This week many of you reached out to me with questions about DeepSeek. There is some fear, confusion, and a little bit of excitement.

The Americans had a bit of a shock, because they thought they were the Kings of AI with all their Executive Orders… but enter tiny DeepSeek.

A New AI Whale Has Surfaced: DeepSeek’s Disruption Implications

Why the Real Competitive Advantage Lies Beyond the AI Foundational Models

AI is no longer just for big tech companies. Powerful AI models are now open-source and widely available, meaning anyone—startups, researchers, or even individuals—can build with them. This changes everything.

Before, AI development was limited to those with massive budgets and specialised teams. Now, with free or low-cost models and accessible tools, the biggest advantage is no longer who owns the technology but who uses it best.

This shift opens up huge opportunities:
✅ Lower costs to build AI-powered products
✅ More innovation in industries that AI giants ignored
✅ New ways to personalise, fine-tune, and apply AI for real-world impact

But AI isn’t fully a commodity—yet. Compute power, specialised data, and fine-tuning still create advantages for well-funded players. The real question is: How will you use AI to create something unique?

1. The Shock of DeepSeek’s R1 Model

DeepSeek, a Chinese startup with fewer than 200 employees, accomplished what few thought possible: building an AI model that rivals OpenAI’s GPT-4 and Meta’s Llama 2 but at a fraction of the cost.

This has sent shockwaves through the industry. Tech leaders assumed AI breakthroughs required billion-dollar budgets and vast computational resources. DeepSeek shattered that illusion. Their approach wasn’t just more cost-effective—it exposed inefficiencies in the way large AI companies operate.

But the shockwaves didn’t stop at the AI industry. DeepSeek’s success has intensified discussions about AI’s role in geopolitics and global competitiveness:

• The $100 Billion AI Race: Former U.S. President Donald Trump and Masayoshi Son (Tanye: A few weeks ago we didn’t know who he was, but now he is a superstar investor )have reportedly discussed a $100 billion AI investment plan to bolster American AI dominance. This reflects growing concerns that U.S. AI companies, despite their vast resources, are proving inefficient compared to lean, focused teams like DeepSeek.

• AI as a National Security Issue: The rise of smaller AI players raises new challenges for governments. With AI shaping defence, cybersecurity, and economic power, the notion that only a few Western companies control cutting-edge AI is fading. The AI landscape is now multipolar.

• The Efficiency Problem: DeepSeek shattered the belief that groundbreaking AI requires billions of dollars and massive teams. Their model proved that small, agile groups—if highly skilled and well-resourced—can outperform even the biggest players. [As an Industrial Engineer, I love it when efficiency wins - the big guys will have to question HOW they work, because more money is not always a good thing]

DeepSeek has proven that AI innovation is no longer just about scale—it’s about efficiency, focus, and execution.

2. The Power of Small and Focused Teams

DeepSeek’s success illustrates the strength of small, focused teams. Unlike massive corporations bogged down by bureaucracy and inertia, smaller teams can move fast, take risks, and adapt quickly.

In DeepSeek’s case, this meant focusing intensely on efficiency—developing their model with resource optimisation as a core strategy. This lean approach wasn’t a limitation; it was their advantage.

The lesson? In the AI revolution, size and investment no longer guarantees success—execution and strategy does.

3. Open-Source: Accelerating Progress

DeepSeek’s decision to open-source its R1 model accelerated its global impact. Researchers, startups, and innovators everywhere gained access to powerful tools, driving rapid collaboration and new applications.

Open-sourced AI means that AI models are freely available for anyone to use, modify, download and improve. Instead of being locked inside private companies, these models can be shared, studied, and adapted by researchers, startups, and developers worldwide.

The advantages of Open Sourcing AI also means:
✅ Lower costs – You don’t need billions to access cutting-edge AI.
✅ Faster innovation – People can build on existing AI models instead of starting from scratch.
✅ More competition – Smaller players can challenge big tech by creating better applications.
✅ Lower barriers to entry – Open-source models and API-driven platforms allow small teams, startups, and even solo developers to create powerful AI-driven solutions without massive upfront costs.
✅ A Surge in Creativity – With AI tools available to a broader audience, new applications are emerging in areas previously overlooked by big tech.

But open-source AI also raises new challenges, like who controls data, who profits from improvements, and how to prevent misuse. The AI revolution is here—but how it unfolds will depend on who builds what with it.

But there are big RISKS too…

4. Challenges of an Open Source Model and the Risks of using DeepSeek

However, open-source isn’t without its CHALLENGES:

• ❓Monetisation Challenges: Giving away technology can make it harder to sustain operations. DeepSeek will still need to figure out how they will make money to compete with big tech.

• ❓Competitor Exploitation: Larger players can repurpose and commercialize open-source models.

• ❓Ethical Concerns: Tools can be misused in ways that are difficult to control.

There are also RISKS concerning DeepSeek that users need to consider:

[These are the risks that we have been made aware of at the time of writing this article. I am sure many more will still surface. It is your responsibility to do the homework before you use a tool]

• ❌ Data Privacy Concerns: Concerns about how DeepSeek collects, stores, and processes user data, including potential risks of unauthorised access or misuse.

