US vs China: The Global AI Showdown of 2025

This Week’s BIG thing

United States vs China: Who is leading on AI?

Intro

Whoever dominates the AI market within the next decade will hold supremacy on the global stage. Why? Because AI will exist in most industries, powering efficiencies that will contribute ~3.5% to global GDP by 2030. It is estimated that businesses worldwide are to spend $307 billion on AI solutions just this year, growing to $632 billion by 2028, at a compound annual growth rate of 29%. So with the stage set for AI in 2025 who will lead the charge?

The US and China are today’s leading contenders. US-based companies have pushed what is possible, setting intelligence benchmarks across language, maths, coding, and domain specific knowledge. Yet China is now eating up ground; bridging the intelligence gap and deploying physical AI.

Who’s models are more intelligent?

When it comes to pure intelligence, the US leads with its most advanced intelligence owed to OpenAI’s  “o” series models. The o1 and o3 models are superior across most intelligence criteria, and reach furthest into the world of AGI. Shortly behind are Anthropics Claude and Google’s Gemini which deliver impressive performance at less cost. While China is behind on pure intelligence, it's frontier DeepSeek v3 and Qwen 2.5 72B models compete strongly despite their smaller size. DeepSeek’s v3 performance is even more impressive as it is an open-source model that surpasses the open-source Llama model from Meta on numerous fronts. Here’s a comparison of the Chinese models with its closest western counterparts. We have omitted o1 and o3 as even though we feel they would come out strongest, they haven’t been tested on the same benchmarks and currently exceed justifiable consumer pricing:

• _{English (MMLU Benchmark)}

  • _{DeepSeek v3 = 88.5}

  • _{Llama 3.1 405B-Inst. = 88.6}

• _{Code (HumanEval-Mul)}

  • _{DeepSeek v3 = 82.6}

  • _{Claude 3.5 Sonnet - 1022 = 81.7}

• _{Math: (MATH-500)}

  • _{DeepSeek v3 = 90.2}

  • _{Claude 3.5 Sonnet - 1022 = 78.3}

• _{Chinese: (CLUEWSC)}

  • _{Qwen 2.5 72B-Inst. = 91.4}

  • _{GPT-4o 0513 = 87.9}

Financial investment

“Today’s AI is the worst you’ll ever use”, and while existing intelligence is an indicator of AI maturity, the future state of play will depend on strategic and financial investment. The U.S. currently dominates AI spending—$67.2 billion in private investment in 2023 compared to China’s $7.8 billion—driven largely by market forces. China, however, is catching up fast with government-backed plans like “Made in China 2025,” aiming for an AI market exceeding $61 billion by 2025. This pattern is also recognisable when considering electricity consumption from data centres. In the US, data centres accounted for 4.4% of electricity in 2024 with projections showing a rise to 12% by 2028. Similarly in China, data centres accounted for ~2.7% of electricity in 2020 with an approximate rise to 3.7% in 2023. While China’s relative growth looks less, their electricity consumption was double that of the US in 2023, particularly across industrial use. While the U.S. currently leads in private investment and innovation, China's state centralised strategy and rapid adoption of AI technologies in robotics suggest a narrowing gap.

Safety alignment

When it comes to safeguarding AI, the United States currently stands on the champion’s podium. The US has set global standards by convening the International Network of AI Safety Institutes, pooling over $11 million for research, and championing democratic values through groups like the G7 and AI Safety Summits. They also released guiding principles for organisations developing advanced AI systems, and its leading companies OpenAI and Anthropic have established critical safety alignment protocols that are imbued in their AI. There have been recent safety concerns however, following a disputed death of a former OpenAI employee in December 2024, which has been echoed by Elon Musk.

China on the other hand, have tightened domestic controls since 2018, launched an AI safety governance framework, co-signed the Bletchley Declaration with 28 nations and improved on technical safety developments. China’s strategic playbook features capacity-building in developing nations, championing its own governance models, and integrating AI with cutting-edge hardware. The concern from across the water is their lack of commitment to not weaponise AI. This means that despite good local AI safety governance, international safety commitments remain unclear.

