Tag: Screen Printing

Singapore’s war on scams has mostly been fought far from the victim: in police operations centres, bank back offices and, more recently, in rulebooks telling platforms what they must block. Now the city-state wants part of that fight to happen inside the phone itself.

Google on Wednesday announced two initiatives with the Singapore Government, unveiled on the sidelines of Singapore International Cyber Week (SICW) 2026. The first brings Singapore into Google’s Fairwind Programme, a gated scheme that gives vetted defenders the company’s most capable cyber models, including its new frontier model, Gemini 4 Argon.

The second pairs GovTech Singapore with Google DeepMind to build an on-device AI model for Android that flags scam attempts in real time without sending private conversations to the cloud.

Also Read: Singapore’s cybersecurity paradox: Leading in digital, lagging in defense

Both deals are framed as defensive. Both also deepen Singapore’s reliance on a single US tech giant at a time when access to frontier AI has become a geopolitical lever.

An unguarded model for trusted hands

Fairwind launched on September 2 as a limited access programme for governments and trusted partners, starting with Gemini 3.8 Flash Cyber paired with Google’s CodeMender agent. Argon, announced on September 30, has since become the programme’s headline act. Fairwind partners and internal Google teams get a version of Argon without its cyber guardrails, the filters that normally block requests that look like hacking.

Organisations may only grant Argon access to internal cybersecurity, incident response, or penetration testing teams, and must track employee access and use.

Google’s pitch is speed. Instead of spending weeks manually fixing vulnerabilities, it says defenders can generate verified, deployment-ready patches in minutes inside their own secure cloud environment. On Google’s own leaderboard, Grok 4.7, Gemini 4 Argon and GPT-6 Astra tie at 68 per cent on the CWE-bench v1 remediation benchmark, which also means the model still misses roughly a third of the test’s tasks.

“Minutes, not weeks” is a vendor claim, not a measured outcome. Analysts have pointed out that the clearest evidence would be merged-patch or time-to-remediation results from Fairwind partners, which Google has not published. An AI that finds bugs faster than humans can review fixes risks lengthening the backlog, not shortening it.

There is also the question of who sees what. Google says it monitors Argon’s chain of thought and its actions and stops execution when necessary. That is a sensible safety measure, though it implies a foreign company has some line of sight into how a government’s defenders probe their own systems. Google does say Argon supports zero data retention when used as a managed model on Gemini Enterprise, but the terms of Singapore’s participation were not disclosed.

A crowded, fragile club

Google is not the only lab rationing its sharpest cyber tools. Anthropic launched Project Glasswing on April 7, 2026, deploying its restricted Claude Mythos Preview model to a vetted consortium, and OpenAI responded with Daybreak on May 11. Each programme decides which governments and companies qualify, making access itself the product.

That access can disappear overnight. In June, US export controls forced Anthropic to suspend its Mythos and Fable models before Washington lifted the restrictions at the end of the month, an episode e27 has argued created a hydra rather than containing risk. For a small state building critical defences on American models, the lesson is clear: the off switch sits in another capital.

The scam problem is a human problem

The second initiative targets a more personal threat. Singapore’s scam-related losses fell 19 per cent to S$913.1 million (about US$715 million) in 2025, but the composition of those losses explains why the government is looking at devices. In 81.8 per cent of 2025 scam cases, victims transferred the money themselves rather than having their accounts taken over. Takedowns and bank freezes mostly come too late; the decisive moment is when a manipulated person presses “send”.

Also Read: Deepfake fraud losses hit US$3.7B as scams spread beyond social media

The GovTech-DeepMind model is designed for that moment. Lightweight Gemma models are fine-tuned on local scam narratives, including impersonation of public institutions, banks and local figures, and run entirely on the handset. When a conversation shows the high-pressure patterns of social engineering, a warning pops up on screen.

“Fighting scams shouldn’t mean sacrificing personal privacy,” said Mark Chen, Director of Trust and Safety at GovTech Singapore, adding that the approach keeps personal communications completely private.

The announcement leaves key questions unanswered. There is no launch date, no accuracy or false-positive data, and no explanation of how the model reads conversations inside end-to-end encrypted apps like WhatsApp or Telegram, which by design hide content from everyone but the sender and recipient. The model is also Android-only in a market where iPhones are hugely popular, leaving a large share of users uncovered.

The timing is not accidental. Singapore has tightened scam rules for messaging, social media and e-commerce platforms, with Google Messages and Google Meet among the designated services that must comply by January 31, 2027, under threat of fines of up to S$10 million (US$7.83 million). Meta has already disrupted 3.7 million scam-linked assets alongside police, and authorities have taken down 30,000 iMessage accounts. A government-endorsed on-device shield is as much Google’s compliance story as its public-safety one.

