If you have been holding off on a GPU upgrade, hoping the market would eventually come to its senses, Jensen Huang has some news for you. The Nvidia chief executive confirmed during the company's latest earnings call that gaming graphics card supply will remain constrained for the foreseeable future. "As much as we would love to have more supply, we do believe for a couple of quarters it is going to be very tight," Huang told financial analysts and investors. The honest brevity of that statement is almost refreshing. Almost.
The timing of the warning matters. The Nvidia chief warned that supply of gaming graphics cards will be tight in the first half of the company's fiscal year, and that it has limited visibility for the second half. In other words, the company's own leadership cannot tell you when things will improve. Huang noted that "if things improve by the end of the year, there is an opportunity to think about what that is from a year-over-year growth, but it is still too early for us to know."
The AI elephant in the room
The root cause of this shortage is not a mystery. The explosion of artificial intelligence and massive data centre build-outs has created an insatiable hunger for Nvidia's chips, a market so profitable that it is rapidly consuming the available supply that would have traditionally gone to gamers. Nvidia manufactures its consumer RTX 50-series Blackwell chips on TSMC's 4nm process node, the same foundry pipeline feeding its AI accelerator business. When wafer production capacity is finite, and data centre clients pay significantly more per chip, the commercial logic is grimly straightforward.
The practical consequences are already visible in Australian retail. The RTX 5090 launched locally with an RRP of $4,039 AUD, with the RTX 5080 at $2,019 AUD. Those figures have not aged well. Prices for the cheapest RTX 5090 cards in Australia rose 15.2 per cent in just two months, from $4,832 in November 2025 to $5,566 in January 2026, with the pace of increase accelerating. For context, that is roughly a $1,500 premium over the official launch price, and the supply situation has not materially improved since.
Across the entire RTX 50 series, the average global price increase has been around 19 per cent over the past three months, effectively pushing each model up by a tier. In November 2025, a $1,000 budget would have secured an RTX 5080; by February 2026, that same money buys only an RTX 5070 Ti. That is a significant slide for consumers who budget carefully for hardware upgrades.
Nvidia's proposed solutions: resurrection and restraint
To his credit, Huang did not simply shrug. When pressed by journalists on possible remedies, he suggested that Nvidia could potentially bring the latest generation AI technology to previous generation GPUs, acknowledging it "will require a fair amount of engineering" but is "within the realm of possibility," adding that he would "go back and take a look." The idea of reviving older architectures on less-contested process nodes is practically sound, though the timeline for any such move remains unclear. Sketchy rumours have floated around about a potential return of the Ampere-based RTX 3060 in 2026.
Nvidia is also now producing chips at TSMC's Arizona facility. "We are now running production silicon in Arizona," Huang confirmed. Reports indicate Nvidia would manufacture Blackwell AI GPUs in Arizona, and even if that is not immediately promising for gaming card availability, it could have a knock-on effect by freeing up capacity elsewhere in TSMC's global network for consumer products. That is a long chain of ifs, but it is the most credible structural improvement on the horizon.
The counterargument: is this really a shortage?
A fair number of industry observers are sceptical of the framing. If you accept that Nvidia controls the overwhelming majority of the discrete GPU market, the argument that supply is being squeezed by external forces starts to look somewhat convenient. The company is, after all, allocating its own wafer orders. Nvidia puts the majority of wafers toward server and AI customers, as it is more profitable. That is a rational business decision, not a natural disaster. Framing it as an unavoidable shortage rather than a deliberate prioritisation choice is a distinction worth keeping in mind.
There is also a competitive dimension that benefits consumers. Across AMD's four RX 9000 series cards, the average price increase over the past three months is around 10 per cent, compared to an average 18 per cent increase across Nvidia's directly competing models, leaving AMD in a more favourable position for now. For Australian gamers who have dismissed AMD's Radeon cards in recent cycles, the current pricing environment is a genuine prompt to reconsider. Intel's Arc lineup has been the least affected by recent price increases, with the Arc B570 actually cheaper in some regions now than it was last year.
What this means for Australian PC builders
The blunt reality for anyone shopping for a new graphics card in Australia right now is that patience is not obviously rewarded. Given the short supply of gaming graphics cards for desktops and gaming GPUs for laptops, expect gaming hardware to come at elevated prices. That hits harder in Australia, where import costs and the exchange rate already push prices above US equivalents before any supply premium is added.
The broader issue here is one of market structure. When a single company commands the kind of market dominance Nvidia currently holds, and that company also serves a far more lucrative enterprise customer base, consumer market outcomes are not going to be the primary consideration. That is not a conspiracy; it is just how capital allocation works. The Australian Competition and Consumer Commission has not flagged any concern about GPU market dynamics, and there is no obvious regulatory lever to pull here.
Reasonable people can disagree about whether Nvidia's prioritisation of AI chip production over consumer GPUs reflects smart strategy, shareholder obligation, or something less flattering. What is harder to dispute is the outcome for the person trying to build or upgrade a gaming PC in 2026: the cards are scarce, the prices are high, and the company most responsible for both conditions has told you, with admirable candour, that it does not expect much to change for at least six months. Plan accordingly.