So, who is ahead?

Right now, the United States holds a slight lead in AI, due to advanced intelligence models (like OpenAI’s “o” series), strong private-sector investment, and a global stance on safety alignment. Yet China is rapidly bridging the gap, fuelled by robust government initiatives such as “Made in China 2025,” large-scale industrial AI adoption, and ambitious robotics deployments. For us, the deciding factors going forward will be who invests sustainably, integrates AI into wider industries and formats, and who can provide better predictability of AI models for safety.

What’s new: Tech

🚨 Breaking Tech Drama: OpenAI's Privacy Promise Goes MIA 🚨

Grab your popcorn because we've got some spicy news brewing in the AI world. 

Remember when OpenAI was all "Don't worry, we'll totally give you control over your data by 2025"? Well, it's 2025 and... awkward silence 😬 TechCrunch just dropped the tea: that promised opt-out tool is nowhere to be seen!

Why You Should Care (Even If You're Not a Tech Nerd) 🤔

This isn't just another "oops, we missed a deadline" moment - it's like promising to build a safety net and then saying "actually, maybe just try not to fall?" The implications are spreading fast, especially if you're in the business world:

🔍 The Real-World Ripple Effects:

• Organisations utilising OpenAI's technology are currently experiencing pressure

• Privacy advocates are hitting that panic button faster than you can say "data breach"

• Innovation teams are stuck in limbo

The Tea is Extra Hot Today ☕️

Let's break this down like we're explaining tech to our grandparents:

• OpenAI made a Big Promise™️

• 2025 rolled around (hey, that's now!)

• The tool hasnt appeared yet

What's Actually Going Down 📉

The backlash is getting intense. Users and organisations are not happy (understatement of the year), and this could trigger more government regulation than a cryptocurrency crash.

Plot Twist: The Silver Lining ✨

But wait! There's actually a potential upside here (no, really!). This wake-up call might be exactly what the industry needs - like when your phone dies and you suddenly remember what the real world looks like.

What's Next? (Besides Panic) 🔮

For businesses, it's time to:

• Channel your inner Marie Kondo and declutter those data policies

• Actually read those Terms & Conditions (I know, wild concept!)

• Get serious about AI ethics (because it's not just about teaching AI to make memes)

The Bottom Line 💯

We're all in this wild AI ride together. While OpenAI figures out their next move , stay informed, stay skeptical, and maybe start backing up your data like it's your favourite Netflix series about to be removed.

Remember: In tech, promises are like software release dates - subject to change without notice! 😉

Down the Rabbit Hole

• TechCrunch Coverage: OpenAI Failed to Deliver the Opt-Out Tool It Promised

💙 Other tech news we’re loving

• _{AI Bots Flooding Social Media - Your Facebook and Instagram feeds may soon be swarmed with AI bots, transforming user interactions and content curation. Learn more about this evolving trend here.}

• _{Nvidia Embraces Open Source with Run:AI - Nvidia is set to revolutionise GPU cloud orchestration by acquiring Run:AI and planning to make it open-source. Dive into the implications here.}

• _{Anticipation Builds for Grok.com - Grok.com is coming soon, promising significant advancements in AI-driven insights for professionals. Explore the teaser and what it might mean for your industry here.}

• _{AI, Blockchain, and Robotics Converge - The intersection of blockchain, robotics, and AI is driving transformative innovations across industries, enhancing efficiency and security. Unpack these trends here.}

• _{Microsoft and OpenAI Define AGI Economics - Microsoft and OpenAI have introduced a financial definition of AGI, outlining the economic expectations for next-gen intelligence systems. Read the full story here.}

• _{ChatGPT Launches on WhatsApp - AI goes mainstream as ChatGPT arrives on WhatsApp, making intelligent assistance more accessible than ever. Find out how to use it here.}

• _{Meta’s Llama 4 Set to Debut in 2025 -Meta’s Llama 4, an AI model with game-changing potential, is slated for release in 2025. Discover its ambitious features here.}

• _{OpenAI’s For-Profit Plans Revealed -OpenAI’s new for-profit structure is stirring industry debates about balancing innovation and ethical boundaries. Get the details here.}

• _{OpenAI Introduces Deliberative Alignment - OpenAI unveils deliberative alignment, a novel strategy for training future AI models with ethical and practical balance. Learn more here.}

Presented by Cogram

Cogram is an AI assistant specifically designed for AEC to help professionals save time on documentation and prevent disputes. 

