Breaking

Sacks Warns U.S. Risks Losing AI Race as China's Moonshot Tops Coding Benchmark With Kimi K3

July 17, 2026 at 11:47 EDT

Foundation Models
Policy & Regulation
Open Source

David Sacks, the AI investor and White House technology adviser, warned Friday that the United States is on track to lose the artificial intelligence race to China, pointing to the strong benchmark showing of a new open-weight model from Beijing-based Moonshot AI as evidence that the American lead is eroding.

July 17, 2026 · AI Policy

"America Is On Track To Lose The AI Race To China"

White House tech adviser David Sacks calls Moonshot AI's new open-weight Kimi K3 model "concerning" — the first Chinese system to top a major coding leaderboard, released free and priced far below U.S. rivals.

First Chinese model to top the Frontend Code Arena leaderboard

~40%

cheaper than comparable U.S. offerings

Jul 27

weights go public — free to download & modify

How narrow is the U.S. lead?

Sacks estimates America is only 3–9 months ahead of China — a gap Kimi K3 appears to be closing.

9 mo

best case

3 mo

worst case

Each block = 1 month of estimated U.S. lead over China

Beijing's playbook, per Sacks

"Stay open while behind — then close the models once you catch up."

Kimi K3 rivals Opus 4.8, GPT-5.6 Sol & Fable 5

GLM 5.2 nears Opus 4.8 on cyber & agentic tasks

Coinbase cut AI costs ~50% using GLM & Kimi

The developer case

Strong price-to-performance and easy customization of weights that can run locally — driving adoption at firms like Shopify and Airbnb.

The safety warning

Freely distributed, highly capable models pose national-security risks that are hard to contain once the weights are public.

The Washington fault line

Speed-first: "permissionless innovation," no pre-release gates

Safety-first: statutory reviews, blocks on most capable models

Sacks warns that FDA-style approvals, data-center limits and release clearances could be a "self-inflicted injury" that hands the race to Beijing.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.