support to supervisor at EVGA, and later worked in video production, product marketing, and sales. His post describes the changes in EVGA's business from the standpoint of someone who sold products and supported customers.
According to his account, EVGA understood that it needed to offer at least one model matching the lowest price Nvidia announced publicly. The company believed that if it did not, Nvidia might reduce the volume of GPUs allocated to it.
As a result, EVGA kept graphics cards that made no profit as budget models, and tried to secure profits with higher-priced models featuring custom boards and cooling designs.
Hedrick himself writes, though, that even if GPU supply had actually been reduced, it would have been difficult to determine whether EVGA's pricing was the reason. No contract terms with Nvidia have been presented, nor any records showing that supply was actually cut because of pricing.
It has therefore not been confirmed that Nvidia actually reduced supply to manufacturers that did not hold to its prices. Nor is there any material that would let a third party verify the cost or profit margin of individual models sold at a loss.
Even so, the profitability problem Hedrick describes is easy to follow.
The cheapest model becomes the price benchmark for buyers, while the products on which EVGA could earn a profit cost far more. According to Hedrick, budget models priced close to Nvidia's minimum were almost always sold out.
Buyers then saw a product advertised as "starting at $599" that they could not get, while the EVGA products they could actually buy looked considerably more expensive.
Graphics card makers do not merely buy GPUs. They design their own boards, power circuits, and cooling systems, and combine them into finished products for sale.
But because their products are compared against rival cards using the same GPU, they must show buyers value that justifies the price gap if they want to pass on that additional cost. Hedrick's recollections suggest this became steadily harder.
Competition with the Founders Edition, seen through official prices
In 2016, Nvidia announced that the GeForce GTX 1080 Founders Edition would go on sale for $699. Custom models sold by partners, it indicated, would start at $599.
The Founders Edition was to be sold not only by Nvidia itself but also by partners including EVGA.
Here, it is necessary to separate the fact that Nvidia itself sold finished graphics cards from the question of whether Nvidia's products were cheaper than partner products.
Comparing the launch prices of the 80-class cards from the May 6, 2016 GTX 1080 announcement and the August 20, 2018 RTX 20 series announcement gives the following.
| Product / year | Lowest partner price | Founders Edition | Difference |
|---|---|---|---|
| GeForce GTX 1080 (2016) | From $599 | $699 | $100 |
| GeForce RTX 2080 (2018) | From $699 | $799 | $100 |
From the GTX 1080 to the RTX 2080, both the lowest partner price and the Founders Edition price rose by $100. The gap between the two stayed at $100.
The prices are US dollar figures at the time of announcement, and the $599 for the GTX 1080 is the expected minimum price Nvidia indicated for partner products. They do not show actual retail prices in Japan, EVGA's purchase costs, or its profit margins. Performance also differs between generations, so the table does not compare the cost of obtaining the same performance.
At least at this point, the Founders Edition was priced $100 above the expected minimum partner price.
The problem Hedrick felt in 2016 was therefore not that "Nvidia was selling more cheaply than its partners."
It was that Nvidia, which also supplied the GPUs, was selling finished graphics cards itself and competing with EVGA for the same buyers.
Nvidia selling its own-brand graphics cards did not begin with the Pascal generation, a point Hedrick also acknowledges in his post.
Hedrick recalls that Nvidia's own improvements to board design and GPU control also affected EVGA's ability to justify its price premium.
EVGA spent money on custom boards and cooling designs. But according to Hedrick, as generations advanced, such custom designs made less of a difference to everyday gaming performance than before.
Mechanisms such as GPU Boost, in which the GPU automatically adjusts its clock speed when power and thermal headroom allow, had also become widespread.
This is Hedrick's own assessment, however, and not a performance comparison of multiple products measured under the same conditions.
That does not mean quietness and temperatures from custom cooling, board quality, overclocking capability, warranty, and after-sales support lost all value. The question was how much extra buyers would pay for differences that are hard to see in frame rates alone.
Falling demand and inventory problems during the shift to RTX
The GeForce RTX 2080 was announced with a release date of September 20, 2018.
However, real-time ray tracing, the centerpiece of the RTX series, only became usable in Battlefield V with an update on November 14 of that year.
There was a time lag before buyers of the new graphics cards could try ray tracing, their signature feature, in an actual game.
Hedrick is not the only one to recall such sales difficulties.
In its revision of financial guidance on January 28, 2019, Nvidia also said that sales of some high-end Turing products had come in below expectations.
Nvidia cited as reasons the possibility that some customers were waiting for prices to fall, and that some were waiting for further demonstrations of RTX technology in actual games.
The same announcement also mentioned channel inventory of mid-range products left over after the end of the cryptocurrency boom, and weaker demand caused by economic deterioration, particularly in China.
This describes market conditions for Nvidia as a whole and is not a figure showing EVGA's own losses. Still, it helps in understanding a period when companies had to sell a new generation of graphics cards while also clearing older inventory.
A concrete example of an inventory problem cited by Hedrick is the 3GB version of the GeForce GTX 1060.
After demand from cryptocurrency mining fell, he says, the company struggled to sell off stock as the generational changeover approached.
Buyers who find a product unappealing can wait until the next generation. Manufacturers holding inventory cannot. They must sell products and turn inventory into cash before their value falls further.
Hedrick is also critical of EVGA's own management decisions.
He believes the company had not prepared sufficiently for a drop in demand, and recalls that after large-scale layoffs, the remaining employees took on even more duties than before.
He himself came to handle sales and could no longer devote time to the video production work he had built up with colleagues. Deteriorating market conditions affected the very nature of the work he was fond of.
Keeping this separate from the 2022 exit from graphics cards
It became clear that EVGA would end its relationship with Nvidia about three years after Hedrick left.
On September 16, 2022, Jon Peddie Research reported what it had heard directly from EVGA: that EVGA would not sell graphics cards with next-generation Nvidia GPUs, while continuing to sell and support existing products.
EVGA said it would continue selling other products such as power supplies, so its exit from the graphics card business did not mean the company itself was shutting down.
At the time, Nvidia gave a different account.
According to a September 22, 2022 PC Gamer article in which the outlet put questions directly to Jensen Huang, Huang said EVGA's Andrew Han had wanted to scale back the business for several years, and described EVGA as a great partner over many years.
This was Huang's view in 2022 and is not a response to Hedrick's current post.
What Hedrick's recollections convey directly are the problems EVGA's front lines faced from 2016 to 2019.
He describes concerns about GPU supply allocation, competition with Nvidia's own Founders Edition, a combination of money-losing budget models and expensive custom models, falling demand and old-product inventory, and a situation in which it had become hard to explain performance differences even after spending on custom designs.
However, it cannot be concluded that these were directly the reasons for the 2022 exit. Hedrick had already left EVGA by then, and he states explicitly that he does not know whether Nvidia intended to push out partner companies.
What this post shows is not the direct reason EVGA decided to withdraw in 2022, but that keeping the graphics card business profitable had been growing harder for several years before that.
Budget models matched to Nvidia's minimum price, the cost of custom boards and cooling, competition with the Founders Edition, and falling demand and inventory clearance at generational transitions. With these factors overlapping, it appears to have become harder than before for a specialist maker like EVGA to secure profits while showing value that justified its price premium.