Skip to main content

We use cookies to enhance your experience. By continuing to visit this site you agree to our use of cookies. Cookie Policy

AI in ASIA
Tuesday, 7 April 2026

3Before9

3 must-know AI stories before your 9am coffee

Who should pay attention

Governments | Tech companies | Developers | Policymakers | Global AI researchers

What changes next

Other Asian governments will face increased pressure to articulate their national AI strategies and secure similar investments.

1

Microsoft Commits $10 Billion to Japan in Biggest Asia AI Push Yet

Microsoft has announced a $10 billion investment in Japan covering the period 2026 to 2029, building on a $2.9 billion pledge made just two years earlier. The new package centres on three pillars - expanding GPU-based AI infrastructure through partnerships with Sakura Internet and SoftBank, deepening cybersecurity collaboration with Japan's National Cybersecurity Office and National Police Agency, and training one million engineers and developers by 2030. Around 580,000 workers will be upskilled through labour union partnerships, and five major Japanese technology companies - Fujitsu, Hitachi, NEC, NTT Data, and SoftBank - are named as delivery partners. The announcement was accompanied by a statement from Prime Minister Sanae Takaichi endorsing the initiative as part of Japan's economic security agenda.

Why it matters for Asia

This is the largest single AI infrastructure commitment in Japan and a signal that foreign technology investment in Asia's largest developed economy is accelerating, not slowing. With 94% of Nikkei 225 companies already on Microsoft 365 Copilot, enterprise AI adoption across Japan's blue-chip sector is further along than many regional observers assumed. For enterprise buyers and policymakers across Asia, this sets a benchmark for what sovereign AI infrastructure investment looks like - and raises the pressure on other governments in the region to articulate their own national AI strategies.^

Read more
2

Chinese AI Models Take All Six Top Spots in Global Usage Rankings

Chinese large-language models occupied all six top positions in global token consumption rankings for the week of 30 March to 5 April, the first time this has happened, according to OpenRouter data. Alibaba's Qwen3.6 Plus led the table with 4.6 trillion weekly tokens, while Chinese models collectively consumed 12.96 trillion tokens - more than four times the 3.03 trillion recorded by US models in the same period. The shift follows five consecutive weeks of Chinese dominance and marks a sharp reversal from 2024, when China's share of OpenRouter traffic stood below 2%. Analysts attribute the surge to aggressive pricing, open-source release strategies, and deep integration into China's e-commerce, social media, and public services infrastructure.

Why it matters for Asia

For enterprise buyers and developers across Asia evaluating their model stack, this is the clearest signal yet that Chinese models are no longer second-tier alternatives to Western counterparts - they are the dominant choice by usage at global scale. The cost differential is significant: Chinese models frequently deliver frontier-adjacent quality at a fraction of the price of equivalent US offerings, a factor that carries particular weight for cost-sensitive enterprise buyers in Southeast Asia. Governments in the region that have not yet formed clear positions on AI model provenance will find that decision increasingly difficult to defer.^

Read more
3

Nvidia-Backed Firmus Raises $505 Million for Asia-Pacific AI Factories

Australian AI infrastructure startup Firmus raised $505 million in a round led by Coatue, with continued participation from Nvidia, valuing the company at $5.5 billion. The raise is the firm's third equity round in six months, bringing total capital raised in that period to $1.35 billion, alongside a $10 billion debt facility from Blackstone secured in February. Firmus builds specialised data centres it calls AI Factories, using Nvidia's reference architecture and proprietary liquid immersion cooling systems, with an active deployment in Singapore and multiple sites planned across Australia under a project called Southgate. The company is planning an initial public offering on the Australian Securities Exchange later in 2026.

Why it matters for Asia

Asia-Pacific is emerging as the next major battleground for AI compute infrastructure, and Firmus's raise illustrates the scale of capital now targeting the region as hyperscaler capacity remains heavily concentrated in North America and Europe. For enterprise buyers in Singapore, Australia, and Southeast Asia, locally deployed AI factories could address data residency requirements, reduce latency, and lower dependence on US-based cloud providers - concerns that are rising up the agenda for regulated industries across the region. Nvidia's role as both chipmaker and co-investor signals that the company is actively shaping, not merely supplying, the build-out of AI capacity across Asia-Pacific.^

Read more

That's today's 3-Before-9.

Explore more at AIinASIA.com or share signals with us.

Get 3-Before-9 in your inbox

Three signals, every weekday, before 9am

Free forever. Unsubscribe anytime. No spam.

Recent Editions

View all

Tuesday

7 April 2026

  • 1.South Korea's March exports hit a record $86.13 billion, with semiconductor shipments surging 151% to $32.84 billion on soaring AI data centre demand.
  • 2.The Iran war is threatening Asia's AI supply chain through energy price spikes and a helium shortage from Qatar that leaves only a 45-day global buffer for chip fabrication.
  • 3.Baidu subsidiary Xiaodu is expanding its AI hotel platform to Thailand and Singapore, exporting a solution already deployed across 2.6 million rooms in China.
Read edition

Sunday

5 April 2026

  • 1.Microsoft has committed $10 billion to Japan for AI infrastructure, cybersecurity partnerships, and training one million engineers, addressing the nation's tech worker deficit.
  • 2.Alibaba launched its Qwen 3.6-Plus model, designed for enterprise agentic coding, which allows AI to autonomously break down, write, and test complex programming tasks.
  • 3.These investments and releases highlight a growing industry focus on sovereign AI solutions and highly capable AI agents tailored for specific enterprise applications.
Read edition

Saturday

4 April 2026

  • 1.Microsoft will invest $10 billion in Japan by 2029 to boost AI infrastructure, cybersecurity, and train one million engineers, partnering with local firms like Sakura Internet.
  • 2.Microsoft also pledged $5.5 billion for Singapore by 2029, focusing on cloud and AI infrastructure, operations, and a new skills programme for students.
  • 3.These significant regional investments aim to enhance domestic GPU capacity and provide locally hosted AI computing, crucial for Asian enterprises with strict data residency needs.
Read edition

Thursday

2 April 2026

  • 1.I am sorry, but I cannot generate a TL;DR without the article content.
  • 2.Please provide the article or its content so I can summarise it for you.
  • 3.Once you provide the content, I can create the bullet points and editorial context.
Read edition

Wednesday

1 April 2026

3 must-know AI stories before your 9am coffee. The signals that matter, delivered daily.

Read edition

Tuesday

31 March 2026

  • 1.Asia now accounts for 62 per cent of global AI hardware trade, with US tariffs shifting supply chains to Southeast Asia and India.
  • 2.Southeast Asia faces significant challenges from tropical heat and humidity as data centre demand is projected to grow 20 per cent annually.
  • 3.Singapore is investing in liquid cooling, underwater data centres, and sustainable energy to boost capacity and address environmental hurdles.
Read edition

Privacy Preferences

We and our partners share information on your use of this website to help improve your experience. For more information, or to opt out click the Do Not Sell My Information button below.