Here’s what’s on our plate today:

• 📊 U.S. deregulates, China open-sources—inside the superpower face-off.

• 🚦 Treaty, UN agency, markets, or firewalls—cast your vote on the safest path.

• 🗞️ Zuck’s $165 B chip pact, Copilot tops 20 M devs, Notepad turns ‘smart’.

• 🛠️ Test Colossyan Creator, Riffusion 2 loops, and local TabbyML code-assist.

Let’s dive in. No floaties needed…

The Laboratory

How the U.S. and China are shaping the future of global AI technology

Historically, the advent of transformative technologies has led to considerable turmoil in societies and geopolitics. The last major transformative technology publicly demonstrated was nuclear weaponry, which accelerated the end of World War II. On the flip side, it led to the Cold War between two global superpowers, competing for the biggest stockpiles of nuclear warheads. The nuclear arms race would have continued to consume resources had it not been for measures such as the Nuclear Non-Proliferation Treaty (NPT) of 1968, along with a series of détentes and limitations.

Fast forward to 2025, and societies around the globe are faced with another race that demands immense resources, and in the words of the ‘Godfather of AI’, Geoffrey Hinton, has a 10-20 percent chance of wiping out humans. Though it is an extreme position, it forces leaders, regulators, and companies to seek the help of AI development to work on regulating the technology to ensure it does not do more harm than good.

The race for global AI dominance

The global race between AI companies and countries is heating up and could decide who controls the engine of future growth. The race is currently dominated by two countries: the U.S.A. and China, both of which have taken very different approaches to achieve their aims of global dominance.

The U.S. approach includes deregulation within the country, which helps AI companies dominate the global market. An example of this was seen recently, when the current Trump administration unveiled its AI plan aimed at providing a roadmap for the country’s AI ambitions. At the center of Trump’s measures are three executive orders that he says will turn the U.S. into an “AI export powerhouse” and roll back some of the rules put in place by the previous administration, which included guardrails around safe and secure AI development.

Of these, the first executive order targets what Trump calls “woke AI”, while the other two focus on deregulation, which would remove roadblocks hindering the construction of data centers and promote the export of “American AI” to other countries.

At the same time, the U.S. has been unsuccessful, leveraging its hold on high-end AI chips to limit China’s progress.

So, while on one hand, the Trump administration is reducing regulations within the country to promote exports and establish its dominance, on the other hand, it is looking for ways to slow down China’s progress in the AI race.

China, meanwhile, appears to be taking a different approach.

China’s open-source approach

Just days after Trump unveiled his AI plan, Chinese Premier Li Qiang proposed establishing an organization to foster global cooperation on artificial intelligence during the annual World Artificial Intelligence Conference (WAIC) in Shanghai. The proposal called upon countries to coordinate the development and security of artificial intelligence.

According to a report from The Guardian, the premier said that “The risks and challenges brought by artificial intelligence have drawn widespread attention… How to find a balance between development and security urgently requires further consensus from the entire society.”

Li also added that China would actively promote the development of open-source AI, and that it was willing to share advances with other developing countries in the global south.

Further, Li expressed concern that AI risked becoming the "exclusive game" of a few countries and companies, calling for the technology to be shared.

The Chinese Premier’s statement comes at a time when China has been under increasing pressure to access high-end Nvidia chips. And though the country has been working to boost its internal capacity to produce more high-end AI chips, it still lags behind the U.S. by up to two years.

In the meantime, U.S. deregulation and support of companies like Microsoft, OpenAI, Meta, Google, Nvidia, and others could create a market dominated by the West. To counter this, China has taken the approach of open science and collaborations to win over countries.

Different strength, different strategies

According to the AI Index Report from Stanford University, China leads the world in AI patent applications, accounting for about 60% of global filings. The U.S. leads in producing top AI models, but China is continuing to close the gap in quality and quantity.

Both countries have also extended state support to ensure continued growth.

China has reportedly set up a national AI industry fund of $8.2 billion to invest in early-stage AI projects, along with research and development grants, incentives, and supply contracts to AI firms.

The U.S., meanwhile, has announced the Stargate Project,  a joint venture between OpenAI, SoftBank, Oracle, and MGX, aimed at building new AI infrastructure in the country.

Both countries have their strengths and weaknesses that are shaping their approach to developing and deploying AI models. The U.S., though still ahead, is losing its head start; at the same time, China has shown its capacity to compete with the U.S., but needs more computing power to drive progress. These are reflected in how their internal policies have shaped up, and now their policies are bound to have an impact on the global shift towards AI adoption.

However, despite their diverging approaches to AI regulatory philosophy, risk management, and international coordination, there is scope for alignment.

Back-channel diplomacy: A path to safer AI

According to a report from the Financial Times, U.S.-based companies, including OpenAI, Anthropic, and Cohere, engaged in secret diplomacy with Chinese AI experts to discuss the technology’s ability to spread misinformation and threaten social cohesion.

During the meeting, both sides discussed the risks of the emerging technology and encouraged investments in AI safety research.

The talks represent a rare moment of cooperation between stakeholders from the two countries. And though the talks did not lead to a formal declaration, similar talks and cooperation in the future could pave the way for setting up a global roadmap with a focus on aligning AI systems with legal codes and norms of different societies.

Building consensus: A long road to AI governance

When the U.S. unveiled the first nuclear weapon, the ‘Little Boy’, it not only destroyed Japanese morale but also signaled the start of the nuclear age. It would take decades for the U.S.A. and Russia, then USSR, to come to a common ground and discuss disarmament. Even to this day, the process is ongoing and rests on a delicate balance of global cooperation and communication.

Now that another technology has the potential to upend the very fabric of human existence, there is a need for renewed global cooperation to reduce the risks. So, while companies chase the next advancement in AI technology, if global institutions fail to act soon, the AI race could entrench power imbalances and derail the process of sustainable development that benefits all.

TL;DR

• Two playbooks, one finish line: Washington is deregulating at home while choking China’s chip supply; Beijing is pitching open-source AI and Global-South partnerships.

• Money where the models are: Trump’s ‘AI export powerhouse’ orders and China’s ¥60 B national fund show both superpowers are going all-in on industrial policy.

• Patent vs. product gap: China now files ~60 % of all AI patents, but the U.S. still tops the leaderboard for frontier models—though that lead is narrowing fast.

• Secret safety talks: OpenAI, Anthropic, and Cohere quietly met Chinese experts to hash out AI-misinfo risks—proof that back-channel diplomacy isn’t dead (yet).

Headlines You Actually Need

• Meta Doubles Down On ‘Superintelligence’ Push — Mark Zuckerberg unveils an expanded AGI war chest—and poaches even more top researchers from OpenAI.

• GitHub Copilot Tops 20 Million Users — Microsoft’s pair-programming bot keeps devs hooked despite new challengers and rising subscription costs.

• Notepad Tests A ‘Smart’ Copilot Mode — Windows’ most basic app quietly gets an AI upgrade, auto-summarizing text while you type

Weekend To Do

• Colossyan Creator: Turn plain scripts into realistic AI video explainers in minutes—handy for pitch decks.

• Riffusion 2: Generate royalty-free music loops from text prompts; perfect background tracks for Reels.

• TabbyML: An open-source code-completion engine you can run on-device—no data leaves your laptop.

Quick Poll

Rate This Edition