# Graylin and Triolo argue in MIT Technology Review that the US-China AI arms race has no winners

[Canonical reading](https://escapemeridian.com/readings/graylin-and-triolo-argue-in-mit-technology-review-that-the-us-71f7b747)

Originally published January 22, 2025, Alvin Wang Graylin and Paul Triolo describe a US-China “AI arms race” and argue that it can have no winners. They trace the US policy focus on winning economically, then say leading AI labs including OpenAI and Anthropic promoted “beating China” in what appeared to be an effort to align with the incoming Trump administration. The authors say the belief that the United States could win rested mostly on its early advantage in advanced GPU compute resources and the effectiveness of AI’s scaling laws. But they write that access to large quantities of advanced compute resources no longer appears to be the defining or sustainable advantage many had thought it would be. The excerpt offers that shift as part of the case against an arms race, but does not give the authors’ full reasoning for why there can be no winners or say what may shape the competition instead. The consequences they foresee are not stated in the available text.

Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.

## Original source

MIT Technology Review · Article

Original source date: Jan 22, 2025 (2025-01-22T13:10:00Z).

Date basis: Date from the publisher’s own records, through its API.

Added to Escape Meridian: 2026-11-16T00:43:27.015Z. First observed: 2026-11-02T08:31:56.169Z. Retrieved: 2026-11-02T08:33:20.405Z.

- [The Download: OpenAI’s lobbying, and making ammonia below the Earth’s surface](https://www.technologyreview.com/2025/01/22/1110416/the-download-openais-lobbying-and-making-ammonia-below-the-earths-surface/)

## Limits

Text as retrieved on November 2, 2026; originally published January 22, 2025. The available text is a newsletter excerpt linking to the full story; it does not provide the authors’ full reasoning for why the arms race can have no winners or what they think will shape it instead.


## Open question

If advanced compute is no longer a defining or sustainable advantage, what do Graylin and Triolo identify as shaping the US-China AI competition instead?


## Source-stated

Alvin Wang Graylin and Paul Triolo frame the US-China AI competition as an arms race in which there can be no winners.

> There can be no winners in a US-China AI arms race

- [Supporting source](https://www.technologyreview.com/2025/01/22/1110416/the-download-openais-lobbying-and-making-ammonia-below-the-earths-surface/)

## Source-stated

The authors say access to advanced compute resources no longer appears to be the defining or sustainable advantage many had thought it was.

> But now it appears that access to large quantities of advanced compute resources is no longer the defining or sustainable advantage many had thought it would be.

- [Supporting source](https://www.technologyreview.com/2025/01/22/1110416/the-download-openais-lobbying-and-making-ammonia-below-the-earths-surface/)

## Source-stated

The authors say the belief that the United States could win was based mostly on an early lead in advanced GPU compute resources and the effectiveness of AI’s scaling laws.

> The belief that the US can win in such a race was based mostly on the early advantage it had over China in advanced GPU compute resources and the effectiveness of AI’s scaling laws.

- [Supporting source](https://www.technologyreview.com/2025/01/22/1110416/the-download-openais-lobbying-and-making-ammonia-below-the-earths-surface/)

## Source and contributor context

MIT Technology Review · Publisher

The publisher is identified here. No individual author has been inferred.


## Source media

Media retrieval status: unchecked.


## Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.


## Graphics processing unit

A chip that performs many calculations at once; the main hardware used to train and run AI models.

- [Definition reference](https://en.wikipedia.org/wiki/Graphics_processing_unit)

## Scaling laws

Observed patterns in which model performance improves predictably as data, model size and computation grow.

- [Definition reference](https://en.wikipedia.org/wiki/Neural_scaling_law)

## OpenAI

OpenAI develops AI models and products and publishes research, release announcements and evaluations.

- [Explore this context](/entities/openai)
- [Reference](https://openai.com/news/)

## Anthropic

Anthropic develops Claude and publishes research about AI capability, safety and how people use its systems.

- [Explore this context](/entities/anthropic)
- [Reference](https://www.anthropic.com/news)
- [Reference](https://www.anthropic.com/research/team/economics)

## Editorial provenance

Version 1. Publication ID: f9913efc-4bf6-dc3d-98ed-a7a0c0ed9da4.

Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.

Source check: passed by anthropic/claude-opus-5.5 on 2026-11-16T00:39:22.340Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: bridge\_admitted\_by\_model, second\_opinion\_categorical.

Observation: aeb4a82ce66ef52cfeb53758b656d378efa428390abe84e96e5de1ea84fbe436. Source text hash: 256cb1f7b01d919cdc8a3aaacdb6cf0e4303e92cbe044b89a6858aceeeda8976.


## Revision and source-check history

### Version 1 — 2026-11-16T00:43:27.015Z

Originally published January 22, 2025, Alvin Wang Graylin and Paul Triolo describe a US-China “AI arms race” and argue that it can have no winners. They trace the US policy focus on winning economically, then say leading AI labs including OpenAI and Anthropic promoted “beating China” in what appeared to be an effort to align with the incoming Trump administration. The authors say the belief that the United States could win rested mostly on its early advantage in advanced GPU compute resources and the effectiveness of AI’s scaling laws. But they write that access to large quantities of advanced compute resources no longer appears to be the defining or sustainable advantage many had thought it would be. The excerpt offers that shift as part of the case against an arms race, but does not give the authors’ full reasoning for why there can be no winners or say what may shape the competition instead. The consequences they foresee are not stated in the available text.

## Original source

MIT Technology Review · Article

Original source date: Jan 22, 2025 (2025-01-22T13:10:00Z).

Date basis: Date from the publisher’s own records, through its API.

Added to Escape Meridian: 2026-11-16T00:43:27.015Z. First observed: 2026-11-02T08:31:56.169Z. Retrieved: 2026-11-02T08:33:20.405Z.

