Global AI Landscape Restructured: OpenAI's First Overseas Lab in Singapore, Meta Cuts 10% and Reassigns to AI
Late May 2026. Three landmark events reshaped the global AI industry.
Three events on three continents. All pointing to the same trend: AI is moving from “technology competition” to “resource reallocation.”
Capital, talent, policy—all are being repriced.
OpenAI’s First Overseas Lab Lands in Singapore
On May 20, OpenAI signed a cooperation agreement with Singapore’s Ministry of Digital Development and Information.
The core: establishing the OpenAI Singapore Applied AI Laboratory. This is OpenAI’s first AI lab outside the United States dedicated to application scenarios.
Planned investment: over SGD 300 million.
Why Singapore?
Three conditions converged:
- Policy friendliness: Singapore maintains an explicitly welcoming stance toward AI industry
- Geography: Southeast Asia’s digital hub, covering a 600-million-person market
- Talent density: NUS, NTU and other top institutions, bilingual English-Chinese environment
For OpenAI, this is the first step in “decoupling from the US, entering Asia.” For Singapore, it is a critical move to solidify its position as “Asia’s AI capital.”
Meta Cuts 10%, Reassigns 7,000 to AI
On the same day, Meta executed a new round of layoffs.
The numbers: 8,000 employees cut, about 10% of total headcount. 7,000 reassigned to AI workflow-related positions.
This is not simple downsizing. It is Meta’s largest internal resource reallocation since 2022.
Zuckerberg’s signal is clear: shrink non-AI business, expand AI business. The 8,000 cuts mainly hit traditional social media operations, hardware peripherals, and non-core product lines. The 7,000 reassigned will enter core battlegrounds: AI model training, agent development, content recommendation algorithms.
A detail worth noting: reassigned headcount is nearly 90% of laid-off headcount. Meta is not shrinking. It is restructuring.
Chinese Tech Giants Follow in Sync
The same week, China’s tech giants released a wave of AI product updates:
Tencent Launches AI Assistant Marvis
- Control your computer from your phone anytime
- Supports celebrity info search, gaming companion, intelligence monitoring, contract review
- Cloud mode and local mode supported
- Covers Windows, macOS, and Android
Marvis’s core selling point is “cross-device agent”—not answering questions inside an app, but executing tasks at the operating system level.
ByteDance Volcano Engine Launches One-Stop AIGC Short Drama Platform “Volcano Drama Creation 1.0”
- Adapted to Seedance, Seedream and other models
- End-to-end coverage from long-form script analysis, full-series asset design, storyboard video generation to final preview
- Production cycle shortened by over 80%
AIGC short drama is the biggest content industry风口 of 2025-2026 in China. ByteDance is entering with a platform approach, aiming to become the “infrastructure supplier” of the short drama industry.
Alibaba Releases Next-Gen AI Chip Pingtouge Zhenwu M890
- Deployed on super-node servers
- Simultaneously released flagship model Qwen3.7-Max
- Launched product website “Qianwen Cloud” adapted for agents
Alibaba’s strategy is three-layer integration: self-developed chips + self-developed models + agent platform. Chips are the cost moat, models are the capability moat, platform is the ecosystem moat.
The New Structure of Global AI Competition
| Dimension | United States | China | Southeast Asia / Others |
|---|---|---|---|
| Model Layer | OpenAI, Google, Anthropic lead | Baidu Wenxin 5.1, Alibaba Qwen3.7, DeepSeek catching up | Import-dependent |
| Chip Layer | NVIDIA monopoly | Pingtouge M890 starting | None |
| Application Layer | Enterprise SaaS dominant | Consumer agent, AIGC content explosion | Singapore as regional hub |
| Talent Layer | Meta reallocation, high-salary competition | Zhejiang University added major, engineer dividend | Singapore absorbs global talent |
| Regulatory Layer | Fragmented, federal contestation | Three-ministry regulation clarified | Singapore policy-friendly |
A Judgment
By mid-2026, global AI competition is no longer a single-dimension contest of “who has the better model.”
It is becoming a five-dimension war: models, chips, applications, talent, regulation.
US advantages: models and chips. But fragmented regulation is slowing deployment.
China advantages: application innovation and cost control. But high-end chips remain the bottleneck.
Hub markets like Singapore are leveraging policy flexibility and geographic advantage to become super-nodes.
For enterprises and developers, the critical judgment is: is your AI strategy tied to a single ecosystem, or cross-ecosystem?
Signals from the Talent Market
BOSS Zhipin Q1 2026 financials revealed a key data point: AI-related positions on the platform drove revenue growth exceeding 100%.
Simultaneously, software engineer hiring grew 10%.
“So far there has been no large-scale reduction in programmer positions that would cause panic,” said BOSS Zhipin founder Zhao Peng on the earnings call.
AI is replacing some jobs while creating more new ones. The question is: are your skills on the demand list for the new positions?
Data sources: OpenAI official announcement, Singapore Ministry of Digital Development and Information, Meta internal notice, Tencent official release, ByteDance Volcano Engine, Alibaba Cloud Summit, BOSS Zhipin financials, Caixin, May-June 2026.