China’s GPU Push, Rising Motherboard Costs and the SSD Price Problem

PC hardware is becoming increasingly difficult to predict in 2026.

A new Chinese GPU from Moore Threads is promising a major jump in 3A gaming performance while bringing hardware-accelerated ray tracing to its latest architecture. At the same time, motherboard manufacturers are facing another wave of rising component costs, and the enterprise SSD market is seeing revenues explode as AI data centers consume more storage than ever.

For gamers, these may look like three separate hardware stories. Put together, they show how quickly the economics of building a gaming PC are changing.

Moore Threads’ Lushan GPU Is Bringing Hardware Ray Tracing to Chinese Graphics Cards

Moore Threads is preparing to launch a new graphics processor that could become one of the company's biggest steps toward competing in modern PC gaming.

During its September 3 half-year earnings briefing, Moore Threads confirmed that its Lushan GPU is expected to launch by the end of 2026, although the company has not yet announced a specific retail release date. Lushan is based on the fifth-generation Huagang architecture and is designed primarily for high-performance graphics rendering.

The numbers Moore Threads is giving it are certainly ambitious. Compared with the previous-generation MTT S80, the company claims that Lushan can deliver up to 15 times the 3A game rendering performance, 64 times the AI computing performance and 50 times the ray-tracing performance. These are official company figures rather than independent benchmarks, so the real comparison with current GeForce and Radeon cards will have to wait until the hardware reaches reviewers and players. (ithome.com)

The most interesting part for gamers is not actually the 15x figure. It is the move to hardware-accelerated ray tracing.

Moore Threads says the Huagang architecture combines its first-generation AI Generative Rendering technology with a second-generation hardware ray-tracing acceleration engine. The GPU supports DirectX Raytracing and DirectX 12 Ultimate, which means the company's latest graphics architecture is no longer relying entirely on software approaches for real-time ray tracing. (ithome.com)

That is a meaningful milestone for China's domestic GPU industry. Modern AAA games increasingly use ray tracing, upscaling and other GPU features that are difficult to support well through software alone. If Lushan can provide useful ray-tracing performance in actual games, Chinese GPUs could start being judged on gaming performance rather than simply compatibility.

Moore Threads is also pushing Huagang as a general-purpose architecture rather than a gaming-only solution. The company says the new instruction set raises computing density by 50% and energy efficiency by 10 times, while supporting precision ranging from FP4 to FP64. Its MUSA software stack also supports AI frameworks including vLLM, SGLang and Ollama.

But none of those specifications answer the question gamers will ultimately care about: how well does Lushan actually run modern games?

That remains unknown. We still need independent testing, mature drivers, broader game compatibility and, perhaps most importantly, a competitive price. A claimed 15x improvement over the MTT S80 is impressive, but it does not automatically tell us where Lushan will sit against NVIDIA and AMD's current products.

For now, the hardware ray-tracing support is probably the bigger story. If Moore Threads can turn that feature into playable frame rates in major AAA games, Lushan could mark a noticeable change in how domestic GPUs are viewed by PC gamers.

Motherboard Costs Are Rising, and Waiting May Not Make a New PC Cheaper

While Moore Threads is trying to push gaming hardware forward, the rest of the PC market is dealing with a very different problem: rising production costs.

Supply-chain reports indicate that motherboard manufacturers are under growing pressure as component prices continue to climb. PCB materials are reportedly up by around 15% to 50%, different types of capacitors have risen by roughly 50%, and Wi-Fi modules have increased by around 10% to 20%. Combined, those increases are already pushing motherboard production costs up by more than 10%.

The biggest concern is the PCB itself. Industry data cited by 博板堂 estimates that the price of PCBs used in consumer motherboards could rise by at least 50% in 2026, while copper, capacitors and controller chips are also becoming more expensive.

The interesting part is that manufacturers have not immediately passed the full increase on to buyers.

ASUS reportedly adjusted channel prices in early August, but some models only increased by around 50 yuan. Maxsun's reported increases were even smaller, at roughly 10 to 20 yuan for selected models.

