Benzinga and Yahoo Finance LLC may earn commission or revenue on some items through the links below.

Nvidia-backed (NASDAQ:NVDA) Reflection AI on Oct. 9 announced it secured $2 billion in funding to build frontier-scale open intelligence systems that rival closed labs like OpenAI, Google, Gemini, and Anthropic. The startup says it will use the capital to develop what it calls frontier open intelligence that anyone can access.

The team includes engineers and researchers who created PaLM, Gemini, AlphaGo, AlphaCode, and AlphaProof at Google’s DeepMind. According to Reflection AI, contributors to ChatGPT and Character AI also joined the venture.

The round included participation from Nvidia, Disruptive, DST Global, 1789 Capital, B Capital Group, Lightspeed Venture Partners, Charles River Ventures, Sequoia Capital, and Citigroup (NYSE:C), according to Reflection AI.

Don’t Miss:

Individual backers include Zoom Communications (NASDAQ:ZM) founder Eric Yuan and former Google CEO Eric Schmidt.

The funding enables Reflection AI to expand development of frontier-scale models alongside leading AI research organizations.

AI is becoming the foundational layer for scientific research, education systems, energy optimization, medical diagnostics, and supply chains, Reflection AI said, warning that allowing a handful of laboratories to control this infrastructure creates a dynamic where everyone else gets locked out.

Major breakthroughs in computing history came from open collaboration, Reflection AI said.

Linux, internet protocols, and foundational computing standards all emerged from transparent development processes. According to Reflection AI, AI reached its current capabilities because researchers shared and published scaling ideas openly.

“There’s a narrow window to change this trajectory,” Reflection AI said in its statement. The company added that without intervention, a small number of entities will control the capital, computing power, and talent needed to build AI.

Trending: If there was a new fund backed by Jeff Bezos offering a 7-9% target yield with monthly dividends would you invest in it?

Reflection AI spent the past year developing a large-scale platform for language modeling and reinforcement learning, according to the company’s statement. The system is designed to train massive Mixture-of-Experts models at frontier scale, a technical capability previously seen only inside top AI research labs.

The company first tested this framework in the field of autonomous coding, a complex domain that demands advanced reasoning. Reflection AI said the experiment confirmed the system’s effectiveness and marked a breakthrough in performance. With that milestone complete, the company plans to extend its methods to broader applications in agentic reasoning.

Reflection AI also said it has established a scalable commercial strategy aligned with its open intelligence mission. The company believes this approach will support sustainable development and continued public release of frontier models.

See Also: Missed Nvidia and Tesla? RAD Intel Could Be the Next AI Powerhouse — Just $0.81 a Share

Reflection AI acknowledged that releasing highly capable models to the public introduces real safety concerns. It said it will address these risks through rigorous evaluations designed to measure each model’s capabilities and potential for misuse before any release.

According to Reflection AI, ongoing security research will focus on protecting the technology from malicious applications while defining responsible deployment standards. The company said these measures form part of a broader commitment to ensure open development does not come at the cost of safety.

“The answer to AI safety is not ‘security through obscurity’ but rigorous science conducted in the open,” the company said.

Reflection AI added that transparency enables independent researchers to identify weaknesses, develop solutions, and hold systems accountable in ways closed development cannot.

Read Next: Microsoft’s Climate Innovation Fund Just Backed This Farmland Manager — And Accredited Investors Can Join the Same Fund

Image: Shutterstock

This article Reflection AI Lands $2 Billion From Nvidia, Eric Schmidt To Build Open Alternative To ChatGPT And Gemini Models originally appeared on Benzinga.com