Photo via TIME

America, China, and the Costs of the Bipolar AI Race

Read the original at TIME ↗
Cite

Vashistha, A. (2026, November 21). America, China, and the Costs of the Bipolar AI Race. TIME. https://time.com/article/2026/11/21/america-china-and-the-costs-of-the-bipolar-ai-race/

Dated from the outlet's own feed. Check the original for its byline.

TIME · collected 2026-11-21 · by Aditya Vashistha

Quick Summary

The article discusses the dominance of the United States and China in the field of artificial intelligence (AI), highlighting a significant power imbalance where these two countries control much of the infrastructure, research, and development of AI technologies. For instance, the U.S. received about 72% of the world’s new AI computing capacity between 2022 and June 2026, while China received 10%, leaving only 18% for the rest of the world. The concentration also extends to energy consumption and research, with the U.S. and China accounting for about 70% of the electricity consumed by data centers worldwide in 2024. Additionally, the article notes significant disparities in language representation in AI models, with languages like Hindi, spoken by over half a billion people, receiving minimal support compared to English.
Written by AI on 2026-11-21, from this article's own text only (the story summary below draws on all the coverage).

AI analysis

Story summary

The United States and China are the dominant forces in the global artificial intelligence (AI) race, with other countries having limited influence over AI development and governance. This bipolar concentration of power exacerbates the digital divide, as decisions about AI safety, user control, and language support are made primarily by the U.S. and China. For instance, Meta’s Community Alignment project, which aims to address cultural preferences in AI, includes participants from only two countries in the Global South and focuses on English and Hindi, excluding hundreds of other languages. This exclusion can result in developers working in less-supported languages like Odia facing higher costs for processing the same content compared to more widely supported languages like English or Mandarin. As the rivalry between Washington and Beijing intensifies, many nations feel increasingly vulnerable and marginalized in the AI landscape.

Written for “Bipolar AI Race Costs” on 2026-11-22, grounded in this article and the 0 other(s) covering the same event.

Compared with similar articles

This article reads leans strongly left and hedges 8% of its claims. Each row says how that neighbour differs.
Fox News
Leans strongly right further right than this Hedges 1 of 33 claims Outlet record 66/100
“The articles discuss the broader context and ongoing competition between the US and China in AI, rather than describing the same specific incident or event.”
The Daily Caller
Leans left further right than this Hedges 11 of 32 claims Outlet record 92/100
“The articles discuss different aspects of AI and its impact, with Article A focusing on the competition between the US and China in AI development, while Article B reports on a poll about American voters' attitudes towards AI and AI chip sales to China.”
Semafor
Leans left further right than this Hedges 1 of 3 claims Outlet record 64/100
“The articles discuss the broader context and implications of the US-China AI competition rather than a specific incident or event.”
Al Jazeera
Hedges 0 of 2 claims Outlet record 60/100
“The articles discuss the broader context of the US-China AI race but describe different aspects and specific actions, indicating they are not reporting on the same specific incident.”

Story

Bipolar AI Race Costs
Technology · 1 article(s) covering the same event.
How this article was measured signals, leaning charts, publisher record, byline, topics and claims

Signals How these are calculated →

Claims extracted
93
sentences that state something as happening or true
Uncertain
8%
7 of 93 hedged ("reportedly", "may", "according to")
Leaning
Leans strongly left
of the writing, not the subject · beta estimate
Publisher signals
60/100
TIME's corrections and hedging, across everything collected from it. Not a measure of accuracy.
Outlets on this story
1
Technology
Same framing elsewhere
1
articles that frame the event the way this one does
Analyzed 2026-11-21 · how these are computed

How this is being covered How these are calculated →

Article leaning vs. publisher reliability
Source leaning vs. consistency

Publisher

TIME · 1082 article(s) · 2 correction(s) detected
Running correction rate · 2 correction(s)

Who wrote this

Aditya Vashistha
1 article(s) here · 1 carrying a prediction
Thus, developers building AI-driven apps in Hindi or Odia can face a language tax, paying more to process the same content than they would in better-supported languages such as English, Mandarin, or Spanish.
2026-11-21 · assertive framing · America, China, and the Costs of the Bipolar AI Race
The only article under this byline in the corpus.

Topics

American China Chinese The United States U.S.

Subjects

China GPE · 7× U.S. GPE · 6× American NORP · 2× Beijing GPE · 1× Carnegie Endowment ORG · 1× Chinese NORP · 1× Epoch AI ORG · 1× The International Energy Agency ORG · 1× The United States GPE · 1× Washington ORG · 1×

Narrative

Even Meta’s more recent Community Alignment project, designed to help models respond to different cultural preferences, draws on participants from only two countries in the Global South and includes English and Hindi, but none of the hundreds of other languages spoken in India. It is also difficult to reliably and systematically measure how safe these models are in local contexts.
framing: assertive · carried by 1 article(s) · first seen 2026-11-21
Thus, developers building AI-driven apps in Hindi or Odia can face a language tax, paying more to process the same content than they would in better-supported languages such as English, Mandarin, or Spanish.
2026-11-21 · TIME
America, China, and the Costs of the Bipolar AI Race · assertive framing

Claims (93 extracted, 7 hedged)

