Our 2025 Christmas message:

Dear Friends, Clients and Staff,

As we enter the heart of the Christmas season, a time often defined by twinkling lights, festive gatherings, and the exchange of gifts, I find myself reflecting on the deeper purpose of this holiday. I believe our success is not just measured by our year-end reports, but by the positive footprint we leave on the world around us.

The reality we must face is that while many of us have plenty — warm homes, full tables, and the security of a steady income—there are countless individuals just outside our doors who have nothing. For many, this season isn’t about celebration; it is about survival.

This year, I want to challenge each of you to give back to society. It is easy to get caught up in the consumerist whirlwind of buying expensive presents for people who already have everything they need. I am asking you to consider a different path this December.

Instead of purchasing that luxury item or another high-priced gadget, I encourage you to redirect those funds. Donate that money directly to the poor or to reputable charity organizations that are working on the front lines to alleviate suffering. Whether it is providing a hot meal, a warm coat, or medical supplies, your contribution can be the difference between despair and hope for someone in need.

We often hesitate to give because we feel our contribution is too small to make a difference. Please remember: a little bit is better than zero. You do not need to be a billionaire to be a philanthropist. A small sacrifice in your holiday budget can provide a monumental blessing to a family in crisis. Let us redefine what "abundance" means this Christmas. Let it not be the number of boxes under our trees, but the amount of help we extend to the helpless.

Thank you for your time reading this, and thank you for joining me in making this a season of true, ethical impact.

With gratitude and warmth,

Daniel Cheah



The AI Race is Shifting from Hardware to Software & Efficiency

Nvidia's dominance has been built on a simple premise - "to implement AI, you need immense computational power, and our GPUs are the best at providing it." For years, this was unquestionably true. Companies like Google, Microsoft, and OpenAI spent billions on Nvidia's H100 and B200 chips to train and run their massive models.

Google's recent "wins"—specifically with its Gemini model and the underlying Tensor Processing Unit (TPU) infrastructure—challenge this premise by proving that the game is no longer just about raw hardware power. It's about creating a more efficient, integrated, and cost-effective "full stack".

Here’s why that's bad for Nvidia's current business model:

1. The Vertical Integration Threat: Google's TPUs

This is the most direct threat. Instead of buying Nvidia chips, Google designs and uses its own custom AI chips called Tensor Processing Units (TPUs).

*  Purpose-Built Efficiency: TPUs are designed from the ground up specifically for the kind of linear algebra operations (matrix multiplications) that dominate AI model training and inference. This can lead to better performance-per-watt and lower cost than a general-purpose GPU for these specific tasks.

* Control the Stack: By controlling both the hardware (TPU) and the software (TensorFlow, JAX), Google can optimize them to work perfectly together. This software-hardware co-design is a significant advantage that a general-purpose chipmaker like Nvidia, which must cater to a wide range of customers, cannot easily replicate for any single one.

* Reducing Nvidia's TAM: Every major AI task Google runs on its TPUs is a task for which it does **not** need to buy a Nvidia GPU. As Google's AI services (Search, Workspace, Cloud, etc.) grow, its internal demand for TPUs grows, directly eating into Nvidia's potential market.

2. The Software Ecosystem Threat: Nvidia's "MoAT" is Being Challenged

Nvidia's true strength has never been just its silicon; it's its **software platform, CUDA**. For over a decade, CUDA has been the indispensable programming model for AI. If you trained a model, you did it with CUDA. This created a powerful "moat."

Google is building a compelling alternative with **JAX** and its ecosystem.

* A New Software Stack: JAX, combined with Google's TensorFlow and optimized for TPUs, is becoming a highly popular and powerful framework for cutting-edge AI research, especially for large-scale models. Many researchers now prefer it.

* Breaking the Lock-In: If the best and most efficient models (like Gemini) are built on a non-CUDA stack (JAX/TPU), it proves that CUDA is not the only game in town. This encourages other companies to explore alternatives, weakening Nvidia's strategic lock-in on the developer community.

3. The Inference Problem: Where the Real Money Is

AI has two phases:

1. Training - Building the model (requires massive compute, Nvidia's stronghold).

2. Inference - Using the model to answer queries (e.g., asking a chatbot a question).

While training is computationally intensive and gets all the headlines, inference is where the vast majority of the long-term computational cost and business revenue lies. Every Google Search, every ChatGPT query, every image generation is an inference task.

*  Inference Favors Specialization: Inference doesn't always need the brute power of a top-tier H100 GPU. It often runs better on cheaper, more specialized, and power-efficient chips—exactly what TPUs are designed for.

*  Cost is King: For a service used billions of times a day (like Google Search with AI), shaving off microseconds and fractions of a cent per query through a more efficient chip like a TPU translates to hundreds of millions of dollars in saved operational costs. Google's vertical integration gives it a massive cost advantage here.