It also enters a crowded field. Singapore’s ScamShield app already filters calls and messages, and caller-ID players such as Taiwan’s Gogolook (Whoscall) and Truecaller have built large regional followings. Google’s edge is distribution: a model built into Android reaches people who would never download a separate app.

Grading its own homework

Google surrounded the announcement with supporting statistics. It says Google Play Protect has blocked more than 3.3 million installation attempts of potentially malicious apps across more than 770,000 devices, without specifying the period. It also cites studies showing that ten minutes of its Be Scam Ready game reduced players’ estimated scam risk by 20 per cent, and is launching a 2.0 version that halves playtime. These figures come from Google, and the effects of a quiz on behaviour months later are hard to measure. Its earlier work with the region, such as the US$5 million ASEAN Foundation drive, followed a similar mix of philanthropy and product.

Also Read: Should cybersecurity be nationalised?

If the on-device model works, the stakes reach well beyond Singapore. Malaysia ranked highest in the region for scam prevalence, with 32 per cent of adults losing money to scams, followed by the Philippines at 31 per cent. In those markets, as e27 has argued on Thailand’s scam epidemic, the problem is now technological and so must be the response, especially as AI phishing turns trust into the weakest link.

Singapore has once again cast itself as the region’s test bed. Whether this test produces a public shield or a deeper dependency will depend on details neither side has shared yet.

The post Singapore wants Google’s AI to catch scammers before victims hit send appeared first on e27.

In early 2018, Yujun Chean and Prateek Jogani were selling car insurance with a cashback twist. Their Singapore startup, Vouch Insurtech, had just raised SGD1 million (then US$755,000) in seed funding as the parent of Thai insurtech FairDee. Eight years on, the pair are back with a much bigger cheque and a far more crowded market.

Their new company, Parallel Loop, has launched with US$6.6 million in seed funding led by Bengaluru-based 3one4 Capital. Picus Capital, Flourish Ventures and Economic Development Innovations Singapore (EDIS) also joined, alongside three high-profile angels: Paytm founder and CEO Vijay Shekhar Sharma, CRED founder Kunal Shah, and Philip Yeo, former chairman of Singapore’s Economic Development Board and A*STAR.

Also Read: Everyone is talking about AI adoption. But are enterprises actually ready for it?

Founded only in June 2026, Parallel Loop calls itself an “AI operating partner” for mid-market and large enterprises. Four months in, it has 30 staff across Singapore, Bengaluru, Bangkok, London and Dubai, and more than 10 clients. These include Indian meat-delivery firm Licious, electricals maker Havells and Tokio Marine Safety Insurance (Thailand). Chean told e27 that Singapore enterprises, including a local healthcare group, are also on the list.

“Our aim is to build a global AI company,” he said.

The gap it is chasing

The pitch rests on a problem most enterprise leaders will recognise. AI tools have made individual employees faster, but turning that into a company-wide capability is far harder. The rules, definitions and know-how that make a business tick are scattered across systems, meetings, inboxes and chat apps. A model that cannot see that context is a clever intern who skipped onboarding.

“Intelligence has become abundant. What many organisations lack is the context, systems and capability to turn this into tangible business outcomes,” said Jogani.

The diagnosis fits the region. Southeast Asia, the argument increasingly goes, does not have an AI adoption problem so much as a scaling problem, and most enterprise AI in APAC is still stuck in the proof-of-concept room.

Three products, one bet

Parallel Loop’s offering has three parts:

• AI Enablement works with leadership teams to identify high-value use cases and build roadmaps.
• The AI Stack is a modular platform that connects a client’s existing systems with its organisational context. It powers everything from live dashboards to deterministic workflows, meaning automations that produce the same result every time rather than the varied output typical of generative models.
• The Software Factory pairs forward-deployed engineers with AI agents to modernise legacy systems and build new applications. Those forward-deployed engineers are the centre of gravity. Popularised by Palantir, the model embeds a vendor’s engineers inside client teams instead of handing over software and walking away.

The stack runs in the client’s own cloud or on premises, with data residency set by the client. It works with multiple model providers, including open-weight models, and keeps an audit trail of prompts, responses, context changes and actions. Clients retain ownership of their data and of the work built for them. That speaks directly to the regional appetite for sovereign AI, where, as one recent analysis argued, sovereignty starts long before the model itself.

Also Read: Why Southeast Asia cannot build sovereign AI on borrowed choices

Licious is the showcase client. Co-founder Abhay Hanjura said the company set out this year to become “fully AI-native”. He said teams across every function now use Parallel Loop daily as “a shared intelligence layer across the organisation”.

A familiar founding duo

The founders’ track record goes a long way towards explaining why investors moved so quickly. Jogani and Chean have worked together for nearly a decade. After Qoala acquired FairDee in 2021, Jogani became Group CTO and Chean Group COO of the Indonesian insurtech.

