Microsoft, Lightstorm Build India–Southeast Asia AI Cable, Z.AI Catches Up to Anthropic & OpenAI

ATP is presented by:

Welcome back to The Asia Tech Podcast. Today we discussed the new India–Southeast Asia AI undersea cable, how China’s Z.AI is catching up to US frontier models, treasury modernization with AI and digital assets, agentic systems for real estate and manufacturing SMEs, the looming post-quantum cryptography threat, and how no-code databases are turning everyday business users into builders.

We were joined by:

• 30:49 Manoj Dugar - Head of Global Payments Solutions, Asia (ex. Greater China) at HSBC

• 46:42 Bernie Devine - Senior Regional Director, APAC at Yardi Systems

• 1:22:00 Jeremy Mah - Founder & CEO at NeuXP Group

• 1:43:48 Lars Jankowfsky - Founder at Gradion

• 2:05:04 Christopher Smith - Founder and CEO at Quantus

• 2:24:05 Jeff Kuo - Founder & CEO at Ragic

Here is an overview of the topics we discussed today:

A Cable Built for AI, Not Just the Internet

Microsoft, Lightstorm, Singtel, and Tata Communications are building a 3,600km undersea cable called I2SEA connecting India, Malaysia, and Singapore. Unlike older cables, this one is described by Lightstorm as AI connectivity — designed for low-latency, high-capacity movement of GPU compute and cloud services. Targeted for Q4 2029, it represents something rare: the physical buildout of AI infrastructure across Asia. What does it mean when hyperscalers don’t just buy capacity after cables are built — they fund the cables themselves?

India’s East Coast is becoming strategically important. The cable lands in Machilipatnam and routes into Hyderabad, one of India’s emerging AI hubs. Microsoft has committed $17.5 billion to India between 2026 and 2029. Google another $15 billion. Malaysia is no longer just a transit point — land and energy constraints in Singapore are pushing data center construction northward, making this corridor increasingly critical for anyone building AI services across the region.

China’s Z.AI: A Mini DeepSeek Moment

Beijing-based Z.AI launched GLM 5.2, an open-weight model that independent benchmarks say is now leading the open-weight category — ahead of DeepSeek V4 Pro and Kimi 2.6. Reuters is calling it a “mini DeepSeek moment.” It runs at roughly one-sixth the cost of leading US closed models, with a one-million token context window built for agentic workflows. Some Silicon Valley developer teams are already testing it seriously.

The real question isn’t capability — it’s trust. Can enterprises in regulated industries run critical workflows through a Chinese open-weight model? And if open-source wins, where does value actually land in the AI chain? The cloud platforms running these models — AWS, Azure, Google — may end up being the real beneficiaries, not the model builders themselves.

Treasury Modernization: Promise vs. Reality

Manoj Dugar from HSBC sees the gap clearly. Treasury teams talk about AI at every roundtable, but actual deployment is slow. The areas getting real traction are tokenized deposits enabling 24/7 settlement, programmable payments, and stablecoin integration for digital commerce. The infrastructure is shifting faster than the processes — and treasury leaders who move now are the ones reducing cross-border friction before their competitors do. What would it take to close the gap between AI ambition and AI execution in corporate finance?

Agentic AI on the Ground — Weeks to Days

Lars Jankowfsky from Gradion gave one of the most concrete examples on the show. A manufacturing company was spending weeks getting component quotes — 30 people manually reaching out to suppliers across Asia. Agentic AI cut that timeline from weeks to days. The bigger lesson: you don’t need six-figure AI budgets to get real results. Focused investments in the $20–40K range, targeting where a business is bleeding time or money, deliver positive ROI faster than large transformation programs. The pressure driving this? Mostly economics — tariffs, pricing pressure, rising costs — not enthusiasm about AI.

Your Encryption Has an Expiration Date

Christopher Smith from Quantus raised something most businesses are not thinking about. Nearly all blockchains and digital financial systems rely on elliptic curve cryptography. That cryptography has a known quantum vulnerability — Shor’s algorithm will eventually break it. When Google published a more efficient version of the algorithm, the open-source community matched and exceeded it in three days using AI. If that’s happening openly, what is happening in classified labs? If you’re building payment infrastructure or holding digital assets today, is there a plan for the post-quantum transition — or is it being ignored until it becomes a crisis?

Real Estate, FinTech, and No-Code: Three More Builders

Bernie Devine showed how Yardi Systems is weaving AI into the core of real estate management — not as a feature, but as a platform-wide foundation. Their example: AI reviewing 200 invoices, flagging 30 for human attention, and approving the rest. Jeremy Mah at NeuXP sees agentic AI reshaping how businesses scale across borders, but warns that most companies still treat it as a software tool rather than a digital workforce layer. And Jeff Kuo at Ragic is closing the gap between technical and non-technical users — a no-code database platform where everyday business users can build real applications using nothing more than spreadsheet logic, with AI agents now able to do the building for them.

Watch the full episode here:

Watch the episode

ATP Is Presented by Twilio:

Twilio is the secure and scalable platform to build amazing customer engagement without the complexity, with the AI-driven workflows you need to connect.

Visit Twilio.com to start building today

ATP Is Made Possible by:

Boomi, the data activation company for AI, powers the agentic enterprise by bringing data to life across the business. Trusted by over 30,000 customers and supported by a network of 800+ partners, Boomi helps organizations of all sizes achieve agility, efficiency, and innovation at scale.

Visit www.boomi.com to learn more