- [The Download: OpenAI’s lobbying, and making ammonia below the Earth’s surface](https://www.technologyreview.com/2025/01/22/1110416/the-download-openais-lobbying-and-making-ammonia-below-the-earths-surface/)

## Limits

Text as retrieved on November 2, 2026; originally published January 22, 2025. The available text is a newsletter excerpt linking to the full story; it does not provide the authors’ full reasoning for why the arms race can have no winners or what they think will shape it instead.


## Open question

If advanced compute is no longer a defining or sustainable advantage, what do Graylin and Triolo identify as shaping the US-China AI competition instead?


## Source-stated

Alvin Wang Graylin and Paul Triolo frame the US-China AI competition as an arms race in which there can be no winners.

> There can be no winners in a US-China AI arms race

- [Supporting source](https://www.technologyreview.com/2025/01/22/1110416/the-download-openais-lobbying-and-making-ammonia-below-the-earths-surface/)

## Source-stated

The authors say access to advanced compute resources no longer appears to be the defining or sustainable advantage many had thought it was.

> But now it appears that access to large quantities of advanced compute resources is no longer the defining or sustainable advantage many had thought it would be.

- [Supporting source](https://www.technologyreview.com/2025/01/22/1110416/the-download-openais-lobbying-and-making-ammonia-below-the-earths-surface/)

## Source-stated

The authors say the belief that the United States could win was based mostly on an early lead in advanced GPU compute resources and the effectiveness of AI’s scaling laws.

> The belief that the US can win in such a race was based mostly on the early advantage it had over China in advanced GPU compute resources and the effectiveness of AI’s scaling laws.

- [Supporting source](https://www.technologyreview.com/2025/01/22/1110416/the-download-openais-lobbying-and-making-ammonia-below-the-earths-surface/)

## Source and contributor context

MIT Technology Review · Publisher

The publisher is identified here. No individual author has been inferred.


## Source media

Media retrieval status: unchecked.


## Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.


## Graphics processing unit

A chip that performs many calculations at once; the main hardware used to train and run AI models.

- [Definition reference](https://en.wikipedia.org/wiki/Graphics_processing_unit)

## Scaling laws

Observed patterns in which model performance improves predictably as data, model size and computation grow.

- [Definition reference](https://en.wikipedia.org/wiki/Neural_scaling_law)

## OpenAI

OpenAI develops AI models and products and publishes research, release announcements and evaluations.

- [Explore this context](/entities/openai)
- [Reference](https://openai.com/news/)

## Anthropic

Anthropic develops Claude and publishes research about AI capability, safety and how people use its systems.

- [Explore this context](/entities/anthropic)
- [Reference](https://www.anthropic.com/news)
- [Reference](https://www.anthropic.com/research/team/economics)

## Editorial provenance

Version 1. Publication ID: f9913efc-4bf6-dc3d-98ed-a7a0c0ed9da4.

Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.

Source check: passed by anthropic/claude-opus-5.5 on 2026-11-16T00:39:22.340Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: bridge\_admitted\_by\_model, second\_opinion\_categorical.

Observation: aeb4a82ce66ef52cfeb53758b656d378efa428390abe84e96e5de1ea84fbe436. Source text hash: 256cb1f7b01d919cdc8a3aaacdb6cf0e4303e92cbe044b89a6858aceeeda8976.


### passed — 2026-11-16T00:39:22.340Z — anthropic/claude-opus-5.5

Publication: f9913efc-4bf6-dc3d-98ed-a7a0c0ed9da4

attribution — pass: The argument is correctly credited to Alvin Wang Graylin and Paul Triolo, whose byline appears on this item in MIT Technology Review's The Download newsletter. The newsletter is one piece in the editors' own words, not a bare listing. The limitations note that the text is a newsletter excerpt linking to the full story. The labs' motive is kept as the authors' hedged view ("what appeared to be").

> There can be no winners in a US-China AI arms race
—Alvin Wang Graylin and Paul Triolo

summary — pass: The reading leads with the actors and their thesis, then lays out the argument: the US focus on winning economically, the labs' "beating China" narrative, the compute and scaling-law basis for that belief, and the shift away from compute as the defining advantage. GPU and scaling laws are defined beside the reading. It plainly states what the excerpt leaves out. The question follows from that gap and does not lead the reader.

> In the early days of this standoff, US policymakers drove an agenda centered on “winning” the race, mostly from an economic perspective.

date — pass: The reading gives January 22, 2025, which matches the UTC day of originalPublished (2025-01-22T13:10Z) and the URL path. The dateModified (12:50 UTC) falls before publication, so there is no later edit. The page's later Deep Dive sidebar items are neither used nor relied on. There is no hindsight framing, and no forecast is presented as an outcome. The evidence is a publisher page, not an archive copy, so no archive disclosure is needed.

support — pass: All three claims quote the source verbatim and keep the authors' hedges ("appears", "mostly"). The reading's framing of the labs' motive matches the source's "what appeared to be" hedge. It invents no benefits, costs or consequences, and it explicitly says the consequences the authors foresee are not stated.

> But now it appears that access to large quantities of advanced compute resources is no longer the defining or sustainable advantage many had thought it would be.

media — pass: No media was selected, and no verified candidates were supplied, so nothing is misidentified.

## Recorded references

Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter.

No public references recorded here yet.

[Reference data (JSON)](/readings/71f7b74722e4bad49d0907aa566466c0ef8eac46e57a34b0d85c74fa130e8a0e/references.json)

## Corrections and challenges

[Send a correction or an opposing source](https://escapemeridian.com/corrections/)

[JSON record](https://escapemeridian.com/readings/graylin-and-triolo-argue-in-mit-technology-review-that-the-us-71f7b747.json)