Qoala grew into five markets, processing millions of policies a month, and raised more than US$130 million from investors, including a US$65 million Series B in 2022. Jogani, who brings around two decades of technology experience, began running hands-on AI programmes for leadership teams in 2025. He noticed that many teams were adopting AI tools, but few were building the systems and processes needed to sustain the gains.

“They’re closing enterprise deals and expanding accounts across multiple geographies with an outcome-led approach,” said Anurag Ramdasan, Partner at 3one4 Capital.

A crowded field

Parallel Loop is walking into a brawl. Global consultancies such as Accenture and Deloitte sell AI transformation programmes to the same boardrooms. AI labs, including OpenAI, have built forward-deployed engineering teams of their own to win enterprise accounts, and Palantir still sets the template. Closer to home, Singtel’s RE:AI has partnered the SIT-NVIDIA centre to tackle the enterprise AI deployment gap.

The model also carries a familiar risk. Forward-deployed engineers are expensive, and the more a startup leans on them, the more it resembles a consultancy: software margins on the pitch deck, services margins on the books. Thirty people serving more than 10 clients across five cities in four months is fast. Whether the AI Stack becomes a reusable product, or each engagement turns bespoke, will decide whether Parallel Loop is valued as a software company or a services firm.

Also Read: Singapore firms embrace agentic AI, but audit trails remain thin

There is a quieter tension, too. A stack that promises no model lock-in also makes it easier for clients to walk away once the capability is built in-house.

For now, the founders have the network, the cheque and a client list that spans three countries. The harder test is whether enterprises that have been told to stop automating broken processes and calling it AI transformation will pay someone to fix the process first.

The post Parallel Loop nets US$6.6M to sell enterprises the one thing AI models lack: context appeared first on e27.

An AI chatbot that drafts a bad email is an embarrassment. An AI agent that sends it, to the wrong customer, with the wrong file attached, after approving the wrong refund, is a liability. Singapore’s regulators have clearly been losing sleep over the second scenario.

On Wednesday, the Infocomm Media Development Authority (IMDA) announced three initiatives at the AI Assurance Exchange: The Agentic Edition, an event organised by the AI Verify Foundation (AIVF) and SGTech that drew more than 300 deployers and testers. They are a starter kit for testing agentic AI systems, an accreditation programme for third-party AI testers with five firms in its first batch, and a new taskforce to defend the infocomm and media sector against AI-enabled cyber threats.

Also Read: What AI safety researchers actually worry about

Individually, each is a modest piece of guidance. Together, they signal something bigger: Singapore is moving from telling companies how AI governance should work to building an industry whose job is to prove that it does.

Why agents need a different exam

Traditional software follows rules. Generative AI produces content. Agentic AI does something more unnerving: it reasons, acts on a user’s behalf, touches sensitive data and changes the systems around it. Testing only the final output, as most companies do with chatbots, misses most of what can go wrong.

The Starter Kit for Testing Agentic Systems for Safety and Reliability tries to fill that gap with a four-step approach: identify risks such as erroneous or unauthorised actions, design realistic tests without causing real-world damage, run the tests and analyse the logs, then monitor and retest continuously.

The more interesting part is the new vocabulary. The kit asks testers to examine an agent’s entire “trajectory”, meaning both its reasoning and its actions, and to track agent-specific metrics. Tool use accuracy measures whether the agent called the right tool. Handoff accuracy checks whether it passed the task to the right agent or human. Human-in-the-loop rates show how often a person had to step in. It also tackles a problem most enterprises will recognise: how to test a system when you have only black-box visibility, because you bought the agent from a vendor rather than built it.

The kit draws on consultations with government agencies and more than 20 deployers and testers, including Google, AWS and Resaro, and includes case studies from Cohere, CyberSierra, Earlybird AI, Thailand’s KBTG and Knovel Engineering. It extends Singapore’s Model AI Governance Framework for Agentic AI, released earlier this year, and an earlier starter kit for testing large language model apps. IMDA calls it a living document, which is a polite way of admitting that nobody has fully worked this out yet.

The demand is real. In August, a Sumsub and Singapore Fintech Association study found that 94 per cent of Singapore businesses were using or piloting multi-step AI systems, yet only 29 per cent could produce an audit trail for AI-driven decisions. Ninety-eight per cent said they were ready to adopt third-party verification. The market for checking agents exists. The question was who would do the checking.

A badge for the testers

AIVF’s answer is the AI Tester Accreditation Programme (AI TAP), which it bills as a first-of-its-kind scheme for third-party AI testing providers. Applicants go through technical, operational and financial evaluations covering testing competency, track record, delivery capability and financial sustainability. That last criterion is telling. Regulators clearly don’t want an assurance firm folding halfway through an engagement.

