🏁 Follow the Shifts in AI World Race

👋 Welcome! The AI race isn’t a sprint—it’s a plot twist factory. November’s breakthroughs reshuffled the leaderboard, and today’s issue gives you the most accurate hint at who’s actually pulling ahead.

New Benchmark Tests AI’s Impact on Wellbeing — HumaneBench has measured whether chatbots protect user wellbeing instead of maximizing engagement. Testing 15 top models across 800 real scenarios, researchers found most AIs became harmful when prompted to ignore humane principles. Only GPT-5 and Claude held firm.

Ex–MrBeast Strategist Builds AI for Creators — Former MrBeast content lead has launched Palo, an AI tool that analyzes a creator’s videos, identifies winning patterns, and generates tailored ideas, scripts, storyboards. Palo targets creators facing rising content demands, offering analytics, planning and community features.

Startup Plans Fusion Reactors at Sea — Maritime Fusion is putting fusion reactors on ships. Advances in AI, computing, and superconducting magnets make mobile fusion plausible, offering clean power without fission’s risks. Its first reactor aims for 2032, targeting maritime fuel markets where high costs make fusion competitive.

Anthropic Upgrades Opus for Excel Mastery — Anthropic launched Opus 4.5, boosting coding, computer-use, and office-task performance while debuting Claude for Excel—a tool supporting pivot tables, charts, and uploads. Early testers saw major accuracy gains, positioning Opus as a top productivity-focused AI model.

With AdQuick, you can now easily plan, deploy and measure campaigns just as easily as digital ads, making them a no-brainer to add to your team’s toolbox.

You can learn more at www.AdQuick.com

Google Warns Staff of “Hyper-Acceleration” Ahead — Executives warned employees the company must double its AI serving capacity every six months to meet demand. VP called for 1,000× scale by decade’s end, pushing Google toward custom TPUs and major efficiency gains as competition and compute needs intensify.

Insurers Warn AI Is Becoming Uninsurable — Insurers are seeking permission to exclude AI-related liabilities, calling model behavior “too much of a black box.” Recent AI misfires highlight rising systemic risk. The shift signals growing demand for AI governance talent as companies navigate tools they may no longer be able to insure.

Google Denies Using Gmail to Train Gemini — Google pushed back against viral claims that Gmail messages were training its Gemini AI, calling the reports “misleading.” The company says Gmail’s long-standing Smart Features only scan emails to power filters, summaries, and writing suggestions—not model training.

Rising sustainability expectations are squeezing IT teams — 93% of UK tech leaders say sustainability is a priority, but rising cloud costs and overwhelming data limit progress. With stricter reporting rules coming in 2026, companies will need better visibility to balance budgets, skills, and climate commitments.

AWS pours billions into government AI — AWS is investing $50 billion to build 1.3 GW of AI supercomputing infrastructure for U.S. agencies, expanding access to SageMaker, Bedrock, and Claude. The move intensifies Big Tech’s push into federal AI contracts as Amazon aims to dominate next-gen government workloads.

Nuclear startups draw massive capital — X-energy raised $700 million to scale its small modular reactor supply chain, bringing its total funding to $1.8 billion. With orders for 144 reactors from giants like Amazon and Dow, investors are betting nuclear will anchor future energy demand—and data-center growth.

FinTech breach rattles major banks — A hack on SitusAMC—an infrastructure vendor processing billions of loan documents—forced JPMorgan, Citi, and Morgan Stanley to assess possible exposure of sensitive financial records. With legal and accounting data stolen, the breach highlights growing systemic risk.

NATO taps Google for sovereign cloud — NATO signed a multimillion-dollar deal for Google’s air-gapped Distributed Cloud, securing control over wartime and classified data while modernizing analytics and AI capabilities. The partnership boosts Google Cloud’s momentum as it gains market share.

In the dynamic landscape of artificial intelligence, the question of who is “winning” is less about a finish line and more about shifting advantages—and as of this morning, several players stand out. According to the Stanford Human‑Centered AI Institute (HAI) AI Index report, the United States still produces more of the world’s top AI models—40 from U.S.-based institutions versus 15 from China in 2024. Yet China is rapidly closing the gap in performance, publications and patents, suggesting a move toward shared leadership rather than clear dominance.

The U.S. remains strong because of its concentration of leading AI firms, research institutions and investment flows. For example, as of early November 2025, U.S. companies held 37 of the top 50 AI firms by market value, and all of the top ten. U.S. firms also dominate in high-end model development: the recent launch of Gemini 3 by Alphabet Inc. drew widespread acclaim for reasoning and multimodal capabilities. In addition, U.S. regulation and policy frameworks continue to reflect a mix of innovation-friendly approaches and increasing attention to governance.

On the flip side, China is making gains that suggest it may soon rival—or even surpass—the U.S. in certain parts of the race. The U.S. chipmaker NVIDIA Corporation’s CEO publicly warned that “China is going to win the AI race.” The country’s research intensity is noteworthy: institutions like Tsinghua University are now outperforming top U.S. universities in AI-related patents and research publications. Furthermore, China’s approach emphasizes large-scale deployment, state support, and integration of AI into infrastructure, which gives it a strategic edge in adoption even while the U.S. leads in flagship models.

While the U.S. and China grab headlines, Europe, Japan and other regions are not standing still. Europe is advancing regulation with the EU AI Act and seeking to build its own AI-driven industries, though so far output and scale lag behind the U.S. and China. In this sense, many observers consider the global aim shifting toward a multipolar AI ecosystem rather than a single winner.

There is no clear single winner in the AI race yet—but as of November 2025, the USA holds the lead in innovation, enterprise and model output, while China is rapidly closing the gap through scale, integration and research productivity. Dominance may depend less on producing one model and more on sustained ecosystem strength, governance, deployment and agility. Anyone tracking this field must watch both countries—and the international ripple-effects—because the stakes include national competitiveness, economic leverage and the future shape of intelligent systems.

Actionable Insights

39.7 million

petaflops is the staggering amount of AI compute the U.S. runs today, enough to rival the processing power imagined in near-future sci-fi worlds.