Lean and mean is how some see the latest major AI platform, and it could lead to a change in perspective when it comes to AI and energy use.
With the advent of Chinese company DeepSeek’s open-source AI platform, the US tech industry momentarily panicked. Big Tech stocks slid after the news that DeepSeek was opening its tech to the world, for free, as well as the claim that the company was able to develop its model cheaply and efficiently.
Dan Ives, a senior analyst at Wedbush, is bullish on tech stock, and downplayed the stock market jitters and pointed to the potential for the future of AI that the new model could mean.
“Open source is a healthy thing, because eventually LLM models are going to be monetized,” Ives said. “The spending is in the data, the algorithms, and the ecosystem; this is just part of a broader technology innovation curve that’s happening.”
With ease. One point of innovation—efficiency—could have ramifications for the US economy far beyond the tech sector, Shane Buckley, Gigamon CEO, told IT Brew, but not necessarily at the cost of consumption.
“TV didn’t kill the radio and the internet didn’t kill TV—it’s exactly the same paradigm we’re dealing with now,” Buckley said. “From an efficiency perspective, I don’t think it changes, necessarily, the amount of consumption that would be required for this technology.”
Top insights for IT pros
From cybersecurity and big data to cloud computing, IT Brew covers the latest trends shaping business tech in our 4x weekly newsletter, virtual events with industry experts, and digital guides.
Energetic ideas. David Wilson, CEO of Energy Exemplar, told IT Brew that while it’s too early to make any definitive predictions on the industry’s path forward, he doesn’t predict that there will be a major dip in energy use and AI deployment. DeepSeek’s slashing of “effort”—another way of saying it’s more energy efficient—may result in an increase of the tech’s capabilities, but the desire for AI computing isn’t going to slow.
“There’s a shortage of compute already; even if you did save that order of magnitude, which, let’s see if it plays out that much, there is still future demand,” Wilson said. “We’re in a compute constrained environment at the minute, and the assumption is that if there’s more compute available, they’ll just train bigger models and put more use cases on all of these.”
One thing Wilson said he was relatively confident of was that moving back to high polluting energy like coal for the demand is unlikely. But President Donald Trump has boosted coal as a possibility for meeting energy demands, telling Davos attendees on January 23 that “we have more coal than anybody.”