Also Read: From copilots to colleagues: How agentic AI is redefining enterprise productivity

The first five accredited firms are AIDX Tech, Knovel Engineering, PwC Singapore, Resaro and Vulcan. Accreditation lasts one year, so firms have to keep earning it as the technology changes.

The stakes go well beyond enterprise IT. AIVF says the testers will assess applications in healthcare, finance, hiring and public services, the areas where an unchecked agent can quietly reject a job applicant, deny a claim or misroute a citizen’s request without anyone noticing.

Vulcan co-founder Alex Leung put the case bluntly: when AI can “send emails, move data or approve transactions, testing isn’t optional.” Resaro co-CEO April Chin argued that organisations need evidence from testing done outside the development team, not just from the people who built the system.

The questions the badge doesn’t answer

The programme deserves some scrutiny. The release does not say whether any sector regulator will require accredited testing. Without that, AI TAP risks becoming a marketing badge rather than a safeguard.

There is also a small-world problem. Resaro helped inform the starter kit, and Knovel Engineering appears in its case studies. Both are now accredited testers. In a market this young, that overlap is probably unavoidable and not improper. But financial auditing offers a clear lesson about what happens when the firms that write the standards, sell the advice and certify the results are the same few names. PwC’s presence in the first cohort makes the comparison hard to ignore.

Cost matters too. Large banks can afford independent AI audits. A Series A startup in Jakarta or Ho Chi Minh City running customer-facing agents may not, which leaves the free starter kit as its realistic option. Singapore has long pitched its governance toolkits as templates for ASEAN neighbours, and Malaysia is preparing an AI Governance Bill. Whether AI TAP travels across the region will depend on whether it stays affordable.

A taskforce on a 2027 deadline

The third initiative, the Frontier AI Resilience and Security Taskforce (FAIRS), brings together more than 20 leaders from telecommunications, media, cybersecurity and AI. Its focus is the risk that frontier models are getting better at finding vulnerabilities in networks that businesses and government systems depend on.

It has three priorities: faster response to AI-accelerated threats, segmenting critical networks to stop disruption from spreading, and pooling sector-wide intelligence on new attack techniques. Its first report is due in the first half of 2027. Attackers armed with the same models, already behind a wave of deepfake fraud, are unlikely to wait that long.

Also Read: Nearly half of APAC enterprises commit US$1M+ to agentic AI

Still, the direction is right. As e27 has argued, AI governance is moving from promises to proof. Singapore has now built the exam, chosen the first examiners and set a renewal date. Whether anyone is required to sit the test is the next policy fight.

The post Singapore’s next AI industry may be the people who check the agents appeared first on e27.

Anthropic has finally given its Singapore customers somewhere to knock. The San Francisco-based AI lab opened an office in the city-state on Wednesday, its fifth in the Asia Pacific, and signed a memorandum of understanding (MOU) with the Ministry of Digital Development and Information (MDDI) that puts AI safety testing at the centre of the relationship.

The MOU was signed on Tuesday, on the sidelines of Singapore International Cyber Week, by MDDI Permanent Secretary Chng Kai Fong and Anthropic co-founder Ben Mann, with Minister for Digital Development and Information Josephine Teo and Anthropic’s Head of APAC Policy Michael Sellitto looking on. It covers three areas: AI safety, security and model evaluation; research on how AI is adopted and what it does to the economy; and responsible adoption across government and industry.

Also Read: Singapore wants Google’s AI to catch scammers before victims hit send

The timing is hardly accidental. The same week, the Infocomm Media Development Authority (IMDA) unveiled new assurance measures for AI agents, a sign that Singapore’s regulators now worry more about software that acts than software that chats.

What is actually in the deal

Strip away the diplomatic language, and the MOU commits Anthropic to technical exchanges with Singapore’s AI safety, security and assurance institutions on model evaluation, testing standards and incident response, the playbook for what happens when a model misbehaves in the wild. Anthropic also says it will share research and data from its Anthropic Economic Index, help small and medium-sized enterprises (SMEs) adopt AI, build AI skills among public officers, and offer policy input as Singapore drafts its AI rules.

None of this is new territory. Anthropic joined the Generative AI Evaluation Sandbox run by IMDA and the AI Verify Foundation in 2023, and later worked with them on red-teaming (deliberately attacking a model to expose its weaknesses) in English, Mandarin, Malay and Tamil. The MOU formalises a relationship that already existed.

What the announcement leaves out is just as telling. There is no investment figure, no headcount target for the office beyond a note that it is hiring, and no word on whether the SME and skills programmes come with money attached. MOUs are statements of intent, not contracts, and Singapore signs plenty of them.

The customer list is the real story

Anthropic’s strongest case for an office here is the business it already has. More than 1,500 engineers and technical staff at the Government Technology Agency of Singapore (GovTech Singapore) use Claude Code, its AI coding tool, to modernise government systems. The Cyber Security Agency of Singapore is exploring how Claude could support the country’s cyber defence.