• ❌ Censorship: Possible biases or restrictions in content moderation that could limit free expression or filter certain viewpoints by the Chinese Government.

• ❌ Intellectual Property (IP) Protection: Potential issues regarding the ownership of generated content, reuse of proprietary data, and compliance with copyright laws.

There are always risks when handling sensitive data. Companies need to educate their employees and put the right frameworks in place.

5. The AI Commoditisation Era: Tools for Everyone

Despite these trade-offs, open-source aligns with the spirit of democratisation—making AI a shared resource for solving global problems.

Shifting the Value Chain

As AI models become commodities, their intrinsic value is diminishing. The real competitive edge no longer lies in owning an AI model—it’s in how the model is applied, fine-tuned, and integrated into real-world workflows. Companies that master this shift will define the next generation of AI-driven businesses.

The Rise of New Business Models

As AI becomes a widely available utility, expect an explosion of new monetisation strategies, including:

• LLM-as-a-Service: AI platforms offering pay-as-you-go pricing tiers based on usage, features, or scalability.

• Industry-Specific AI: Highly specialised AI models fine-tuned for sectors like legal, healthcare, finance, and urban planning. [I think there is massive opportunity here. Do you have industry expertise? Turn it into an asset it - chat to us]

• AI-Driven Collaboration: Intelligent tools that enhance human-machine interaction, optimising workflows across industries.

• Proprietary Datasets: While foundational models are becoming widely available, high-quality, proprietary, and well-structured data is what will truly differentiate AI-powered businesses. Check out our Article on Data.

  • Do you have your data sorted out? Then you have an advantage!

We are rapidly approaching a future where foundational AI models are freely available, and differentiation happens at the levels of application, infrastructure, and fine-tuning.

However, AI hasn’t fully become a commodity—yet. Here’s why (Tanye:Yes this is technical- ask DeepSeek or ChatGPT to explain it- they do a good job.):

• Open-Source Pressure: Models like DeepSeek, Llama, and Mistral are driving the cost of AI down, but this doesn’t mean all models are free. Proprietary models still offer advantages in performance and security.

• Compute & Data Constraints: Training state-of-the-art (SOTA) models still requires immense computational power and proprietary datasets—giving an edge to well-funded AI labs like OpenAI, Anthropic, and Google DeepMind.

• Customisation & Fine-Tuning: While base models are open-source, creating high-value AI applications requires additional fine-tuning, Reinforcement Learning from Human Feedback (RLHF), and access to proprietary industry data.

• Infrastructure & Inference Costs: Even free AI models require significant resources to run at scale. Companies that control optimised AI infrastructure (NVIDIA, AWS, Microsoft) still maintain a strong competitive advantage.

So, while foundational AI models are moving toward commoditisation, the real value—and competitive advantage—will come from how AI is applied, customised, and integrated.

4. What Does This Mean for the Future?

With AI models becoming widely available, the real opportunity shifts to those who can build on top of them.

This is what excites me. We have this powerful new technology, now we have to figure out how we improve the way we work using AI —> There are so many opportunities. [I will write about this next week].

Here’s what’s next:

• Increased Accessibility – AI tools will become as easy to access as cloud computing, allowing individuals, startups, and businesses to innovate without deep technical expertise.

• The Rise of Builders – The competitive edge will no longer come from owning AI models but from how they are applied. The best ideas, not the biggest budgets, will win.

• Shifting Value Chain – AI’s value will migrate away from the models themselves and toward the applications, integrations, and user experiences built around them.

• New Business Models – We will see new monetization strategies, such as:

  • LLM-as-a-Service – Pay-as-you-go AI access with different tiers for performance and customisation.

  • Industry-Specific AI Solutions – AI fine-tuned for legal, medical, financial, or creative industries.

  • AI Infrastructure Providers – Cloud providers will compete on inference efficiency, cost, and accessibility.

In short: AI is no longer the differentiator—what you build with it is.

4. The Real Opportunity: Applying AI

While AI models are moving toward commoditization, true differentiation will come from how AI is applied, refined, and integrated into real-world solutions. The companies that master this shift—by developing practical, scalable, and user-friendly AI applications—will be the ones that win in the long run.

And this is what excites me most. [I will write about this next week].

We now have access to incredibly powerful AI technology, but the real challenge—and opportunity—is figuring out how to use it to improve the way we work, create, and innovate. There are endless possibilities, and I’ll be diving deeper into them next week.

The Rise of Builders

The competitive edge in AI is no longer about owning the model. It’s about how well you apply it.

✅ The best ideas—not the biggest budgets—will win.
✅ The AI race is shifting from model training to building experiences, workflows, and businesses on top of AI.
✅ AI itself is no longer the differentiator—what you build with it is.

The playing field is open. The AI models are here. The tools are available. The real differentiator is what you choose to build.

The world is waiting. What will you create?

Contact us is you need support bringing AI into your business:

• Workshops

• AI Needs Assessment

• AI & Automation Solutions