Skip to main content
AI in Asia
3 Before 9: March 19, 2026
3 Before 9

Thursday, 19 March 2026

3 Before 9: March 19, 2026

3 daily AI stories and 1 bold opinion before your 9am kopi

Listen to today's 3 Before 9

Who should pay attention

Enterprise AI buyers in Asia-Pacific | Semiconductor and memory investors | Cloud infrastructure strategists

What changes next

If Nvidia's China H200 restart holds and volume caps are raised, Chinese cloud providers will accelerate training workloads that have been on hold since mid-2025, intensifying competitive pressure on US AI labs.

1

Nvidia Restarts H200 Chip Shipments to China After Ten-Month Freeze

At Nvidia's GTC 2026 conference, CEO Jensen Huang confirmed the company is restarting manufacturing of its H200 processors for shipment to China. The announcement ends a freeze of roughly ten months on advanced chip supplies to the world's second-largest economy, a period during which Chinese AI developers were left scrambling for domestic alternatives or older, less capable silicon. Huang confirmed that Nvidia has received purchase orders from multiple Chinese customers and that the supply chain is now actively ramping. The H200 shipments will be subject to a 25 per cent US duty and a government inspection regime. Officials are reportedly considering a cap of 75,000 chips per customer and up to one million processors in total. Nvidia's top-tier Blackwell and Rubin architectures remain off-limits for Chinese buyers under current export control rules.

Why it matters for Asia

China was once responsible for roughly a quarter of Nvidia's total revenue. Its return as a buyer reshapes the competitive calculus for the entire Asia-Pacific semiconductor supply chain. Korean and Taiwanese chip packaging and memory suppliers stand to benefit from renewed H200 production volume. Chinese cloud providers, many of which have been rationing compute capacity since mid-2025, will finally gain access to training-grade silicon. But for the Asian semiconductor ecosystem, even a partial reopening of the China market is significant news.^

Read the full story
2

Micron's Memory Supercycle Delivers Record-Breaking Q2 Results

Micron Technology's fiscal Q2 2026 results were extraordinary by any measure. Revenue of US$23.86 billion came in roughly three times the year-ago figure and beat analyst expectations by nearly US$4 billion. Earnings per share of US$12.20 demolished the US$9.31 consensus. Cloud memory revenue alone surged more than 160 per cent to US$7.75 billion, driven by insatiable demand for high-bandwidth memory used inside AI accelerators. Micron's entire 2026 HBM production capacity is already sold out, and CEO Sanjay Mehrotra has pointed to structural supply constraints that show no sign of easing before 2028 at the earliest. Micron guided Q3 revenue to roughly US$33.5 billion, implying year-on-year growth above 200 per cent.

Why it matters for Asia

South Korea's SK Hynix and Samsung are locked in a three-way race with Micron for HBM market share, and both are aggressively expanding fab capacity across Korea, Japan and China. With HBM shortages expected to persist until 2028 or 2029, Asian memory manufacturers and their equipment suppliers are looking at a prolonged earnings tailwind that will reshape balance sheets across the sector. The memory shortage is compounding the AI chip access problem for smaller players in Southeast Asia and India who lack the procurement scale of hyperscalers.^

Read the full story
3

Nebius Opens Asia-Pacific Headquarters in Singapore

Nebius, the Nasdaq-listed AI cloud provider, has appointed John Haarer as general manager for Asia-Pacific and Japan and established its regional headquarters in Singapore. Haarer brings experience from Cloudflare and Twilio, both companies with strong Asia expansion track records, and will lead commercial growth across Singapore, Japan, South Korea and India. Nebius's Asia-Pacific expansion comes off the back of 479 per cent year-on-year revenue growth in 2025, a contract backlog exceeding US$20 billion, and multi-year infrastructure deals with both Microsoft and Meta. The company is targeting annualised run-rate revenue of US$7 billion to US$9 billion by end of 2026.