In finance, DBS uses Claude for software development and productivity in its Singapore operations, while sovereign wealth fund GIC is rolling it out across investment and corporate services teams. Regulated firms can now run Claude with in-region inference on Amazon Bedrock, meaning their data is processed within the region. That matters now that the Monetary Authority of Singapore has given financial firms a year to prepare for AI risk rules.

Also Read: Why Southeast Asia cannot build sovereign AI on borrowed choices

The consumer internet giants are in too. Grab says BriX, its Claude-powered sales assistant, saves its sales teams around 40,000 hours a quarter. Sea uses Claude to make Shopee’s customer service chatbot faster and more accurate, and thousands of its employees write and debug code with it. Supabase, the developer platform with Singaporean roots, is building with Claude Code.

A caveat: these figures come from Anthropic and its customers, not independent audits. So does the claim that Singapore ranks second of 121 countries for Claude.ai usage relative to population, at nearly six times what its size would predict. Usage per head also says little about value created, a gap visible in research showing messy data keeps many SME pilots stuck.

“Our customers here hold AI to an extremely high standard, especially when it comes to reliability and security,” said Dale Finlay, Anthropic’s General Manager for ASEAN.

Rivals are circling the same prize

Anthropic is not alone. OpenAI last month hired former Elastic and Snowflake executive Sanjay Deshmukh to lead Asia Pacific sales, and has backed a public-private AI accelerator in Thailand. Google is betting on schools, telcos and local languages to scale Gemini across the region. Databricks has unveiled a US$350 million investment plan for the city-state. And Singapore’s own SEA-LION model offers agencies a homegrown option for Southeast Asian languages.

What sets Anthropic’s pitch apart is that it leads with safety evaluation rather than consumer reach, the area where Singapore, through AI Verify, has made itself a global reference point. For a lab whose brand rests on pairing capability with safety, a government that tests models for a living is a natural partner.

The questions an MOU cannot answer

Two uncomfortable issues sit beneath the handshake photos.

The first is dependency. In June, Anthropic suspended access to its Fable and Mythos models for nearly three weeks to comply with US Department of Commerce export controls, restoring it on 1 July. That episode was a reminder that Washington, not Singapore, holds the off switch for America’s frontier AI. The MOU says nothing about that risk, and it is fair to ask how GovTech would cope if a similar order hit the models its engineers now rely on.

The second is commercial. Anthropic’s IPO prospectus shows two customers generated nearly a quarter of its 2025 revenue, while the company plans to spend US$518 billion on compute and infrastructure. Spreading revenue beyond a few big US clients is a necessity, and a market of wealthy, AI-hungry institutions is where that starts. The office is a safety partnership, but it is also a sales operation.

Also Read: Singapore’s AI dividend will depend on what happens after the pilot phase

Neither point makes this a bad deal for Singapore. Closer access to a frontier lab’s evaluation methods strengthens its claim to be the region’s AI assurance hub, and local developers, more than 300 of whom attended Anthropic’s Builder Summit this week, gain a shorter line to the company.

The real test will come later, in what the MOU produces: published testing standards, incident-response protocols that get used, and SME programmes with funding behind them. Until then, Singapore has a new tenant with an impressive customer list, and another piece of paper.

The post Anthropic’s Singapore MOU bets on safety testing, not consumer reach appeared first on e27.

Anyone who has spent 10 minutes pressing “0” to escape a voice bot knows how the last automation wave ended. Call centres were promised dramatic savings from IVRs and chatbots. Customers got phone trees that went in circles, and companies got a fresh layer of complaints to manage.

Sunny Rao sold that promise from the inside. Before becoming SVP of APAC at JFrog, he spent years at Avaya, Genesys and Nuance, three firms that defined the contact-centre automation era. Now he is watching Southeast Asia fall for AI agents, and he thinks the region risks repeating an old mistake: confusing a faster task with a more productive business.

Also Read: Singaporean founders’ Lightsage bags US$4M to decode how AI agents choose software

“An agent completing a task faster does not necessarily translate into greater productivity, particularly if it creates additional review, remediation or security work downstream,” he tells e27 in an email interview.

The adoption-readiness gap

The enthusiasm is not in doubt. A recent Agoda AI Developer Report found that 53 per cent of developers in the region say agents are already in production or broad use. Only 38 per cent believe their codebase is ready for that autonomy, a readiness gap e27 has examined before, and one that explains why regional tech leaders are learning to trust agents, but not with production.

Rao, who speaks regularly with CIOs in Singapore, Jakarta and Bangkok, frames the problem as one of control rather than capability. Coding agents now select dependencies, pull tools, modify code and interact with enterprise systems at machine speed. A company may know an agent is running without knowing what it consumed, which permissions it used or whether its actions complied with policy.

