Chip firms fully commit to CES AI spam

Perhaps seeking to stoke the hype even further, Open AI CEO Sam Altman chose to blog about what a big deal AI is on the eve of CES. After reflecting on how great his company (and by inference he) is, Altman couldn’t resist a utopian turn. “We are now confident we know how to build AGI as we have traditionally understood it,” he proclaimed.

AGI stands for Artificial General Intelligence and is an ill-defined term. On the whole, it seems to refer to artificial intelligence that is equivalent to the human brain, although even that poses a whole bunch of questions, not least since we still don’t fully understand how the brain works. Having offered that teaser, he didn’t elaborate or even offer a timeline.

But apparently unsatisfied with just the one bombshell, Altman added “We are beginning to turn our aim beyond that, to superintelligence in the true sense of the word.” To be clear, that refers to AI that significantly exceeds human capabilities. For anyone who finds that concept a tad ominous, the science fiction community has been offering glimpses into possible futures containing ASI for decades.

Even more than Open AI, the company currently most intimately associated with AI is US GPU-maker Nvidia. It front-loaded a bunch of press releases for day one of CES, nearly all of which have AI in the headline. For us, the most interesting of these was titled “NVIDIA Puts Grace Blackwell on Every Desk and at Every AI Developer’s Fingertips.”

“The GB10 Superchip is a system-on-a-chip (SoC) based on the NVIDIA Grace Blackwell architecture and delivers up to 1 petaflop of AI performance at FP4 precision,” explains the press release. Nvidia’s AI platforms have been in such demand since the advent of genAI that the company has struggled to produce them quickly enough and is now the second biggest company in the world by market capitalization, behind only Apple.

That demand has been largely driven by the likes of Open AI and anyone else looking to build a large language model, since Nvidia seems to uniquely provide the enormous processing power required to do so. The novel thing about this announcement, called ‘Project DIGITS’ (not, apparently, an acronym, but that never stops Nvidia from capitalizing words), is that it brings that same platform to the desktop (pictured), thus opening up access to a lot more people.

“AI will be mainstream in every application for every industry,” said Nvidia CEO Jensen Huang. “With Project DIGITS, the Grace Blackwell Superchip comes to millions of developers. Placing an AI supercomputer on the desks of every data scientist, AI researcher and student empowers them to engage and shape the age of AI.”

Intriguingly, the CPU part of this platform is provided by Taiwanese firm MediaTek. While a disinclination to collaborate with long-time rivals Intel and AMD is understandable, it’s intriguing that Nvidia also chose to give Qualcomm the cold shoulder too. “The combination of MediaTek’s industry-leading CPU performance and power efficiency with NVIDIA’s accelerated computing technologies will drive the next wave of innovation,” said Huang.

While it made sure to tick the AI box, Qualcomm has mainly focused on automotive at CES. To be fair, the current manifestations of AI are perhaps most significant to the in-car environment as they significantly improve the voice UI experience. It’s chillingly easy to imagine a near future in which people converse with their cars as they would a passenger, making the classic 80s series Knight Rider more prescient than ever.

Intel, meanwhile, reckons it has extended its ‘leadership in AI PCs and edge computing’ at CES 2025. Inevitably, this took the form of a bunch of new chips. “Intel Core Ultra processors are setting new benchmarks for mobile AI and graphics, once again demonstrating the superior performance and efficiency of the x86 architecture as we shape the future of personal computing,” said Michelle Johnston Holthaus, interim co-CEO of Intel.

The edge angle is well summarised by Light Reading. It focuses on AI inference, as opposed to training, which is the process by which LLMs are called upon to answer questions and provide solutions. While training can be done in massive datacenters, the need for speed demands the inference processing be done closer to the end user. Having said that, Nvidia seems to be dominating that market too.

Lastly, AMD has chosen to focus on what it calls the ‘AI PC’ market at this year’s show, for which it reckons you need its processors. “As consumers and professionals increasingly recognize the productivity benefits of AI PCs, AMD is further increasing its performance leadership in the market,” said Jack Huynh, GM of AMD’s computing and graphics group. “With the next generation of AI-enabled processors, we are proliferating AI to devices everywhere and bringing the power of a workstation to thin and light laptops.”

For all the spammy hype, this clearly will be a big year for AI, especially in terms of its ubiquity. “We believe that, in 2025, we may see the first AI agents ‘join the workforce’ and materially change the output of companies,” said Altman in his blog post. While there’s little doubt AI will significantly improve productivity, the societal implications of having to coexist with such increasingly intelligent machines are likely to be very complex.