Reflection AI launches open-weight model in U.S. push against Chinese rivals

Traders Union
All News
Financial News
Reflection AI open-weight model US launch

Highlights

Reflection AI, valued at $25bn and backed by Nvidia's $800mn investment, launches its first open-weight model Beam, rivaling China's GLM-5.2.
Open-weight models process 56 per cent of Vercel's AI Gateway tokens in August, up from 7 per cent in December, signaling rapid adoption over closed models.
Reflection's partnerships with the Pentagon and U.S. allies and rising Washington scrutiny on Chinese models highlight escalating policy and pricing competition in AI.

Beam debut sharpens U.S. open-model race

As reported by Financial Times, Reflection AI on Monday unveils its first open-weight model, Beam, and says it is on par with Z.ai's GLM-5.2, one of the leading Chinese competitors. The start-up says Beam outperforms U.S. peers on coding and agentic tasks at a fraction of the token cost, although it has not yet published full technical details of the model's performance.

Reflection is backed by Nvidia and has raised more than $4bn since its 2024 launch, reaching a $25bn valuation before releasing its first model. Nvidia has invested $800mn in the company and provided access to chips, while other backers include Sequoia Capital, Disruptive and 1789 Capital.

Open-weight models let users customise systems and run them on their own hardware, often making them cheaper for businesses than advanced closed models such as OpenAI's ChatGPT or Anthropic's Claude. Reflection has also partnered with the Pentagon and the U.S. Department of Energy, and has struck agreements to develop AI models for U.S. allies including South Korea.

Policy pressure and pricing stakes grow

The commercial and strategic importance of open-weight systems is increasing as their use accelerates. In August, they account for 56 per cent of all tokens processed through Vercel's AI Gateway, up from 7 per cent in December, highlighting how quickly demand is shifting toward public, customisable models.

That growth is feeding a wider Washington debate over technology competition with China. Some Trump administration officials, including Treasury secretary Scott Bessent, have floated sanctions on Chinese models over alleged intellectual property theft, while Bessent tells lawmakers last month that more American open-weight models are needed to curb the dominance of closed systems from OpenAI and Anthropic.

If Reflection's launch gains traction, it could intensify price competition among leading AI labs. U.S. companies using Chinese open-weight models are also facing scrutiny from Republican lawmakers, who are examining national security risks tied to reliance on models developed under People's Republic of China jurisdiction.

In our earlier article on Sentient’s SENT token and its OML framework, we examined how the project aims to make open AI model marketplaces monetizable through payments, staking, and governance while preserving ownership and licensing links. We also outlined key investor risks, including low token circulation, future unlock overhangs, and the need for clearer evidence of real network usage amid intense competition across decentralized AI platforms.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.