Why it matters for Asia

Singapore's continued rise as the preferred regional hub for AI infrastructure is underscored by this move. The city-state's regulatory clarity, data connectivity, and access to regional talent make it the default choice for cloud and compute providers establishing an Asia-Pacific presence. For enterprise buyers across the region, Nebius's arrival introduces a well-funded, GPU-native competitor into a market currently dominated by AWS, Azure and Google Cloud. More competition means more pricing pressure, and that is unambiguously good for buyers.^

Read the full story

That's today's 3 Before 9.

Explore more at AI in Asia or share signals with us.

Recent Editions

View all

Sunday

26 April 2026

  • 1. Sony AI's Project Ace becomes the first robot to beat elite and pro table tennis players, published on the cover of Nature.
  • 2. Geely, AFARI Technology and CaoCao Mobility debut China's first purpose-built robotaxi - the EVA Cab - at Auto China 2026, with commercial launch slated for 2027.
  • 3. Taiwan and South Korea leapfrog the UK in global equity rankings as the AI chip rally lifts TSMC, Samsung and SK Hynix to structurally larger weight.
Read edition

Sunday

26 April 2026

  • 1. Sarvam AI is closing a $300-350M Series B at a $1.5B valuation led by Bessemer with Nvidia, Amazon and HCLTech, India's first major sovereign-AI unicorn round.
  • 2. Tencent launched Hy3 Preview, a 295B-parameter MoE model, and swapped DeepSeek out of its Yuanbao chatbot in favour of the new in-house technology.
  • 3. Nissan committed to fitting AI-driven Level 2-plus hands-free driving from UK startup Wayve into 90 per cent of its future vehicle line-up.
Read edition

Friday

24 April 2026

  • 1. DeepSeek launches V4-Pro and V4-Flash with a 1-million-token context window, running on Huawei's new Ascend 950 supernode clusters and priced at roughly a tenth of OpenAI's output-token rate.
  • 2. SoftBank is seeking a $10 billion two-year margin loan collateralised by its OpenAI stake, pushing its total commitment to the ChatGPT maker to about $64.6 billion.
  • 3. TSMC unveiled A13, a 1.3nm-class shrink of A14 aimed at AI accelerators, with production slated for 2029 alongside a 1.2nm A12 variant, while A16 slips to 2027.
Read edition

Thursday

23 April 2026

  • 1. Microsoft pledges A$25 billion (US$17.9 billion) to make Australia its largest AI infrastructure hub, including skills training for three million workers.
  • 2. SK Hynix posts a record 72 per cent operating margin and US$27 billion quarterly profit as HBM demand keeps Asia's memory duopoly in the driver's seat.
  • 3. Japan's Nikkei 225 breaches 60,000 for the first time, but only 17 per cent of Tokyo stocks rose on the day, exposing a dangerously narrow AI-driven rally.
Read edition

Wednesday

22 April 2026

  • 1. LINE Yahoo launched Agent i on 20 April, a single AI agent now embedded in both the LINE app's 100m-plus user base and Yahoo! JAPAN, with enterprise and no-code versions due over the summer.
  • 2. SoftBank named Arm CEO Rene Haas to also run SoftBank Group International from 21 April, tightening coordination across the group's overseas chip and AI bets including Arm, OpenAI, Stargate and Project Crystal Land.
  • 3. Meta's $2 billion acquisition of Singapore-based Chinese AI start-up Manus is still under Beijing's national-security review, with co-founders barred from leaving China and the 'Singapore washing' route for China AI firms now under serious pressure.
Read edition

Tuesday

21 April 2026

  • 1. Qualcomm CEO Cristiano Amon met Samsung and SK Hynix executives in Seoul to pitch 2nm wafer orders and lock in memory supply for Snapdragon and AI platforms.
  • 2. Chinese humanoid robot makers shipped roughly 90 per cent of global units in 2025 but still trade at a valuation discount to pre-production US rivals.
  • 3. Volkswagen confirmed its China cars will ship with an on-device AI voice agent from H2 2026, built on tech from Tencent, Alibaba and Baidu.
Read edition