“The real readiness test is not whether an enterprise can deploy autonomous agents, but whether it can detect and control unintended actions before they become business risks,” he says. The worry has only sharpened since OpenAI shelved GPT-6.1 Astra, a model built to handle complex tasks with less human supervision.

Tokens are the wrong yardstick

Most companies still measure AI spend in seats or tokens. Rao thinks both miss the point. If a CFO could track one number, he would pick cost per successful, production-ready outcome.

The logic is simple arithmetic. An agent that is cheap to run but needs repeated prompting, produces unusable work or creates hours of human clean-up may cost more than a pricier one that gets it right first time. Tokens and licences are inputs, not value, a point that echoes e27‘s look at the hidden economics of autonomous AI agents.

Also Read: Agentic AI ambitions in Singapore run into legacy systems and data quality gaps

Take a pull request an agent writes in two minutes that a senior engineer then spends 45 minutes fixing. For software teams, Rao says, the better measure is Mean Time to Release, how long code takes to reach production, weighed against the company’s overall security posture. Code that introduces vulnerabilities or regressions can delay a release rather than speed it up.

In regulated industries, checks once run after deployment are becoming mandatory release gates, and two-minute code that cannot clear them is worth little.

The blind spot: what agents eat

Ask Rao where APAC enterprises are most exposed, and he points not at access but at consumption. Access is increasingly treated as an identity and permissions problem, which security teams broadly understand. Less visible is what agents pull in to finish a job: open-source packages, MCP servers (connectors built on the Model Context Protocol that plug agents into tools and data), skills and plugins. Any of these can carry vulnerabilities or malicious code into enterprise infrastructure before anyone notices.

“The critical control point is where an agent consumes a component,” he says.

When that control fails, the blame does not get diluted. If an agent pulls a compromised MCP server into production at a fintech regulated by the Monetary Authority of Singapore, Rao says the institution deploying it remains accountable, not the model provider or the governance vendor. That means being able to reconstruct which agent acted, under whose authority, what it consumed and why it was allowed. Many firms cannot do that yet: a Sumsub study found Singapore companies’ AI audit trails remain thin.

The vendor in the room

There is an obvious catch. JFrog sells exactly the controls Rao describes, through its MCP Registry, AI Catalog and Agent Guard. “Agents are risky, so buy our governance” is a convenient story for a vendor to tell.

Rao’s defence is that headcount no longer tracks risk. A 30-person startup running agents may operate workflows once associated with a far larger engineering team. If it builds for a bank or insurer, it inherits that client’s security and regulatory demands regardless of size. He concedes startups do not need bank-grade architecture, but argues the minimum, verifying what agents consume, limiting permissions and keeping evidence of how software was built and approved, becomes a commercial requirement long before any breach.

Pricing is harder to pin down. JFrog does not publish list prices for AI Catalog or Agent Guard; Rao says they are integrated capabilities within the JFrog Platform, not standalone products. That may be so, but it leaves buyers calculating governance ROI without a visible line item, an awkward gap for a company urging CFOs to measure cost per outcome.

JFrog is not alone in this race. Sonatype and Snyk have long pitched software supply chain security, and Snyk bought agent-security startup Invariant Labs in 2025. GitHub and Docker run their own MCP catalogues inside platforms developers already live in. For Southeast Asian teams, the question is less whether to govern agents than whose platform gets to hold the keys.

Singapore’s next exam

Rao expects Singapore to get agent governance right first, citing its regulatory track record and concentration of multinationals in regulated sectors. JFrog’s own 2026 Software Supply Chain State of the Union research found 71 per cent of Singapore respondents carefully review AI-suggested security fixes, the highest of eight markets surveyed.

Also Read: SEA’s insurers face a new question: what happens when customers have agents?

The same data shows the cracks. Some 18 per cent have policies against unauthorised AI tools but no way to detect violations, and 59 per cent wait a week or more for open-source package approvals. That is fertile ground for Shadow AI, employees quietly bringing in tools without IT’s knowledge, which Rao says worries organisations more than any rush to hand agents critical decisions. The one call he would keep with humans: which sensitive systems an agent may touch, and how far its permissions stretch.

The lesson from the IVR years still holds. Automation that merely made old processes cheaper disappointed; automation that rethought them paid off. Singapore’s next test, Rao says, is turning governance rules into automated enforcement inside developer workflows. Until then, the agent that writes code in two minutes may still be the slowest route to production.

The post Tokens are the wrong yardstick for AI agents, says JFrog’s Sunny Rao appeared first on e27.

When Singapore’s offices emptied out in 2020, flexible work was sold as the cure for an overworked city. Employees would get their commutes back, set their own hours and finally see their families at dinner.

