This article first appeared in Forum, The Edge Malaysia Weekly on October 5, 2026 - October 11, 2026

Coming back from travelling in the UK and China, I had time to reflect on how AI is transforming the macroeconomy and how small companies or economies like Malaysia can survive in this war of giants.

Artificial intelligence (AI) and robotics are not just new technologies; they are a revolutionary reorganisation of how value is created, work is performed, and resources are allocated across the global economy. Leading thinkers are forecasting that generative AI could lift global output by between 3% and 7% over a decade, adding trillions of dollars to world GDP. But this promise is not without its perils, as AI thought leaders are already warning about the dire risks of loss of human control over AI processes in terms of hallucination, hacking or cyber-meltdown. Even as agentic AI promises widespread efficiency gains, power, capital and know-how are being concentrated in the few, while job cuts are for the many and mass re-training of the existing workforce becomes key to future social stability. At the macro-finance level, AI is increasing capital expenditure and increasing energy, water and rare mineral consumption, while being funded through unprecedented debt expansion that suggests growing systemic leverage and financial instability.

Geopolitically, AI, robotics and tech operating systems are bifurcating between the two leading economies, the US and China. Simply put, we may be seeing a huge fight between Apple/Windows operating systems and Huawei HarmonyOS that will dominate future devices and tech operating ecosystems. For middle-powers and smaller and open economies, such as Canada and Malaysia, the challenge is not simply to adopt AI, but which system to choose, how to avoid being dominated or sanctioned by either dominant rival systems or creating interoperable ecosystems that avoid choosing sides.

Broadly speaking, leading think-tanks agree that AI will profoundly lift productivity, but how it diffuses gains and losses remains contentious. Goldman Sachs is most optimistic, estimating that generative AI could accelerate annual global productivity growth by 1.2 to 1.5 percentage points, double recent trend rates, with the full effect visible within a decade. The International Monetary Fund (IMF) and Organisation for Economic Co-operation and Development (OECD) project more modest gains of 0.8 to 1.2 percentage points annually, with uneven benefits accruing mostly to hyper-connected firms and advanced economies. The Bank for International Settlements (BIS) is most cautious, noting that earlier general-purpose technologies — electricity and internet — took seven to 15 years to translate into measurable economic improvement. Smaller players, such as small and medium enterprises (SMEs) and emerging and developing market economies (EMDEs), will suffer “implementation lag”, because they lack the capital, access to know-how and funding to reorganise workflows, re-train workers and build infrastructure and new rules and standards before efficiency gains appear in the macroeconomy. The digital divide will widen with huge political and social implications.

If, as estimated, between 12% and 30% of global jobs may be transformed or automated, particularly in administrative, routine knowledge-work and entry-level roles, youth and older workforce unemployment will cause a huge political backlash. Without deliberate policy and government spending even as government debt levels are at historically high levels, AI will not be a rising tide that lifts all boats; it will be a force that amplifies existing injustices.

At the same time, there is a financial peril on top of the productivity promises. The global AI buildout in hugely expensive data centres, semiconductor fabrication, model training and infrastructure requires an estimated US$1 trillion to US$1.3 trillion in capital through 2030. Historically, technology leaders funded expansion from internal cash flow and high stock market valuation. Today, as the capex and revenue gap is widening, AI investors are increasing their debt, even as global interest rates are rising. Japanese tech investor SoftBank is borrowing US$11.1 billion through junk-bond issuance to fund its US$64.6 billion AI investments. Hyperscalers alone may be issuing more than US$1 trillion in new debt over the next few years, while private credit funds are now major lenders to mid-tier AI firms.

This scale of borrowing assumes that AI will deliver massive revenue and productivity gains. If adoption slows or new technology compresses margins, or real-world benefits arrive later or are smaller than priced, refinancing will fail and write-downs will cascade. Furthermore, true leverage may be larger than reported, as operating leases, project-financed data centres and take-or-pay capacity contracts are mostly off balance sheet and can hide leverage risks.

For smaller players, higher global borrowing costs increase the cost of servicing their debt; and those that adopt the wrong technology or too expensive capex will face not only structural marginalisation, becoming dependent on the major hyperscaler platforms, but also become trapped consumers rather than beneficiaries in the AI boom.

Small players therefore face the inevitable choice of either adopting one dominant system or going for a dual-system solution of “interoperability” to maintain sovereignty. In a global digital landscape that is visibly bifurcating between Apple or Windows-led ecosystems and HarmonyOS-style distributed frameworks as parallel alternative standards, with differing governance, supply chains and geopolitical alignments, smaller economies choosing one system to the exclusion of the other means accepting dependency, surrendering data control and locking local development into a foreign roadmap.

Thus, the historical solution is not to pick a winner, but towards interoperability. Historically, Android — specifically its open-source foundation, AOSP — provides a proven blueprint for this approach. Applications built for Android can run within Windows compatibility environments, on HarmonyOS devices and on independent local platforms. Building on open-source systems allows local developers to write once and reach multiple markets.

For Malaysia and Asean, this translates into a practical resilience strategy. National digital infrastructure can be built on open, interoperable foundations rather than proprietary lock-in. The Malaysian MyInvois e-invoicing platform, simplified ESG reporting and Islamic finance digital tools can be designed as universal services accessible across all major environments. Devices and systems can support both global ecosystems while retaining local sovereignty over data, privacy and critical functions. This is “polycentric resilience”, not allowing a single authority or system to be the sole point of failure.

In practical terms, interoperability carries economic advantages of reducing the cost of digital adoption for SMEs, and if the software proves efficient and resilient, it creates space for local innovation and can be sold to other SMEs and EMDEs that face exactly the same digital dilemma of dual-system bifurcation.

In short, AI is reshaping the macroeconomy, but trapped destiny is not a foregone conclusion. Despite all its risks, AI is a tool that can drive productivity gains and open up new products, services and markets. For small players, dual-system interoperability is not just a technical choice, it is a survival strategy. This can happen only if all stakeholders at the local level — governments, businesses, universities and civil societies — work together to design and develop open, connected, sovereign digital infrastructure that bridges parallel global ecosystems. Connecting the dots and disparate systems can only be achieved locally and eventually globally. History has shown that it is not just giants that thrive, but small, nimble and adaptive species that survive.

Tan Sri Andrew Sheng writes on global issues from an Asian perspective

Save by subscribing to us for your print and/or digital copy.

P/S: The Edge is also available on Apple's App Store and Android's Google Play.