When video game publisher Valve released its new small form factor Steam Machine in June, reviewers and Redditors complained about the modestly-specced device’s incongruously high price tag.
At $1,049 USD, the price for an entry-level Steam Machine had jumped an estimated 43 percent since its unveiling in November of last year. Over roughly the same period, price hikes on gaming hardware became the norm. Current-gen video game systems and once-plentiful gaming PC components saw sharp markups, tanking sales as consumers delayed expensive purchases, while gaming executives hinted that the next cycle of Xbox and PlayStation consoles could sell for over a grand.
“This is the games industry's worst nightmare. Their costs are going so high that they have to launch a machine at a thousand bucks,” Dean Takahashi, veteran journalist at GamesBeat, told The Forecast.
“A lot of people would opt out of that, even the hardcore gamers,” he said. A few generations ago, “these gaming machines all used to cost around 250 bucks.”
In part, Takahashi blames the rocketing costs on the rise of artificial intelligence (AI). The construction of hyperscale AI data centers has created excess demand for advanced chips, pushing prices higher and pinching the supply of powerful gaming components.
The threat is expected to grow: according to a Deloitte report, global AI compute capacity is projected to quadruple or quintuple annually through 2030. Those figures alarm game companies that rely on modern cloud computing technologies to build, scale and process huge volumes of game-play data.
Takahashi and other industry watchers fear that ballooning prices will be the first of many disruptions for a beleaguered industry as wave after wave of AI workloads crashes down on plans and infrastructure strategies. They describe “a new reality” in which these two heavyweight technologies, AI and gaming, increasingly contend for the same chips, physical data center space, and engineering talent.
From RAMflation to Bigger Problems
One shortage for games is compute. Games rely on high-performance graphic processor units (GPUs) to generate their immersive 3D environments. These high-speed components have become harder to come by as suppliers reportedly prioritize production of specialized GPUs used for model training and inference.
“NVIDIA would much rather sell a $25,000 chip than a [gaming] GPU. There's a real shortage for gaming on that front,” said Takahashi.
Another is “RAMflation,” the consumer version of the memory and storage crises rocking the IT industry. The cumulative shortfalls have buffeted console makers who must procure all three components in bulk to assemble ready-made systems.
“AI doesn’t just compete for one component. It competes across the entire compute bill of materials,” said business strategist Scott Steinberg, who wrote a recent article for The Forecast titled Memflation Wakes the Mother of Invention for IT Teams. “GPUs,…DRAM and NAND are increasingly part of the same structural demand stack.”
Asked how he sees the physical hardware shortages panning out, he said that they could be an early indicator of a broader problem facing the games industry. He explained how game companies’ infrastructure strategies are at risk of disruption as AI hogs an ever-larger slice of global compute capacity, a figure that’s expected to quadruple or quintuple annually through 2030, according to the Deloitte report.
“Supply will eventually catch up. What may persist is a new reality where compute becomes a strategic resource, [and] AI companies compete directly with gaming companies for GPUs, data center space, and power,” he said.
It could result in “infrastructure spending rises as a percentage of budgets,” he said. “The more lasting effect may actually be power and data center constraints rather than semiconductor shortages,” he said.
Friction Between Multiplayer and Inference
Several prominent game hosting companies recently started offering cloud AI services, illustrating how data center space could become crowded. Originally designed to provide a “lag-free” experience for online multiplayer games like first-person shooters and esports, these bare metal businesses are leveraging their unique strengths to tackle AI workloads.
“Not all traditional data center providers are suited to run tasks at scale like this. So I think we will see a shift where we see new providers in the mix,” said i3d.net co-founder Stefan Ideler during a webinar titled Real-Time AI: Speed, Spend & Sovereign Control.
He says that game server providers “are built in such a way as to have the compute very close to the inference capabilities,” making them ideal for tasks like token generation that require quick turnaround. Balancing capacity between multiplayer games and inference workloads is a “logistical puzzle” that requires careful management.
“Most games have a peak in the evening of each time zone, and, then, during the day, you will have more capacity free on your dedicated inferencing hardware to utilize for enterprise workloads which peak during business hours,” he said.
“That’s the puzzle many compute providers will have to play,” he said, acknowledging that things could go sideways.
“We will also see compute providers who overbook the hell out of it, and that will lead to a degraded experience,” he said.
Fishing in the Same Pool for Talent
Beyond chips and racks of servers, there’s growing concerns about how the AI boom is draining the games industry of technical knowhow and institutional knowledge as top experts and young talent are wooed away by more lucrative paydays at AI-related companies. GamesBeat’s Takahashi points to the acqui-hire of Hathora, a gaming infrastructure provider who switched sides to focus on AI development.
Before the pivot, Hathora ran a game server orchestration platform with global reach. Its serverless model spun up multiplayer matches in containers and scaled them across a distributed edge network. Those were exactly the skills AI companies now prize and pay handsomely to acquire.
“It was clear that these guys looked at the gaming market opportunity and the AI market opportunity, and they chose to go into the AI market opportunity,” he said.
“That's another outcome where games lose out because they can't pay nearly as well as some of these AI companies can. And yet they need the same kind of talent.”
Steinberg said the Hathora example illustrates competition not merely for talent, but for infrastructure expertise itself. The same low-latency, distributed systems skills that support multiplayer games are highly valuable for AI inference platforms,” he said.
He says the acquisition suggests that game makers and AI ventures are fishing in the same pool for relatively scarce infrastructure specialists.
He says don’t be surprised if the talent poaching extends to the development side of the games business. In that case, an exodus of senior leadership could upset the intricate process of game-making, leading to greater consolidation among major publishers and a slower pace of new releases.
“Ironically, AI may partially offset that by making junior engineers more productive. But replacing a distinguished engine architect or infrastructure lead is much harder than replacing routine coding labor,” he said.
“The most consequential brain drain issue…may be the migration of systems-level engineering talent into AI infrastructure companies,” he said.
Related:
Jason Johnson is a contributing writer. He is a longtime content and copywriter for tech and tech-adjacent businesses. Find him on Linkedin.
© 2026 Nutanix, Inc. All rights reserved. For additional information and important legal disclaimers, please go here.
Related Articles