Six years on, the cure appears to have come with side effects. According to the Jobstreet by SEEK Singapore Workplace Culture Study 2026, 52 per cent of employees say workplace stress has increased since 2019. Only 27 per cent say they have gained more flexibility and work-life balance.

Also Read: Building burnout-resistant teams

Employers broadly agree. Some 48 per cent of them say stress has risen, while 30 per cent think the balance has improved. Nearly half of both groups (46 per cent of employees and 45 per cent of employers) say job scopes have become more demanding.

In short, Singaporeans gained control over where they work, not over how much they work.

The flexibility paradox

The study’s central finding is a contradiction that many remote workers will recognise. Flexible arrangements changed the location and timing of work, but not its volume or complexity. Once the office walls disappeared, so did many of the natural boundaries that used to end the working day.

The commute that once marked the start and finish of work is gone. So is the shared understanding that leaving the building means being off duty. What replaced them, for many, is a state of constant, low-grade availability: the laptop on the dining table, the Slack ping at 9.30 pm, the “quick call” during school pickup.

This happened even as Singapore formalised flexible work. The government’s Tripartite Guidelines on Flexible Work Arrangement Requests, which took effect in December 2024, require employers to properly consider workers’ requests for flexible arrangements. The guidelines govern access to flexibility. They say nothing about how much work fills those flexible hours.

The finding fits a wider pattern e27 has tracked: burnout is no longer just a personal problem but an operations problem.

Same storm, different worries

The study suggests both sides see the same forces reshaping work, but feel them in very different places.

AI and automation are cited as a primary driver of workplace change by 56 per cent of employees and 53 per cent of employers. Economic uncertainty, inflation and the cost of living are named by 53 per cent of employees and 46 per cent of employers.

Also Read: The Philippines uses AI as a burnout shield; Singapore uses it for deep work

From there, the concerns split along predictable lines. For employees, the pressure is personal. Sixty per cent cite salary growth and the cost of living as their main worry, 48 per cent are focused on job security, 37 per cent struggle to protect work-life balance, and 33 per cent feel squeezed by competition for roles. It is the anxiety e27 contributors have described as global uncertainty turning into workplace anxiety.

For employers, the pressure shows up on the balance sheet and the org chart. Fifty-two per cent are preoccupied with how AI and automation will change workflows and team structures. Forty-nine per cent are wrestling with salary expectations in an inflationary environment, and 24 per cent cite restructuring and running leaner teams.

That last figure is the hinge of the whole story. When nearly a quarter of employers are running leaner teams, the work does not vanish with the departed headcount. It gets spread across whoever is left. That would explain why job scopes have grown, and why stress has risen alongside flexibility. Working from anywhere has become a convenient way to absorb the extra workload.

It is a local version of what e27 has called Big Tech’s efficiency paradox of record profits and record layoffs. There is also a regional warning attached: the AI layoff trap points straight at Southeast Asia.

Where the money is going, and where it isn’t

The most revealing data in the study concerns employers’ spending plans. The money is overwhelmingly flowing into technology and skills: 34 per cent of employers plan to invest in AI and digital skills, 25 per cent in new workplace technology tools, and 23 per cent in upskilling and reskilling.

The measures most likely to relieve the pressure employees describe are far lower on the list. Only 14 per cent plan wellbeing and mental health initiatives, 13 per cent plan workload management improvements, and just 12 per cent plan to hire more people to ease the workload.

Put bluntly, companies are trying to solve a capacity problem with software. Training a stretched employee to use an AI assistant may make them faster. It does not give them fewer things to do. Without changes to workload or headcount, technology investment risks simply raising the speed at which already exhausted people are expected to work.

That may be a rational short-term bet for employers watching costs. But the bill tends to arrive later, in the form of burnout, declining work quality and the quiet departure of the high performers who have the most options elsewhere.

Fixing the design, not the location

The study’s recommendations are less about where people work than about how the work is designed. They are practical rather than revolutionary.

• First, companies should run regular workload and priority audits that compare targets against actual headcount, so that scope creep is caught before it becomes chronic stress.
• Second, when business goals change, managers should state explicitly what will be paused or dropped, rather than stacking new priorities on top of old ones.
• Third, middle managers need proper training to run distributed teams and protect boundaries. Other research has found that the manager has an outsized effect on wellbeing at work. Fourth, performance should be judged on outcomes rather than responsiveness or how visibly “online” someone appears. That matters in a city where two in five employees already feel watched at work.

None of this is new to HR professionals. The hard part is that each step costs management time and, in the case of hiring, real money, which is exactly what most employers are not budgeting for.

A caveat on the numbers

The figures presented here do not come with sample sizes, survey dates or a methodology, so they are best read as directional rather than definitive. The comparisons with 2019 are also based on what respondents remember, not on data tracked over time.

