KUALA LUMPUR, Dec 5 (Bernama) -- Bursa Malaysia closed lower on Friday amid mixed regional market performance as investors turned cautious over a possible rate hike by the Bank of Japan (BOJ) and upcoming US economic data that may influence the Federal Reserve’s (Fed) interest rate decision next week. At 5 pm, the FTSE Bursa Malaysia KLCI (FBM KLCI) pared most earlier losses to settle 4.55 points easier, or 0.28 per cent, to 1,616.52 from Thursday’s close of 1,621.07. The benchmark index, which opened 0.37 of-a-point lower at 1,620.70, moved between 1,609.67 and 1,621.25 throughout the day. The broader market was negative, with decliners outpacing advancers 604 to 439. A total of 550 counters were unchanged, 1,151 untraded, and 18 suspended. Turnover declined to 3.17 billion units worth RM2.24 billion from 4.48 billion units worth RM2.75 billion yesterday. Rakuten Trade Sdn Bhd vice-presiden...
Amazon has officially stepped deeper into the AI arms race, launching Trainium 3, a new chip built to challenge Nvidia’s dominance in the GPU market.
The latest hardware — available through Amazon Web Services — promises 4x faster model training than its predecessor while using the same energy footprint. Each cluster of Amazon’s new “UltraServers” can run up to 144 Trainium 3 chips, enabling large-scale LLM training at hyperscaler level.
The move positions Amazon directly against Google and Nvidia, both of which dominate today’s AI infrastructure landscape. Google’s lead in the AI model race has reportedly pushed OpenAI CEO Sam Altman into a “code red” mode.
Crypto Miners Become AI’s New Powerhouses
The AI boom has created a problem even Big Tech struggles with: massive energy and space requirements.
Enter crypto miners — companies already running large-scale, power-heavy data centers. After the 2024 Bitcoin halvingcut mining rewards, major firms began repurposing their operations into AI-ready facilities.
Now, firms traditionally viewed as bitcoin plays are evolving into AI utilities:
Core Scientific (CORZ)
CleanSpark (CLSK)
Bitfarms (BITF)
TeraWulf (WULF)
IREN (IREN)
IREN surged last month after securing a massive US$9.7 billion AI cloud deal with Microsoft, while TeraWulf locked in a US$9.5 billion joint venture with Google-backed Fluidstack.
These miners control gigawatts of existing power capacity, complete with cooling systems and grid stability — making them ideal partners for hyperscalers hungry for compute.
But the AI-Minor Convergence Comes With Bubble Risks
Despite the excitement, analysts warn that the AI infrastructure boom is flashing early bubble signals:
1. Heavy Borrowing
Miners are taking on substantial debt to retrofit their facilities for AI workloads — a risky move if demand cools.
2. Risk-Asset Correlation Breaks Down
Tech and crypto have fallen sharply:
Bitcoin: -17% in 30 days
CoinDesk 20 index: -19.3%
NASDAQ 100: -1.5% (after a 7% drawdown)
Markets are turning wary of the massive capex cycle behind the AI trade.
3. The Trillion-Dollar Question
OpenAI alone has committed to trillions in infrastructure spending — funding it still needs to secure.
Bain & Co. warns that:
The AI industry could face an US$800 billion capital shortfall
Companies may need US$2 trillion in annual revenue by 2030 to meet projected compute demand
If AI demand slows, both Big Tech and hybrid miner-AI operators could face a liquidity crunch reminiscent of crypto’s 2022 meltdown.
Such a shock would likely trigger heavy selling across risk assets.
The Takeaway: A New GPU Gold Rush — but With Fragile Foundations
For now, crypto miners are betting their future on GPUs, not ASICs, transforming into the backbone of the next wave of AI infrastructure.
But the same energy that fuels this AI boom also heightens the risk:
if demand falters even slightly, the fallout could ripple across tech, crypto, and broader markets.
The AI arms race is accelerating — but so are the warning signs.
Comments
Post a Comment