Many are comparing the technology to electricity and describing it as the industrial revolution of this century. asserted
Many → compare → century
The United States has long dominated the field, but China’s meteoric rise has made it the strongest challenger to American dominance. asserted
it → dominate → dominance
The AI race is increasingly bipolar, and much of the power to shape this critical technology is concentrated in these two countries. asserted
much → shape → countries
With the American and Chinese dominance of this emerging technology comes a widening geographic divide that begins with who builds AI technologies and where. asserted
who → emerge → technologies
As the great power rivalry between Washington and Beijing intensifies, most other countries increasingly rely on AI technologies but have little say in how those systems are built, tested, and governed. asserted
systems → intensify → say
These technologies are entering schools, hospitals, and government offices around the world, but decisions about what makes these systems safe and how much control people retain are made largely elsewhere. asserted
people → enter → control
So are the decisions about whose values frontier models reflect and which languages they support. asserted
they → reflect → languages
The bipolar concentration of AI power has heightened the sense of vulnerability outside the U.S. and China and given a new urgency to the efforts of various countries to build sovereign models of their own. AI by the few, AI of the few, AI for the few For now, the future of AI in the rest of the world depends on the U.S. and China. asserted
future → heighten → U.S.
Between 2022 and June 2026, the U.S. received roughly 72% of the world’s new AI computing capacity, according to estimates by Epoch AI, a research organization. uncertain
U.S. → receive → AI
China received 10%. asserted
China → receive → %
That left a mere 18% for all other countries combined. asserted
That → leave → countries
This concentration extends beyond AI chips. asserted
concentration → extend → chips
The International Energy Agency estimates that the U.S. and China accounted for about 70% of the electricity consumed by data centers worldwide in 2024, while data centers in the rest of the world consumed a meager 30%. asserted
centers → estimate → %
Research and development infrastructure for AI is similarly concentrated in the U.S. and China. asserted
infrastructure → concentrate → U.S.
A Carnegie Endowment study of more than 10,000 researchers who published at NeurIPS, a leading AI conference, in 2025 found that three-quarters were affiliated with employers in the U.S. or China. asserted
quarters → publish → U.S.
Even if we set aside the energy and physical infrastructure needed to build frontier models, significant gaps in training data and language coverage remain. asserted
gaps → set → data
Most of the world’s roughly 7,000 languages remain poorly supported by AI, even though about 3.4 billion people speak them. asserted
people → support → them
The data used to train frontier models and the benchmarks used to evaluate them largely leave these languages out. asserted
data → use → languages
A recent snapshot of Common Crawl, a web archive widely used to train language models, revealed that roughly 43% of the pages were in English. asserted
% → use → English
A mere 0.2% were in Hindi, a language spoken by more than half a billion people. asserted
% → speak → people
In a 2025 study, my colleagues and I found that training data for Quechua, a language spoken today by about 8 million people, primarily across South America, included texts from the era of Spanish colonization. asserted
data → find → colonization
Developers of most frontier models rely heavily on publicly available data on the internet, which contains biases and harmful content my colleague Claire Wardle described as the “armpit of the internet.” asserted
colleague → rely → internet
In fact, my research has shown that frontier models can act as a powerful force for cultural and linguistic homogenization while reproducing gender, disability, caste, and religious biases. asserted
models → show → biases
I have been studying and evaluating how AI affects the way people in different cultures express themselves in writing. asserted
people → study → writing
I have consistently found that when users in India write with AI assistants, their writing styles shift to mimic those of Americans, and their language loses some of its cultural nuance. asserted
language → find → nuance
The suggestions of AI assistants also influence and change how these users describe their cultural norms and express their views on a range of social issues. asserted
users → influence → issues
Even the cost of using AI can depend on the language you speak. asserted
you → use → language
AI companies often charge by the token, a unit of text. asserted
companies → charge → text
The same message can require more tokens in Hindi and Odia than in English. asserted
message → require → English
Newer systems have narrowed this gap but have not eliminated it. asserted
systems → narrow → it
Thus, developers building AI-driven apps in Hindi or Odia can face a language tax, paying more to process the same content than they would in better-supported languages such as English, Mandarin, or Spanish. asserted
they → build → English
Sovereign AI models of their own The rest of the world has little control over American and Chinese frontier models, and even access to them can depend on decisions made in Washington or Beijing. asserted
access → have → Washington
In June, citing national security concerns, the U.S. government ordered Anthropic to block foreign nationals—including those on its own staff—from accessing Fable 5 and Mythos 5 models. asserted
government → cite → models
Anthropic responded by taking both models offline for all users. asserted
Anthropic → respond → users
The restrictions were lifted later that month, but the episode exposed how quickly a single government’s decision can cut off access to technologies used around the world. asserted
decision → lift → world
The incident invigorated the efforts of several middle powers to build sovereign models of their own. asserted
incident → invigorate → own
A handful of middle powers have charged ahead and developed their own models: Sarvam in India, Cohere in Canada, Mistral in France, and Jais in the United Arab Emirates. asserted
handful → charge → Emirates
Their developers have sought to make these models more responsive to local needs, contexts, and languages. asserted
models → seek → needs
To gain more control over the technology and protect their national interests, the governments of India and Canada are moving quickly to invest in domestic AI firms, including through equity stakes. asserted
governments → gain → stakes
In the UAE, Jais was developed by a partnership involving Inception, a subsidiary of G42, which is chaired by Sheikh Tahnoun bin Zayed Al Nahyan, the country’s national security adviser and a member of Abu Dhabi’s ruling family. asserted
which → develop → family
…and 53 more, not listed.
Give feedback · no sign-in needed

No sign-in needed for now. Your note is sent without a name, so it cannot be answered.

History Support