Trending Now
We have updated our Privacy Policy and Terms of Use for Eurasia Group and its affiliates, including GZERO Media, to clarify the types of data we collect, how we collect it, how we use data and with whom we share data. By using our website you consent to our Terms and Conditions and Privacy Policy, including the transfer of your personal data to the United States from your country of residence, and our use of cookies described in our Cookie Policy.
{{ subpage.title }}
A reporter films at a data centre supported by Huawei, at Hongliulin Coal Mine during a Huawei-organised media tour, in Shenmu of Yulin city, Shaanxi province, China April 25, 2023.
When banned US chips are cheaper in China
Not only are Nvidia’s high-end chips getting through to the Chinese market, despite stringent export controls levied by the Biden administration, but they’re actually cheaper to access in China than in the US.
According to a new Financial Times report, Chinese data centers offering Nvidia’s AI chips charge just $6 an hour to use a server with eight Nvidia A100 processors. That same setup in the US costs about $10 an hour.
Nvidia’s H100 and A100 models are industry-standard for training and running large language models. But the Biden administration has restricted the sale of those chips and others, along with semiconductor manufacturing equipment, so that Chinese industry cannot take advantage of made-in-America technology.
It’s unclear how the chips are getting into China, but tech companies and resellers said that chips and data center inventories are openly advertised on Chinese social media and e-commerce sites — a sign that there are still gaping holes in the Biden administration’s attempts to wall off China. The Commerce Department, which oversees export controls, can’t remove what’s already gotten through, but it’ll try to keep plugging holes with future measures.FILE PHOTO: A smartphone with a displayed NVIDIA logo is placed on a computer motherboard in this illustration taken March 6, 2023.
Nvidia’s high-flying earnings aren’t good enough
Nvidia’s earnings reports have become a cultural phenomenon, with super-fan investors even throwing watch parties to tune into how high-flying the chip maker’s marks will be each quarter.
Last week, Nvidia, whose chips have fueled the current AI boom, reported more than $30 billion in sales in the second quarter of its fiscal year, up 122% from the same quarter last year. But even though it beat Wall Street analyst predictions, the stock sagged 7% after the report.
Nvidia’s stock is up 147% since the start of 2024, leading the hottest part of the stock market. While some wonder if AI is little more than a speculative bubble with AI stocks soaring, Nvidia is now the third-most-valuable company globally, and the biggest question each quarter is: How good is good enough for investors?A computer chip with the letters AI on top of it.
How China smuggles sought-after chips
The US has placed strict limits on the sale of powerful chips to China. But in the underground electronics market in Shenzhen, the southeastern port city, vendors reportedly claim to be moving hundreds or even thousands of banned chips. These include Nvidia’s A100 and H100 series chips, their most advanced models.
One vendor said he arranged a $103 million shipment to a nearby warehouse. “The Shenzhen market cannot be restricted,” he told the Times. That these middlemen are getting their hands on powerful chips is a serious threat to US economic and national security priorities, as the Biden administration is dead set on limiting Chinese access to any technologies that can fuel the government’s AI ambitions. Commerce Secretary Gina Raimondo said recently that she’s under “no illusions” that her department is executing their goals perfectly and told the Times she’s limited by budget constraints.
A Chinese military with artificial intelligence at its fingertips is a nightmare scenario for the US, and while its export controls have limited what China can make, it might never be able to fully plug leaks in the mechanism.A photo illustration of a smartphone displaying the NVIDIA Corporation stock price on the NASDAQ market, with an NVIDIA chip visible in the background.
A chip bottleneck
Margrethe Vestager, the European Union’s competition chief, has warned of a “huge bottleneck” involving Nvidia. The US semiconductor company plays a pivotal role in designing chips necessary for training and running artificial intelligence models and applications — good for 80% of the market. In recent months, Nvidia has become a $3.1 trillion company — now the third-most-valuable firm in the world behind only Microsoft and Apple.
It’s too much of a good thing: NVIDIA’s chips are so in demand that it can’t make enough for AI firms looking to train bigger and better models. EU regulators are starting to wonder whether that bottleneck raises concerns over fair competitive markets.
In Singapore, Vestager told Bloomberg that the EU’s watchdogs are asking preliminary questions of Nvidia but haven’t made up their mind about any further regulatory steps. Vestager said a robust secondary market could relieve competitive concerns, implying that as long as Nvidia respects smaller firms it should stay in antitrust regulators’ good graces. Meanwhile, AMD and Intel are looking to close the gap with Nvidia, something that intense regulatory scrutiny on the market leader might aid.
Apple wants its own chips
Apple is leveling up its chip ambitions. The Silicon Valley technology giant has spent years designing chips for its own hardware — for Macs, iPhones, iPads, and more. But, running AI models requires higher-grade chips like NVIDIA's graphics processors, which have become industry standard.