That is understandable. PC builders are already dealing with high memory prices, and the overall DIY market has been under pressure as consumers delay upgrades. A much larger motherboard price increase could make the situation even worse.

But manufacturers can only absorb higher costs for so long.

If PCB and component prices keep rising, more motherboard brands may eventually have to raise retail and channel prices again, particularly on models with thinner margins or more expensive specifications.

For gamers, that changes the usual logic of waiting for a better deal. Hardware prices often fall over time as products become older, but that only works when manufacturing costs are stable or declining. When the supply chain itself becomes more expensive, waiting does not guarantee that the final product will become cheaper.

A motherboard is not normally the component that dominates a gaming PC budget, but it is still one more place where costs can increase. Combine that with expensive memory, and an upgrade that looked affordable a few months ago can quickly become harder to justify.

There is still no indication that every motherboard is about to receive a dramatic price increase. The more likely scenario is a gradual adjustment across selected models as manufacturers reach the point where they can no longer absorb higher costs.

For anyone building a PC later this year, however, the old strategy of simply waiting for everything to become cheaper is starting to look much less reliable.

AI Is Making Enterprise SSDs a Huge Business, and Gamers Are Feeling the Pressure

Storage is facing its own version of the same problem, although the biggest numbers are currently coming from the enterprise market.

According to TrendForce, the world's five largest enterprise SSD suppliers generated a combined $37.59 billion in revenue during the second quarter of 2026, an extraordinary 103.6% increase from the previous quarter. The growth was driven by stronger shipments as well as higher contract prices.

The market is also extremely concentrated. Those five companies controlled 91.9% of the enterprise SSD market during the quarter.

Samsung remained the largest supplier with $14.35 billion in revenue and a 35.1% market share. SK hynix and Solidigm followed with more than $8.63 billion, while Micron reached approximately $6.98 billion and recorded the fastest quarterly growth among the top five, increasing revenue by 126.3%.

Kioxia and SanDisk completed the top five, with both companies benefiting from stronger demand for high-capacity enterprise SSDs.

The reason behind much of this demand is obvious: AI infrastructure needs enormous amounts of storage.

Cloud providers are upgrading data-center infrastructure, moving toward faster PCIe interfaces and deploying increasingly large AI systems. High-capacity QLC SSDs are becoming particularly important, while NAND manufacturers are also focusing more heavily on enterprise products where demand and pricing are stronger.

TrendForce expects enterprise SSD demand to remain high through the third quarter, driven by the continued expansion of generative AI services, cloud data centers and large-scale NVIDIA GB-series AI server deployments.

For gamers, the concern is not that a gaming SSD is suddenly becoming an enterprise drive. The issue is what happens to the broader NAND supply chain when data-center customers become such a massive source of demand.

This is already part of a much bigger hardware trend. AI is no longer affecting only GPUs. It is increasing demand for memory, storage and other components throughout the server ecosystem, which can eventually put pressure on consumer products as manufacturers adjust capacity toward the more profitable enterprise market.

That does not mean every consumer SSD will suddenly double in price. But it does explain why storage prices are becoming increasingly difficult to separate from the AI boom.

And when motherboard costs, memory prices and storage demand are all moving in the wrong direction at the same time, building a new gaming PC becomes much more complicated than simply choosing the best CPU and GPU.

At the same time, Moore Threads' Lushan shows that the other side of the market is changing too. New domestic GPU companies are trying to close the technology gap, while the established hardware supply chain is being reshaped by AI demand and rising component costs.

For PC gamers, 2026 is becoming a strange combination of progress and pressure. Graphics technology is moving forward, new GPU architectures are becoming more ambitious, but putting together a complete gaming PC is getting harder to budget for.

The next few months may ultimately tell us which trend matters more: whether new hardware can deliver enough performance to justify its cost, or whether the rising cost of the components underneath it will keep pushing PC upgrades further out of reach.