Tech & AI
Every major AI story has an Asia-Pacific chapter, whether Silicon Valley admits it or not. The platforms, chips, and startups building what comes next in technology, first.
-
Aug- 2026 -18 August

AI cloned the prime minister’s voice to steal $7.4 million from Australians
ASIC removed 19,400 fraudulent sites in a single year, yet deepfake videos of Anthony Albanese promising $40,000 monthly returns still convinced investors to hand over their savings.
-
17 August

Singapore’s first wetware rack runs on neurons that expire in six months
A 20-unit biological computer using 16 million lab-grown human neurons draws just 850–1,000 watts, but replacement demands sterile labs and trained cell biologists every half-year.
-
17 August

A Wellington startup built an AI accountant that actually works
Sterling raised NZ$3.8 million from Blackbird to automate month-end close across existing finance systems, not replace them—a harder sell than any chatbot demo.
-
16 August

Australia’s cable shield faces its first test as two undersea lines fail
Shunt faults in the Perth protection zone on August 8–9 arrived days before an ANU paper urged Canberra to market itself as the Indo-Pacific's safest cable hub, turning abstract vulnerability…
-
14 August

Apple surrendered its AI to China. Beijing approved it anyway.
The Cyberspace Administration of China registered Apple Intelligence in July 2026, but the service will run on Alibaba's Qwen model, not Apple's own technology—a template for how Beijing forces foreign…
-
14 August

A Chinese chipmaker nobody knows just became worth more than Intel
CXMT's $540.5 billion valuation rests on a 6.73% free float, making the ranking misleading—but the $8.6 billion capital behind it is real.
-
14 August

Australia’s sovereign AI just bought itself two years of speed
Centuria locked A$165 million in GPU debt, 72MW of off-grid power, and federal colocation access this week, compressing its deployment timeline from 2030 to early 2028—if customer contracts follow before…
-
13 August

Taiwan’s nuclear agency fell to free AI agents in four days
Open-source frameworks breached 21 government systems by falsely declaring themselves authorized penetration testers, exposing a design flaw that cannot be patched away.
-
12 August

China just unwound Meta’s $2 billion AI deal after it closed
The NDRC ordered Manus to reverse its acquisition on August 11, establishing Beijing's power to retroactively unwind cross-border tech deals regardless of offshore incorporation—a capability foreign acquirers assumed did not…
-
10 August

India is redefining what makes a platform a publisher
The government argues Meta's algorithmic recommendations constitute active publishing, a legal theory that could strip safe-harbour immunity from every ranked service operating in the country.