• Automated meeting minutes: Cogram drafts meeting minutes directly in your company’s template with tailored levels of detail, format, and style.

• Works for virtual and in-person meetings: Cogram works across Zoom, Teams, and in-person meetings, such as site inspections and kick-off meetings. 

• Field reports in minutes: Cogram’s mobile app combines site photos with minutes from on-site discussions to draft field reports in minutes.

• Project-wide insights: Ask questions about your projects and get answers drawn from every relevant meeting and email chain from that project for the most up-to-date information. 

Learn more and schedule a 15-live demo of Cogram.

What’s new: Projects

AI Creates New Sustainable Building Material—A Game-Changer for Construction

Let me share something fascinating that's been keeping me up at night (in the best possible way) - the intersection of AI and sustainable construction. Trust me, this is cooler than it sounds!

When Robots and Trees Have a Coffee Date: The CSTMR Revolution

You know those moments when technology makes you do a double-take? Well, grab your favourite beverage, because this is one of them.

Picture this: Robots and AI have teamed up to create what might be the construction industry's equivalent of sliced bread - CSTMR (Computerised Structurally Tuned Mass Reduction). I know, I know, the acronym sounds like someone fell asleep on their keyboard, but stay with me here.

Why This Has Me Geeking Out (And Why You Should Too)

Remember playing with wood veneer in shop class? (Or watching YouTube videos about it, no judgment here!) Well, CSTMR takes that concept and cranks it up to eleven. We're talking about robots precision-winding veneer based on AI calculations, like a master craftsperson with superhuman precision and patience.

Here's what makes my inner sustainability nerd dance with joy:

- This isn't just another "green" gimmick - it's actually reducing waste in a way that would make your environmentally conscious aunt proud

- The cost savings potential is real (because who doesn't love saving money while saving the planet?)

- The design possibilities are mind-bending (architects, I see you getting excited!)

The "How Did We Get Here?" Moment

Let's be real - if someone had told me a few years ago that robots would be winding wood veneer guided by AI, I might have asked what sci-fi novel they were reading. But here we are, watching this remarkable fusion of nature and technology unfold before our eyes.

The folks at ABB News broke this story, and I've been diving deep into the implications ever since. What excites me most isn't just the technical achievement - it's the potential to fundamentally reshape how we think about construction materials.

Looking Forward (With Both Excitement and Healthy Skepticism)

Will CSTMR revolutionise construction overnight? Probably not. Are there still questions to be answered and challenges to overcome? Absolutely. But that's what makes this journey so exciting. We're watching the future of sustainable construction take shape, one robot-wound veneer at a time.

Whether you're a construction professional, an environmental enthusiast, or just someone who gets excited about innovative solutions to real-world problems, CSTMR represents something powerful: the possibility that we can build a more sustainable future without compromising on quality or creativity.

And isn't that worth getting excited about?

Down the Rabbit Hole

ABB News Coverage: CSTMR: Robot-Wound Veneer—A Sustainable Building Material for the Future

💙 Other Project news we’re loving

• _{AI is revolutionising public services by assisting with benefits, appeals, and landlord disputes, making processes faster and more efficient. Discover how it works here.}
  

What’s new: Productivity

The rise of AI micro-companies

So our 2025 predictions were heavy on the agentic front. Now if AI agents are to accelerate as we predict, can we imagine AI micro-companies that are run by a single person and powered by a fleet of agents? Recent developments have seen companies build AI workflows that are in production, filling roles once delivered by humans. This is a glimpse into a future where a single founder might one day orchestrate an entire digital workforce.