4. The Cloud Power Shift: Competing with Your Supplier

Google Cloud Platform (GCP) is a major seller of Nvidia GPUs to its customers. But it's also the primary showcase for its own TPU v5e chips.

*  Offering an Alternative: Google can now offer cloud customers a choice: "You can rent Nvidia GPUs from us, or for many workloads, you can use our cheaper, more efficient TPUs." This positions TPUs as a direct competitor *within* Nvidia's own distribution channel.

*  The "Apple vs. Microsoft" Dynamic: This is akin to Apple controlling its entire hardware and software stack (like Google with TPU+JAX) versus Microsoft/PC makers relying on Intel (like other AI companies relying on Nvidia). The integrated model can often be more efficient and profitable.

Conclusion (Why Nvidia Isn't Doomed)

It's crucial to understand that this is a long-term threat, not an immediate collapse.

*  Nvidia is Still the King and the Pace-Setter: Nvidia's latest GPUs (like the Blackwell B200) are still arguably the most powerful AI chips on the market. The demand for AI compute is so immense that the market can support multiple winners for the foreseeable future.

*  The Broader Market: Nvidia sells to everyone: other cloud providers (Azure, AWS), sovereign nations, research institutions, and startups. Google's success does not directly impact these sales. In fact, it fuels the overall AI arms race, which benefits Nvidia.

*  Nvidia is Evolving: Nvidia isn't standing still. It's building its own cloud AI services (DGX Cloud), investing in software, and its hardware roadmap remains aggressive. It's also expanding into new areas like robotics and autonomous vehicles.

Google is proving that the path to AI dominance may not run exclusively through Nvidia's GPUs. By successfully vertically integrating with its TPUs and building a world-class software stack, Google is breaking Nvidia's perceived monopoly on high-performance AI computation. It demonstrates that superior algorithms and a tightly integrated hardware-software stack can be a more powerful and cost-effective advantage than simply buying the most raw compute power from a third party.

For Nvidia, this means the competitive landscape is shifting from being the sole provider of the "picks and shovels" in the AI gold rush to being one major player in a more diverse and competitive ecosystem. That, by definition, is bad news for a company that has enjoyed near-total dominance.

AI's Transformative Impact on Business Revenue: Opportunities and Challenges in 2026

Artificial Intelligence (AI) is reshaping the business landscape, driving unprecedented opportunities for the Asia Pacific market, especially. Strong AI adoption in many countries has led to revenue growth, increased lead generation, customer retention, and market expansion. However, it also presents challenges, such as AI talent shortages. Let's explore how this affects every business.

1. Boosting Business Revenue Through AI

AI is a powerful catalyst for revenue growth, enabling businesses to optimize operations, personalize customer experiences, and make data-driven decisions. According to recent industry insights, companies leveraging AI can see revenue increases of up to 20% by improving efficiency and customer engagement. Here’s how:

- Operational Efficiency: AI automates repetitive tasks, such as inventory management, supply chain optimization, and customer service, reducing costs and freeing up resources for revenue-generating activities. For example, predictive analytics can optimize pricing strategies, boosting profit margins by 5-10%, as seen in retail and e-commerce.

- Personalization at Scale: AI-driven tools analyze customer data to deliver hyper-personalized experiences, increasing conversion rates. Companies like Amazon use AI to recommend products, contributing to an estimated 35% of their revenue from personalized suggestions.

- Predictive Insights: AI’s ability to forecast trends and customer behavior helps businesses anticipate demand, reduce churn, and identify new revenue streams. For instance, financial services firms use AI to detect fraud, saving billions annually while enhancing customer trust.

 2. Revolutionizing Sales Lead Generation

AI is transforming how businesses generate and qualify sales leads, making the process faster and more precise. By leveraging machine learning and natural language processing, companies can identify high-potential leads with greater accuracy.

- Lead Scoring and Prioritization: AI algorithms analyze historical data, website interactions, and social media activity to score leads based on their likelihood to convert. This allows sales teams to focus on high-value prospects, increasing conversion rates by up to 30%, according to Salesforce data.

- Automated Outreach: AI-powered tools like chatbots and email automation platforms engage leads in real-time, nurturing them through the sales funnel. For example, Drift’s AI chatbot has helped businesses reduce response times and increase lead engagement by 50%.

- Behavioral Targeting: AI analyzes customer behavior across platforms, enabling businesses to create targeted campaigns. B2B companies using AI-driven lead generation tools report a 20% increase in qualified leads, as they can pinpoint decision-makers more effectively.

 3. Targeting Key Existing Customers for Upselling

Upselling and cross-selling to existing customers is a cost-effective way to boost revenue, and AI makes it more effective by identifying the right customers and tailoring offers.

