Yesterday the Wall Avenue Journal reported that OpenAI CEO Sam Altman needs to lift as much as $7 trillion for a “wildly-ambitious” tech venture to spice up the world’s chip capability, funded by traders together with the U.A.E. — which in flip will vastly broaden its means to energy AI fashions.
Whereas this will merely be a dreamy moonshot on Altman’s half, or an Elon Musk-like hype-generator, what is just not doubtful is the environmental impression of such an enormous effort, based on Sasha Luccioni, local weather lead and researcher at Hugging Face.
“If it does work out, the quantity of pure assets that shall be required is simply mind-boggling,” she instructed VentureBeat. “Even when the vitality is renewable (which it isn’t assured to be), the amount of water and uncommon earth minerals required is astronomical.”
For comparability, in September 2023 Fortune reported that AI instruments fueled a 34% spike in Microsoft’s water consumption; Meta’s Llama 2 mannequin reportedly guzzled twice as a lot water as Llama 1; and a 2023 examine discovered that OpenAI’s GPT-3 coaching consumed 700,000 liters of water.
VB Occasion
The AI Impression Tour – NYC
We’ll be in New York on February 29 in partnership with Microsoft to debate steadiness dangers and rewards of AI purposes. Request an invitation to the unique occasion beneath.
And past the environmental impression, shortages of uncommon earth minerals similar to gallium and germanium have even helped inflame the world chip warfare with China.
Luccioni criticized Altman for not specializing in extra environment friendly AI strategies to develop AI. As an alternative, she mentioned, “he’s taking a brute power method and persons are calling it.. visionary?”
GPU entry has turn out to be key Silicon Valley AI battle
However the reality is, Altman’s need to sort out the present GPU shortages and reshape the semiconductor panorama is just not uncommon. Final summer season, VentureBeat reported on how entry to Nvidia’s hard-to-come-by, ultra-expensive, high-performance computing H100 GPU for big language mannequin (LLM) coaching was turning into the “high gossip” of Silicon Valley.
And simply final week, Meta supplied a deep dive into its AI technique in its newest earnings name, CEO Mark Zuckerberg mentioned that to construct AI “full basic intelligence” the primary key requirement is “world-class compute infrastructure.” Zuckerberg went on to repeat what he had just lately disclosed in a latest Instagram Reel: that by finish of this 12 months Meta can have about 350k H100s — together with different GPUs the entire shall be round 600k H100 equivalents of compute.
The corporate plans to proceed investing aggressively on this space, he defined: “As a way to construct essentially the most superior clusters, we’re additionally designing novel knowledge facilities and designing our personal customized silicon specialised for our workloads.”
Luccioni has been vital about Nvidia’s transparency concerning the carbon footprint of its merchandise (that are designed by the corporate however manufactured by the Taiwan Semiconductor Manufacturing Firm: “Nvidia has but to publish any details about the environmental footprint of their manufacturing,” she mentioned, including that e-waste as an entire can also be a “big situation as a result of folks need the brand new GPUs and so they’re primarily throwing out the previous ones after a 12 months or two.”
In Nvidia’s 2023 Company Accountability Report, the corporate mentioned “emissions are generated at each stage of our product lifecycle, together with manufacturing inside our provide chain. Since 2014, we’ve anticipated our key silicon manufacturing and techniques contract manufacturing suppliers to report their annual vitality and water utilization, waste, greenhouse gasoline (GHG) emissions, and discount objectives and goals by way of the RBA Environmental Survey or CDP. We additionally count on suppliers to have their GHG emissions verified by a 3rd celebration. We use this provider knowledge to higher perceive our product manufacturing impression and allocate carbon emissions to our prospects.”
Extra transparency from OpenAI is unlikely
Total, Luccioni maintains that there’s much less transparency right this moment on the subject of the environmental impression of AI — and is unlikely to alter anytime quickly with Altman’s new fundraising march.
“Should you have a look at the PaLM 1 paper from Google, which was in 2022, after which Palm 2 [released in May 2023], the quantity of data they offered drastically dropped,” she mentioned. Within the unique paper, she defined, Google shared sufficient info in order that energy-use estimates may very well be made.
“Now [companies] don’t even say how lengthy it took [to train], what number of chips they used, there’s completely no info offered anymore,” she mentioned.
However general, Luccioni says she isn’t too fearful: “I believe that is only a moonshot venture that received’t truly pan out,” she mentioned. “However that may put [Altman] on par with Elon when it comes to outlandish tasks that appeal to consideration and generate hype.”
VentureBeat’s mission is to be a digital city sq. for technical decision-makers to achieve data about transformative enterprise know-how and transact. Uncover our Briefings.
