Singapore AI Week runs June 8-14, with SuperAI 2026 as its anchor event at Marina Bay Sands on June 10-11. If you follow the US-centric AI conference circuit — NeurIPS, GTC, Build, Google I/O — SuperAI is worth knowing about for a reason that goes beyond speaker names: it is explicitly positioned as the place where AI development happens outside the US-China axis.
That framing is doing real editorial work in 2026.
Quick Facts
- SuperAI dates: June 10-11, 2026
- Location: Marina Bay Sands, Singapore
- Singapore AI Week: June 8-14
- Scale: 10,000 attendees, 1,500 AI companies, 150+ speakers, 100+ exhibitors, 150+ countries represented
Why Singapore, Why Now
The conference’s own framing is “where the world’s AI powers converge on neutral ground.” That phrase isn’t just marketing — it’s a reading of the current moment.
As US export controls on AI chips tighten and China’s AI development proceeds on a parallel track, Singapore has positioned itself as a jurisdiction that can work with both. The city-state has invested heavily in AI infrastructure (including a national compute cluster), relaxed talent visa frameworks for AI researchers, and explicitly courted both US and Chinese AI labs. Singapore AI Week is the institutional expression of that bet.
For builders who work in Asia-Pacific markets, or who are building infrastructure that will need to operate under multiple regulatory regimes, the Singapore AI Week speakers and sessions reflect a different axis of concern than US-centric conferences. The tracks on governance, compute access, and fintech AI have a different flavor when the audience includes builders from Southeast Asia, Japan, India, and the Gulf.
Speaker Highlights for Builders
The most directly relevant speakers for the technical builder audience:
Andy Hock (CSO, Cerebras Systems) — Cerebras makes wafer-scale compute specifically for AI inference optimization. Hock’s session will likely cover the practical tradeoffs between GPU clusters and purpose-built AI silicon for inference workloads — a question that matters if you’re running inference at scale and want alternatives to NVIDIA.
Evan Conrad (CEO, San Francisco Compute) — San Francisco Compute is a newer entrant in the GPU access/compute marketplace. Conrad’s session is expected to address the access layer: how builders without enterprise GPU contracts can get compute at scale.
Zixuan Li (Head of Z.ai) — Z.ai has been one of the more aggressive frontier model developers outside the US-Europe-OpenAI axis. Li’s session covers frontier scaling, which is increasingly relevant as inference costs drop and the “which model do I use” question expands beyond GPT-4 and Claude.
Balaji Srinivasan (keynote) — Srinivasan’s keynote is framed around personal, private, and programmable AI — the intersection of personal AI assistants, privacy-preserving inference, and programmable AI agents that builders control rather than platforms. His talks typically function as an essay with a specific thesis; even if you disagree with his conclusions, they’re worth reading as a signal of where the decentralized/sovereign AI movement is heading.
Benedict Evans (annual AI outlook) — Evans does an annual “state of AI” deck that’s widely circulated in product and strategy circles. His SuperAI session will likely cover his current read on where AI development stands, where the hype is overextended, and what the next 12-18 months look like from a market structure perspective.
Ramine Tinati (Google DeepMind) — The Google DeepMind presence covers their research and applied AI agenda; expect framing around safety, capabilities, and infrastructure.
The Six Conference Tracks
SuperAI runs six tracks, each covering a different vertical:
- Robotics and Humanoids — the physical AI layer; humanoid robot deployment, industrial automation
- Frontier Scaling and Safety — model capabilities, alignment research, safety frameworks
- Compute Infrastructure — hardware, data centers, energy, supply chain for AI compute
- Fintech AI — AI in financial services, trading, fraud, compliance, payments
- Health and Biotech AI — AI in drug discovery, diagnostics, clinical workflows
- AI Governance and Policy — regulation, international frameworks, ethics
For most builders, the Compute Infrastructure and AI Governance tracks will be the most useful if you’re watching the recordings. The governance track is particularly interesting given Singapore’s regulatory environment and the multi-jurisdictional audience.
The NEXT Hackathon
SuperAI runs an attached hackathon — NEXT — with $130,000+ in prizes backed by AWS and Vercel. The format is a 36-hour build sprint with embedded technical mentors; the top 5 teams present on the main SuperAI stage to the full 10,000-person audience. Applications for 2026 are closed (a waitlist is open for 2027), but if you’re in Singapore this week and were selected, this is one of the more visible hackathon stages available outside the US.
What This Means for Builders Who Aren’t Going
If you’re not in Singapore this week, the practical value of following SuperAI is:
Compute alternatives are consolidating. Cerebras, San Francisco Compute, and the Singapore national compute cluster represent different paths to AI compute that aren’t NVIDIA/AWS. If GPU access is a constraint for you, the SuperAI compute track is worth reading for emerging options.
Asia-Pacific AI development is not a derivative. The presence of Z.ai, Singapore’s sovereign AI infrastructure, and the multi-jurisdictional governance track reflects genuine AI development momentum in markets that US-centric coverage often underweights. If your users are in APAC or your data must stay in-region, this track is worth attention.
Regulatory divergence is accelerating. The EU AI Act, China’s AI governance framework, and the patchwork of US state laws are all heading in different directions. The Singapore governance track typically produces the clearest cross-jurisdictional analysis available in conference format.
What We’re Watching
The headline news from SuperAI 2026 will likely be a mix of partnership announcements, compute capacity disclosures, and one or two model or product reveals from the non-US AI labs in attendance. We’ll cover anything builder-relevant that comes out of the keynotes — particularly if Hock (Cerebras) or Conrad (SF Compute) announce pricing or availability changes that affect the GPU access market.
Singapore AI Week continues through June 14. SuperAI keynotes on June 10-11.
This preview is based on pre-event research. Speaker topics and session details may change. Event information accurate as of June 8, 2026.