- Customer Segmentation: AI clusters customers based on purchase history, preferences, and behavior, allowing businesses to target high-value segments with personalized upsell offers. For instance, Netflix uses AI to recommend premium plans, contributing to a 15% increase in subscriber upgrades.

- Predictive Upselling: AI predicts which customers are likely to purchase additional products or services based on their engagement patterns. Retailers using AI-driven upselling strategies report a 10-20% increase in average order value.

- Real-Time Personalization: AI enables dynamic pricing and real-time offer adjustments. For example, airlines use AI to upsell premium seats or ancillary services during booking, increasing ancillary revenue by up to 25%.

 4. AI Talent Shortages: A Critical Challenge

While AI offers immense potential, the shortage of skilled talent is a significant barrier. The demand for AI professionals—data scientists, machine learning engineers, and AI strategists—far outstrips supply, creating a competitive market for talent.

- Current Landscape: According to LinkedIn, AI-related job postings grew by 74% annually from 2020 to 2024, yet only 1 in 5 organizations has sufficient AI expertise. This gap delays AI adoption and increases implementation costs.

- Impact on Businesses: Companies without in-house AI talent often rely on third-party vendors, which can increase costs by 20-30%. Smaller businesses, in particular, struggle to compete for talent against tech giants offering high salaries and advanced projects.

- Solutions: To address this, businesses are investing in upskilling programs, partnering with universities, and leveraging low-code AI platforms that require less technical expertise. For example, Google’s AI training programs have helped over 2 million professionals gain basic AI skills since 2022.

 5. Market Potential and Revenue Opportunities

The global AI market is poised for explosive growth, presenting vast revenue opportunities for businesses across industries. According to recent projections, the AI market is expected to reach $1.8 trillion by 2030, growing at a CAGR of 37.3%.

Industry-Specific Opportunities: 

 - Healthcare: AI-driven diagnostics and personalized medicine are projected to generate $150 billion in annual revenue by 2026.

 - Retail: AI-powered e-commerce solutions, such as dynamic pricing and inventory management, could contribute $500 billion to global retail revenue by 2030.

 - Financial Services: AI applications in fraud detection, trading, and customer service are expected to save banks $447 billion annually by 2028.

- Emerging Markets: AI adoption in developing economies is accelerating, with Asia-Pacific and Latin America projected to contribute 40% of global AI revenue by 2030, driven by digital transformation initiatives.

- SMEs and AI: Small and medium enterprises (SMEs) are increasingly adopting AI through affordable SaaS platforms, with 60% of SMEs reporting revenue growth after implementing AI tools, per a 2024 Gartner study.

 Navigating the AI Revolution

To maximize AI’s impact on revenue, businesses must act strategically:

- Invest in AI Infrastructure: Adopt scalable AI tools tailored to your industry, such as CRM platforms with built-in AI or predictive analytics software.

- Focus on Data Quality: AI’s effectiveness depends on clean, structured data. Invest in data governance to ensure accurate insights.

- Address Talent Gaps: Partner with AI vendors or invest in training to build internal capabilities.

- Ethical AI Adoption: Ensure transparency and fairness in AI applications to build customer trust and comply with regulations.

In conclusion, AI is no longer a futuristic concept—it’s an immediately-available critical driver of business success. From using AI chatbots to AI Agents to the latest Agentic-AI apps, we are unlocking new market opportunities. AI offers a evolution in transformative potential for revenue growth. Businesses that strategically embrace AI will not only boost revenue but also gain a competitive edge in an increasingly AI-driven world.

Something was missing. So we didn't hire that person...

Sounds familiar? Some companies whinged that there's a limited number of Asia-based university graduates with real-world, practical Big Data and AI skills. This has a huge impact on hiring the right AI talent, leading to challenges for companies looking to improve using AI strategies in Asia.

Tight Talent Pool:

With a lower number of these Business+digital savvy graduates, the pool of available AI talent is reduced, creating a tighter labor market. This leads to increased hiring competition among employers for the available candidates, potentially(most likely) driving up salaries and benefits to attract the best candidates. (also, intense competition for talent will shorten the interview process, which could be good or bad)

Skills Mismatch:

If the graduates produced by universities do not have the robust data analysis skills, business acumen and technical AI skills that employers are looking for, this skills mismatch can result in graduates being underemployed or unemployed. Employers struggle to find suitable candidates for AI-related roles.

In conclusion, skills mismatches and challenges to adapt to changes caused by AI innovations can be crucial for sustainable growth of companies. After all, AI is everything now. (until the next level when GenAI becomes sentient...)

"Success breeds complacency. Complacency breeds failure. Only the paranoid survive." - Andy Grove, ex-CEO of INTEL

https://lnkd.in/gaZYB9jz