The noise fades, but the pattern remembers.
I’ve seen this movie before. In 2017, it was Telegram groups buzzing about ICOs promising to “decentralize the world.” In 2020, it was DeFi protocols with triple-digit APYs that vanished overnight. Now, the script is being rewritten for Malaysia’s data centre boom—a narrative spun by governments, global tech firms, and venture capitalists who want you to believe that Southeast Asia’s next AI hub is being built in Johor. But the pattern remembers. And the pattern says: when the infrastructure narrative takes centre stage, the real game is elsewhere.
Context: Why Malaysia, Why Now?
The story is simple on the surface. Singapore’s data centre moratorium, imposed in 2019 due to energy and land constraints, forced hyperscalers to look for alternatives. Malaysia’s southern state of Johor, just across the causeway, offered cheap land, lower electricity costs, and a government hungry for foreign investment. By 2023, Google, Microsoft, Amazon, and ByteDance had announced billions in data centre investments. The narrative was set: Malaysia is emerging as a key AI hub, riding the wave of global AI compute demand.
But the pattern remembers.
Back in 2017, every second-tier city in China was an “AI hub.” In 2021, every crypto-friendly jurisdiction was a “Web3 hub.” The term “hub” is a marketing label, not a technical reality. What Malaysia is actually building is a physical infrastructure node—a massive cluster of servers running on subsidized electricity and cheap labour. The AI part is a convenient branding exercise. The real driver is the same as it was for Bitcoin mining farms in 2018: energy arbitrage.
Core: The Numbers Behind the Noise
Let’s look at the data. According to industry reports, Malaysia’s data centre capacity is projected to reach 2-5 GW of IT load by 2027, up from roughly 200 MW today. That’s a 10x to 25x increase. But the devil is in the details.
- Planned vs. Delivered: Most of these announcements are “Memorandums of Understanding” (MoUs), not shovels in the ground. I’ve tracked similar MoUs from the 2020 “DeFi Summer” era—projects that promised “liquidity mining” but never launched. The pattern is the same: hype precedes execution, and the gap between announced and operational is where value gets destroyed.
- Energy Reality: Malaysia’s state-owned utility, Tenaga Nasional, has flagged that the national grid may not support the additional load without massive upgrades. Water scarcity in Johor is another risk—data centres consume enormous amounts for cooling. The “cheap electricity” argument only holds if the supply can scale. Based on my experience in real-time trading signal analysis, I’ve learned that unverified supply assumptions are the fastest way to get rekt.
- GPU Allocation: How many of these data centres will actually house AI-specific hardware like NVIDIA H100s or B200s? The honest answer is very few. Most will be filled with traditional cloud servers, storage, and networking gear. The AI narrative is a marketing overlay to attract venture capital and government subsidies. We didn’t just watch the chart, we lived it—and the chart shows that the majority of “AI data centres” announced in 2023 have yet to order a single GPU.
Contrarian: The Unreported Angle
Everyone is focused on the AI boom. But the contrarian take is that Malaysia’s data centre boom is actually a crypto mining revival in disguise. Hear me out.
- The same factors that attract AI data centres—cheap power, lax regulations, and proximity to Singapore—are also attractive for Bitcoin mining operations. Several mining firms have already established sites in Malaysia, taking advantage of the same infrastructure. The line between “AI compute” and “proof-of-work hashpower” is blurring, and the narrative is being carefully curated to avoid the regulatory stigma of crypto.
- LayerZero’s cross-chain model relies on trust assumptions. Similarly, the “AI hub” narrative assumes that the data centres will be used for cutting-edge AI training. But the reality is that many of these facilities will be repurposed for high-frequency trading, crypto exchange hosting, or even tokenized compute networks. The same static streams of data, just with a different label.
- The VC Playbook: I’ve seen this pattern in DeFi and Layer2. VCs manufacture a narrative—like “liquidity fragmentation”—to justify new products. Here, the “AI hub” narrative is manufactured to justify massive infrastructure spending. The real beneficiaries are the landowners, construction firms, and energy providers, not the AI startups or local communities. Shiny objects distract, but dry powder preserves. The dry powder here is the billions in capital that will be locked into long-term infrastructure contracts before the demand materializes.
Takeaway: What to Watch Next
The signal is not the press release. The signal is the power purchase agreement. The first real test will be whether Malaysia can deliver reliable, cheap electricity at scale. If the grid fails, the boom becomes a bust. The second test is whether any of these data centres actually host meaningful AI workloads—or if they become glorified warehouses for crypto mining rigs.
Trust the code, verify the art, ignore the hype. The code here is the energy contracts, the GPU utilization rates, and the actual data on data centre capacity delivered. The art is the shiny narrative of “AI hub.” Ignore it.
From static streams to living liquidity. The liquidity in this market is not compute—it’s attention. And attention is already shifting to the next shiny object. Malaysia’s data centre boom will be a footnote in the bigger story of how AI infrastructure is being built on the backs of taxpayer subsidies and regulatory arbitrage. The pattern remembers. And the pattern says: when the infrastructure narrative peaks, it’s time to sell the shovel suppliers, not buy the hype.