To keep up, and to fuel its own AI ambitions, Apple is working on its own chips, according to a report in the Wall Street Journal, designed to support AI applications from servers in large data centers. Internally, the project is code-named ACDC, short for Apple Chips in Data Center, though it has no set timeline for completion.
Apple's chips are reportedly meant for running AI applications, rather than training them, which makes sense given Apple's consumer focus. Apple has yielded the first leg of the AI race to upstarts like OpenAI and Anthropic, as well as to incumbents Microsoft and Meta, but the view from Cupertino is clearly better-late-than-never.
A student of National University of Malaysia walks past displays of the country's "Stripes of Glory" flags at its campus in Bangi outside Kuala Lumpur August 22, 2007.
Hard Numbers: Microsoft takes Malaysia, Massive (and unknown) startup, Safety first, Don’t automate my news
2.2 billion: Microsoft has its eye on Southeast Asia. The computing giant announced it’ll pour $2.2 billion into Malaysia’s cloud infrastructure over the next four years and will establish a national AI center with the government. This investment is the latest in a string of Microsoft infusions in local economies to help develop AI: In the past month, the company announced a $2.9 billion investment in Japan, $1.7 billion in Indonesia, and a new data center in Thailand, plus a $1.5 billion stake in the UAE firm G42.
19 billion: There’s a $19 billion AI startup that you’ve likely never heard of. It’s called CoreWeave, and it started as a small crypto company that stockpiled powerful graphics chips. Now, it runs data centers that are in high demand from AI companies that need to access those chips to run their models. It’s a company that has quickly “come out of nowhere,” as its cofounder said, to play a major role in the booming AI economy.
2: AI safety research comprises only 2% of total research about artificial intelligence, according to a new report from Georgetown University’s Emerging Technology Observatory. That’s dwarfed by global research into subjects such as computer vision (32%), robotics (15%), and natural language processing (11%).
42: In the run-up to the 2024 presidential election, 42% of Americans are concerned that news organizations will create stories with generative AI, according to a new poll from the Associated Press and the American Press Institute. While news organizations have been using AI to write simple stories — such as earnings-related stories and sport recaps — for years, those that have turned to generative AI in recent years to replace human-written stories have received public pushback and condemnation.
US Secretary of Commerce Gina Raimondo announces a major grant at the Samsung semiconductor plant in Taylor, Texas, on Monday, April 15, 2024.
Samsung hands Biden another chip win
The Biden administration is busy courting global semiconductor manufacturers to build stateside, recently handing billions to Taiwan Semiconductor Manufacturing Company to expand its chip fabrication plant in Phoenix, Arizona.
On Monday, Commerce Secretary Gina Raimondo announced that the Biden administration is giving out another award as part of its CHIPS Act budget — this time to TSMC competitor Samsung, the South Korean electronics giant. Samsung will receive $6.4 billion to put toward its new manufacturing hub in Taylor, Texas, and expand its existing plant in Austin. In return, Samsung will pour $45 billion into its US projects and commit to producing cutting-edge two-nanometer chips.
Biden has made so-called silicon nationalism a tenet of his economic and national security-focused public policy, desperate to control the slow but crucial supply of chips used for everyday technologies as well as new artificial intelligence applications.
Google's Gemini home page displayed on a smartphone in a photo illustration.
Hard Numbers: Pay for Google?, Indonesian investment, Amazon walks out on AI, Scraping YouTube
175 billion: Google said it made $175 billion in revenue from its search engine and related advertising last year, but is it ready to risk the golden goose? The company is reportedly considering charging for premium features on its search engine, including AI-assisted search (its traditional search engine would remain free). We’ve previously tested Perplexity, one of the companies trying to uproot Google’s search dominance with artificial intelligence, and you can read our review here.
200 million: The chipmaker Nvidia is teaming up with Indonesian telecom company Indosat to build a $200 million data center for artificial intelligence in the city of Surakarta, according to Indonesia’s communications minister. This news comes weeks after AI played a central role in the country’s presidential election, and it represents a major investment from one of the world’s richest tech companies in a key emerging market as Indonesia seeks to modernize its economy.
1,000: Amazon’s Just Walk Out in-store AI system for cashier-less grocery store checkout relied heavily on more than 1,000 contractors in India manually checking that the checkout transactions were accurate. Now, Amazon has announced it’s ditching the technology, which was being used in 60 Amazon-branded grocery stores and two Whole Foods stores.
1 million: One OpenAI team reportedly transcribed more than 1 million hours of YouTube videos to train its GPT-4 large language model. The company built a speech recognition tool called Whisper to handle the massive load, a move that may have violated YouTube's terms of use. YouTube parent company Google is a major rival to OpenAI in developing generative AI. Google hasn’t filed suit yet, but legal action could eventually come.