AI Policy Reshapes Landscape, Stargate Advances, Physical AI Emerges, Meta Hires ChatGPT Co-Creator, & Open-Source Models Get Smarter
The White House published their AI Action Plan, OpenAI & Oracle advance Stargate, AEye's Physical AI powered by Nvidia, Meta hires ChatGPT Co-Creator, and Alibaba open-sources leading reasoning LLM
AI Action Plan Pushes Deregulation and Global Dominance
The White House launched a sweeping "AI Action Plan", signing three executive orders aimed at accelerating U.S. AI development by slashing regulations, expediting data center permits, and promoting American AI exports.
The plan prioritizes "free speech and American values" in AI, seeking to eliminate references to "misinformation" or "Diversity, Equity, and Inclusion" from federal AI guidelines, including NIST's AI Risk Management Framework.
Supporters, including Nvidia’s Jensen Huang and Microsoft, praise the deregulation, but critics like the AI Now Institute warn it’s a “giveaway to Big Tech,” potentially exacerbating job losses and unchecked AI risks.
This policy marks a clear ideological divergence from previous administrations, emphasizing rapid deployment over precautionary safety frameworks, with significant implications for how AI is built and governed in the U.S.
OpenAI & Oracle Advance Stargate with 4.5 GW Partnership
OpenAI and Oracle announced a new agreement to develop an additional 4.5 gigawatts of AI data center capacity in the U.S. This latest commitment brings the total Stargate capacity under development with Oracle to over 5 GW, including the existing Stargate I site in Abilene, Texas.
This substantial build-out is designed to house over 2 million AI chips, primarily NVIDIA GB200s, with Oracle having already begun delivering racks and early training workloads running at Stargate I.
While the original $500 billion, 10 GW vision for Stargate was announced in January, this 4.5 GW update marks a concrete step forward, signaling a more structured approach to achieving OpenAI's ambitious compute goals. The expansion is projected to create over 100,000 jobs across construction and operations.
However, despite OpenAI's momentum with Oracle, reports indicate that SoftBank's involvement in the broader Stargate initiative is facing headwinds. While SoftBank was initially slated to be a major equity funder, some of its investors are reportedly getting cold feet due to the sheer scale of the investment, the intense capital demands, and the long payback periods associated with such massive AI infrastructure projects.
This hesitation could force SoftBank to reassess its financial commitment or seek greater syndication with other investors, potentially shifting more of the financing burden onto OpenAI or its other partners. The evolving dynamics of this colossal project highlight the immense financial and logistical challenges of building the next generation of AI compute.
AEye Launches OPTIS, a "Physical AI" Solution Powered by NVIDIA
AEye, creator of the Apollo lidar sensor, officially launched OPTIS, a complete "physical AI" solution designed to transform smart transportation, safety, and security.
Powered by NVIDIA's Jetson Orin platform, OPTIS combines AEye's long-range lidar with advanced computing to deliver high-resolution 3D perception and real-time responsive action in physical environments.
This launch signifies a crucial step in bringing AI beyond the digital realm and into tangible, real-world applications. OPTIS is being field-deployed with partners, targeting autonomous vehicles, traffic management, and rail safety.
It leverages NVIDIA's edge AI capabilities, positioning physical AI as a new frontier for innovation and offering industries the ability to enhance legacy infrastructure with intelligent, connected systems. The success of these solutions will be key to unlocking new revenue streams and operational efficiencies in complex physical settings.
Meta Hires ChatGPT Co-Creator
Meta continues their aggressive talent acquisition spree, announcing the hiring of Shengjia Zhao, a co-creator of ChatGPT and former lead scientist at OpenAI, as Chief Scientist of their Superintelligence Labs.
Zhao's move to Meta highlights the allure of the company's significant resources and its ambitious push into cutting-edge AI research.
And while the numbers seem big, Zuckerberg isn’t exactly betting the company on this talent. We previously wrote that he hasn’t even spent 1% of Meta’s enterprise value bringing these people on board. If Meta doesn’t succeed, investors have only lost 1% from this hiring.
For Meta, securing a talent of Zhao's caliber is crucial for accelerating its AI development roadmap and competing directly with OpenAI and Google at the frontier. This acquisition could provide a substantial boost to Meta's efforts to build advanced AI systems, but it also reflects the extraordinary lengths companies are going to gain a competitive edge in the global AI race.
Alibaba Open-Sources A New Reasoning Powerhouse
Alibaba has significantly advanced the open-source AI landscape with the release of their latest LLM, Qwen3-235B-A22B-Thinking-2507. This specialized model, part of the broader Qwen3 series, is exclusively built for "thinking mode," enabling natively extended reasoning chains for maximum depth and accuracy without manual intervention.
Their new open-source model has surpassed competing reasoning models across several critical benchmarks, including logical reasoning, mathematics, science, coding, and academic evaluations that typically demand human expertise. Its "thinking" specialization proves particularly adept at multi-step reasoning, logical deduction, strategic planning, causal inference, and abstract reasoning.
This release underscores the escalating competition in the open-source AI landscape, with non-Western tech giants like Alibaba making significant strides. By providing advanced reasoning capabilities, Qwen3 aims to democratize access to powerful AI tools, potentially challenging the dominance of proprietary models from companies like Google and OpenAI in various analytical and problem-solving applications.
The rapid progress in open-source AI suggests the gap between closed and open models is quickly narrowing, fueling a dynamic and competitive global AI ecosystem.
AI is fragmenting fast: U.S. deregulation accelerates Big Tech, Stargate shows the scale (and fragility) of AI infrastructure, “Physical AI” moves intelligence to the edge, Meta locks down scarce talent, and Alibaba’s reasoning LLM proves open-source is closing the gap. Winners will align compute, talent, and trust.