A surprising “must-have” device

Hey y’all, it’s Austin Carr in Boston. The hottest artificial intelligence product of 2023 was a computer chip that now retails for about th [View in browser]( [Bloomberg]( [by Austin Carr]( Hey y’all, it’s Austin Carr in Boston. The hottest artificial intelligence product of 2023 was a computer chip that now retails for about the price of a BMW. But first... Three things you need to know today: • China-based Xiaomi showed off its[first electric vehicle](. • Investor excitement for AI is driving the [Nasdaq 100’s best run in decades](. • Funding for Israeli startups[crashed in the quarter](. A chip that’s hard to find Usually the most sought-after hardware is a sold-out-everywhere phone or gaming console, but this year it seemed everyone in the tech industry was willing to wait months and shell out tons of cash for a product you’ll likely never see in stores or even in person: [Nvidia Corp.’s H100]( artificial intelligence accelerator. Nvidia’s chip has arguably become the most crucial piece of technology powering the AI boom. With its 80 billion transistors, the H100 is the go-to workhorse for training the large language models undergirding apps like OpenAI’s ChatGPT and has helped Nvidia dominate the AI chip market. But with such insatiable hunger for the H100, and with rivals Advanced Micro Devices Inc. and Intel Corp. slow to produce chips with comparable performance, the dependency forced big tech players in 2023 to spend increasing gobs of money in a processor arms race. A single H100 now lists for $57,000 on hardware vendor CDW’s online shop — and data centers are filled with thousands of them. When Nvidia Chief Executive Officer Jensen Huang delivered the company’s [first AI server with an older generation of graphic processing units to OpenAI in 2016](, few could’ve predicted the role these kinds of chips would play in the coming ChatGPT-inspired revolution. Nvidia’s graphics cards were then synonymous with video games, not machine learning. But Huang recognized early on that their unique architecture, adept at what’s called parallel computing, would be better suited to handling the massive and simultaneous data crunching that AI models require compared with traditional computer processors from the likes of Intel.  To support OpenAI’s efforts, investor [Microsoft Corp. ended up building a supercomputer]( with some 20,000 Nvidia A100 GPUs, the H100’s predecessor. Amazon.com Inc., Alphabet Inc.’s Google, Oracle Corp. and Meta Platforms Inc. [soon placed similarly large orders for H100s to build out their cloud infrastructure]( and data centers, which Huang now calls “AI factories.” [Chinese companies]( even raced to [stockpile weaker versions of Nvidia’s GPUs](, whose performance was limited due to US export controls on semiconductors. Chip delivery times could stretch more than six months. “GPUs at this point are considerably harder to get than drugs,” Elon Musk [joked this past spring](. Gripes aside, the blockbuster product line catapulted Nvidia’s valuation over $1 trillion and sent its revenue soaring. In its most recent quarter, sales for its [data center division reached $14.5 billion](, nearly quadruple the same period a year earlier. But the GPU bottleneck has also woken up the industry to the risks of relying on a single entity for such a critical component of their AI portfolios. To try to ease costs and find performance gains, Google has invested [significantly in its in-house TPU chips](, while [Amazon]( and [Microsoft recently showed]( off their own custom AI accelerators. Meanwhile, Intel is touting its Gaudi 2 processor as an H100 alternative, and AMD has [said its new MI300]( will open it up to a $400 billion AI chip market in the coming years.  For some tech giants, the shift into homegrown chips could make for an awkward frenemy dynamic if their products catch on. On the one hand, Amazon and Google don’t want to become overly reliant on Nvidia, but they also don’t want to harm their relationship with the world’s most valuable chipmaker and potentially jeopardize future access to the latest-and-greatest GPUs. (Huang told Bloomberg earlier this year he doesn’t mind if his biggest customers also become his competitors and that he’ll treat them no differently.) In any case, it’s not yet clear whether any of this new AI chip competition will put a significant dent in Nvidia’s lead in 2024. Last month, Nvidia announced an upgrade to its AI processor, [named the H200](, which it said will ship in the second quarter of next year. Already Amazon and Google have signed on as initial customers. —[Austin Carr](mailto:acarr54@bloomberg.net) The big story Software companies, led by Salesforce, faced a choice in 2023: boost profits or continue to grow at all costs. Investors showed there really was [no choice at all](. Looking back at 2023 OpenAI Chief Executive Officer Sam Altman was fired by his board, then reinstated and [the board was ousted in a chaotic five days](. Uber reported its first-ever GAAP operating profit, [but investors still seemed unhappy](. Take-Two Interactive thrilled Grand Theft Auto fans with just the hint of a trailer for the [next blockbuster in the game’s franchise](.  Elon Musk made 2023 all about himself. Listen to the [end-of-year episode]( of the Elon, Inc. podcast. More from Bloomberg Get Bloomberg Tech weeklies in your inbox: - [Cyber Bulletin]( for coverage of the shadow world of hackers and cyber-espionage - [Game On]( for reporting on the video game business - [Power On]( for Apple scoops, consumer tech news and more - [Screentime]( for a front-row seat to the collision of Hollywood and Silicon Valley - [Soundbite]( for reporting on podcasting, the music industry and audio trends - [Q&AI]( for answers to all your questions about AI Follow Us Like getting this newsletter? [Subscribe to Bloomberg.com]( for unlimited access to trusted, data-driven journalism and subscriber-only insights. Want to sponsor this newsletter? [Get in touch here](. You received this message because you are subscribed to Bloomberg's Tech Daily newsletter. If a friend forwarded you this message, [sign up here]( to get it in your inbox. [Unsubscribe]( [Bloomberg.com]( [Contact Us]( Bloomberg L.P. 731 Lexington Avenue, New York, NY 10022 [Ads Powered By Liveintent]( [Ad Choices](