Asia Is Building Its Own AI Stack

ATP is presented by:

Welcome back to The Asia Tech Podcast. Today we discussed Asia’s move to build its own AI stack, the evolving cybersecurity landscape for SMEs, AI-powered credit intelligence, AI-driven marketing automation, the human side of AI adoption in enterprises, programmable money for banks, and blockchain as a coordination layer for distributed AI.

• 28:00 Steve Hunter - Director of Engineering at Arctic Wolf

• 42:23 Jamie Twiss - Chief Executive Officer at Carrington Labs

• 1:03:09 Chrissy Lim - Founder & CEO at Protaigé

• 1:24:10 Luke Salway - Master Facilitator, MCC Certified Coach, CEO at Coachology

• 1:44:04 Julia Demidova - Head of Digital Assets Product & Strategy at FIS

• 2:05:49 Levi NGO - COO at Hola Tech

Here is an overview of the topics we discussed today:

Something has quietly shifted over the last few months. Asia is no longer just consuming AI — it is assembling every layer of the stack to produce it domestically. Models, low-cost inference, domestic chips, developer distribution, and policy control over access are all being pieced together. China is the center of gravity right now, but India is not far behind.

The clearest signal: Beijing is reportedly weighing restrictions on overseas access to its most advanced AI models, including open source releases. Companies like Alibaba and Zhipu AI are part of those conversations. Zhipu’s GLM 5.2 already ranks fifth on a major AI intelligence leaderboard and second on Coderina’s front-end coding benchmark — at materially lower cost than leading US models. Those benchmarks were not created in China. AI is being treated less like software and more like strategic infrastructure, in the same policy category as semiconductors and cryptography. Europe, by contrast, has nothing comparable at scale.

Steve talked about how the attack landscape has fundamentally changed — and it is more urgent than most organizations realize. The old model of a secure perimeter has given way to what he called “a thousand holes.” Agentic AI is not just being used inside organizations to boost productivity. Attackers are using it too.

A recent example: a threat actor used an agentic AI system to exploit a vulnerability in LangFlow, a popular AI orchestration framework. The attack, dubbed Jade Puffer, is a modern version of “Patch Tuesday, Exploit Wednesday” from 20 years ago — but compressed into a much tighter window. The gap between a vulnerability being discovered and weaponized is shrinking fast. And the organizations most at risk are SMEs, who face the exact same threat landscape as the largest banks but without the security teams to match. Steve pointed to Singapore’s CyberMark framework as a practical starting point — government guidance that SMEs can actually use.

Jamie’s take on AI and lending was one of the more striking conversations. The core idea: you can learn more about someone’s creditworthiness from how they manage their money than from any credit score.

One example he gave was surprisingly specific. If someone barely made rent one month, the interesting question is not that they struggled — it is how far in advance they saw it coming. Did they start compressing discretionary spending 21 days before the payment? Or 3 days? That one number tells you something a credit score never could. Carrington Labs builds thousands of these signals from raw transaction data. The goal is not a black box decision — it is an explainable one. And the lending workflows themselves? Jamie described them as something a Charles Dickens character would recognize. The AI opportunity here is enormous and almost entirely untouched.

Luke works with enterprises trying to bring AI into their organizations, and he sees the same pattern repeatedly. The technology gets introduced before the mindset does.

Leaders with a fixed mindset — “this is how we’ve always done it” — are the biggest obstacle. The issue is pace: organizations keep trying to roll out AI all at once, without testing or iteration, driven by board-level pressure that trickles down from a conversation at a golf club. The smarter path, Luke argues, is to shorten the distance to acceptance. Communicate the change clearly, address the very real fear of job displacement directly, and carve out protected time for teams to experiment without consequences. Failure should be expected, not avoided.

Julia’s framing was refreshingly direct. Banks thinking about stablecoins only as payment products are missing most of the value. The real opportunity is in FX, liquidity management, treasury services, custody, and cross-border settlement — particularly in Asia, where trade flows are enormous.

Dollar-denominated and local-currency stablecoins will both play a role, driven by use case rather than competition. What banks are increasingly asking for is a single infrastructure layer that handles stablecoins, tokenized deposits, and tokenized assets all in one place. The most in-demand tokenized assets in Asia right now: gold, money market funds, and bonds. And yes — the move toward integrating digital assets directly into standard bank accounts is already underway in several markets.

Watch the full episode here:

Watch the episodoe

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

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