We Spent Decades Cleaning Up Industrialisation. From a Finance Lens, AI Must Not Repeat the Same Costly Mistake.✎ Edit

I began my finance career in the 1990s, when Malaysian manufacturers were expanding fast and environmental compliance was still treated as a deferred cost rather than a core line item. One of the industrial belts I worked in had rivers that regularly failed discharge standards and air that carried a visible load.

Malaysia was not the only one. Across Asia, factories scaled rapidly, agriculture intensified, livestock operations grew, and waste controls remained weak for years.

By the 1980s and 1990s, environmental awareness became harder to ignore. Cleaning rivers, improving air quality and changing industrial practices required regulation, capital works, monitoring systems and, most importantly, time that had already been lost.

I still remember a supplier audit in Busan, Korea in 1998. Heavy rain mixed with severe pollution made the working air almost unbreathable, and the respiratory illness that followed kept me away from the office for weeks. It was a clear lesson: environmental risk is a productivity and personnel cost, not just a regulatory one.

From the 2000s onward, progress became measurable. Wastewater treatment plants, emissions controls, renewable energy capacity, cleaner logistics and electric vehicles all expanded. Governments and enterprises in China, Korea, Malaysia and elsewhere poured capital into correcting the environmental balance sheet created during decades of uncontrolled industrialisation.

Today, many cities are far cleaner than they were thirty years ago.

That progress should remind every SME owner and finance manager of one thing: environmental damage is avoidable, but preventing it at the design stage is materially cheaper than remediating it later.

This is why I now scrutinise the way Artificial Intelligence is being procured and deployed. Large models are frequently routed to routine tasks, inference is repeated without reuse, and systems are assembled without proper routing, segmentation or deterministic processing. Behind every AI request sit GPUs, servers, electricity, cooling systems, water and depreciating physical infrastructure. AI may look like pure software, but its total cost of ownership and environmental footprint are very real.

The problem is not AI. The problem is using AI without financial and operational discipline. We spent decades and enormous capital cleaning up after uncontrolled industrialisation. We should not repeat the same mistake with Artificial Intelligence. At AINNA, we apply the NeuralOps principle: use advanced AI only when advanced intelligence is genuinely required. Smart Routing, specialised models, detached systems and deterministic processing reduce unnecessary compute, lower energy opex, extend hardware life and protect against carbon liabilities. For Malaysian SMEs, that discipline translates into lower utility bills, better asset utilisation and compliance-ready carbon accounting. Build more intelligence. Use less unnecessary compute. Design responsibly from the start.