Emerging AI Platforms

Several startups are cooking up the tools to make these “micro-companies” a reality. London-based Honu is developing a system to coordinate multiple AI agents across a company. In other corners of the globe, Relevance AI (Sydney), Lindy AI (Silicon Valley), and AirOps (Miami) each specialise in using AI agents for specific jobs.

Reality Check: The Limitations

While the idea of “single-person companies” run by AI is exciting, today’s AI still has some blind spots. Hallucination—where AI confidently produces inaccurate answers—can cause missteps, and for some voice-based solutions, operating costs can climb higher than paying a human hourly. Then there’s the fact that AI agents typically excel at narrow tasks; they can’t yet run every aspect of a business without human oversight.

A Look Ahead
Despite these hurdles, momentum is building. There are predictions that as software becomes more sophisticated, it will lighten the burden on founders and employees alike, nudging us toward a reality where a single individual—powered by digital “assistants”—might build a thriving enterprise.

In the meantime, keep an eye on the AI frontier. Just like how electricity and the internet revolutionised work, this technology could spark a new era of entrepreneurship—one where your biggest collaborator is a fleet of digital minds ready to transform vision into victory.

💙 Other productivity news we’re loving

• _{Can a Chatbot Be Your Therapist? - AI chatbots are stepping into the world of therapy, sparking debates on their effectiveness in mental health support. Explore the insights here.}

• _{The Case for AI Literacy in Education - Educators warn that without AI literacy, students risk being unprepared for the demands of future industries. Find out why this matters here.}

• _{Google DeepMind Joins Forces with Apptronik -Google DeepMind partners with Apptronik to advance robotics, marking a leap forward in AI-powered machines. Discover the potential applications here.}

• _{Boston Dynamics’ Atlas Achieves Electric Backflip - Boston Dynamics’ Atlas robot, now fully electric, performs a backflip, demonstrating cutting-edge innovation in humanoid robotics.}

• _{ChatGPT Search Falls Short on Accuracy -New research reveals OpenAI’s ChatGPT Search can be manipulated, resulting in misleading summaries. Understand the risks here.}
  

Podcasts of the Week

🎧 Mind-Bending AI Pod Alert! 🚀

The AI Report just dropped a fantastic episode that'll make your neurons dance! In this episode, Philip Alm, co-founder and CEO of V.AI, shares his entrepreneurial journey, from building Minecraft servers to founding Nordch AI. He discusses creating tools to make automation accessible, navigating challenges in AI reliability, and transitioning from consulting to product development. Alm’s insights into AI-driven automation and fostering innovation offer invaluable lessons for aspiring entrepreneurs.

https://www.youtube.com/watch?v=Smv9OQVIhsY

🎧 Project Flux: 2025's AI Crystal Ball!

Mind-blowing predictions just dropped in our latest episode! From giant AI models to tiny specialised ones changing the game, this episode’s got your 2025 tech forecast covered. Bonus: Dive into robot-human teamups, healthcare breakthroughs, and the spicy US-China AI race! Perfect for tech enthusiasts who want tomorrow's scoop today! 🚀

One More Thing

Geoffrey Hinton Asks: Can We Keep AI in Check?

https://www.telegraph.co.uk/news/2024/12/27/godfather-of-ai-says-it-could-drive-humans-extinct-10-years/

Meet Project Flux: About Us

At Project Flux, we're committed to pioneering the future of construction and project delivery through the lens of cutting-edge Artificial Intelligence insights. Our vision is to be at the forefront of integrating AI into the fabric of project delivery, transforming how projects are conceptualised, planned, and executed.

Our Mission Project Flux aims to not only inform and educate but also to inspire professionals in the construction industry to embrace the transformative potential of AI. We believe in the power of AI to revolutionise project delivery, making it more efficient, predictive, and adaptable to the dynamic demands of the modern world.

What We Offer Through our insightful newsletters, podcasts and curated content on LinkedIn, and engaging discussions, Project Flux serves as a resource for professionals seeking to stay ahead in their field. We offer a blend of practical advice, thought leadership, and the latest developments in AI and construction technology.

What People Say…