Also Read: SEA’s CEOs want innovation, but employees feel stuck in execution mode

Still, the overall shape is hard to dismiss when both employees and employers report it. Neighbouring markets have treated flexibility as a competitive advantage; in the Philippines, flexible work is no longer a perk but the price of talent. Singapore’s experience suggests flexibility alone is not enough. Unless companies also limit how much work it carries, the freedom to work from anywhere can turn into the obligation to work from everywhere.

The post Work from anywhere, work all the time: Inside Singapore’s flexibility paradox appeared first on e27.

Southeast Asian companies are not short of money for AI. Most of them simply can’t get their projects past their own IT and security teams.

That is the uncomfortable picture painted by the Southeast Asia findings of Adobe’s 2026 State of AI Readiness Report, released today. The study surveyed 461 senior executives in the region and found that 73 per cent of companies are increasing their AI investment year-on-year. Yet 48 per cent are still working on their first AI use case, or have given up after a failed attempt.

Also Read: Everyone has adopted AI. Almost nobody is ready for it

In other words, budgets are going up while results are standing still.

About 57 per cent of organisations are stuck in what the report calls the “Frozen Middle”. They are piloting or deploying AI without the coordination needed to scale it. For agentic AI, meaning systems that can carry out multi-step tasks on their own rather than just answer prompts, the figure is 40 per cent. Some 79 per cent of the region’s companies sit at an early stage of AI readiness, the highest share anywhere in Asia Pacific.

The findings echo a theme that keeps resurfacing on these pages: Southeast Asia doesn’t have an AI adoption problem, it has a scaling problem.

Where the good ideas go to die

The most striking number in the report is the attrition rate. Only 15 of every 100 AI ideas make it into production.

The biggest losses do not happen in the lab or the boardroom. About 36 per cent of AI proposals fail to get past IT evaluation. A further 18 per cent are rejected at security review, and that rejection rate nearly doubles in highly regulated industries such as banking and healthcare.

Read one way, that is a story about bureaucracy strangling innovation. Read another, it is a story about gatekeepers doing their jobs. Many AI proposals arrive at IT with no clear data plan, no security model and no owner. e27 has previously reported on the hidden risk of unchecked agent privileges, and a security team that waves every agentic pilot through would be failing at its own job.

The more useful question is why so many proposals reach IT in a state that deserves rejection. The report’s answer is that nobody owns them.

The ownership vacuum

Only 12 per cent of Southeast Asian companies have an AI owner with budget authority, meaning someone who can both make decisions and pay for them. Among the companies Adobe classifies as Leaders, that figure is 89 per cent.

Measurement is just as thin. Only 8 per cent of companies measure AI return on investment comprehensively, and most track usage (how many people logged in) rather than quality (whether the output was any good). Meanwhile, 81 per cent of businesses say their AI skills shortfall stayed the same or got worse over the past year. That is consistent with what we have  seen in Malaysia’s AI skills gap and elsewhere in the region.

Also Read: Scaling beyond AI pilots: Six-move Capability Cycle

“Organisations that establish clear ownership, shared measures of success and stronger collaboration between business and IT will be best placed to move beyond isolated pilots and achieve sustainable impact,” said Jeremy Willmott, Senior Director, Digital Strategy Group Asia Pacific at Adobe.

It is a sensible prescription, if not a new one. Many consultants have offered similar advice, and e27 contributors have argued that Singapore’s AI adoption problem is not worker resistance but weak execution.

What the Leaders do differently

The Leaders are a very small club: 6.3 per cent of Southeast Asian respondents, or just 27 organisations. They have scaled AI across multiple workflows and built an operating model to sustain it. According to Adobe, they convert pilots to production at nearly four times the rate of other companies.

They also outscore the regional average by 67 per cent on ownership and accountability, and by 59 per cent on embedding AI into everyday processes. They keep their “destructive failure” rate, failures that cause real damage rather than lessons, below 1 per cent. By contrast, more than a third of companies repeat the same AI mistakes. And 41 per cent of Leaders say their single most effective move was funding the employees who were already making AI work.

That last finding is perhaps the most practical in the report. The fastest route to scale often runs through the staff member who quietly built a working tool in a spreadsheet, not through the vendor’s pitch deck. It fits other regional research showing that workers are sprinting ahead of their bosses in AI adoption.

Read the fine print

Before boards rewrite their AI strategies around these numbers, a few caveats are worth noting.

First, the sample. The research was conducted by Incisiv on Adobe’s behalf. It covered only organisations with at least US$100 million in annual revenue, 60 per cent of which earn more than US$1 billion. Startups and the SMEs that make up most of Southeast Asia’s economy are not in the picture.

Second, the Leaders’ statistics rest on just 27 companies in the region. When the report says 89 per cent of Leaders have a budget-holding AI owner, that comes to roughly 24 companies. That is enough to suggest a pattern, but not enough to prove it.

Third, the headline conclusion is partly built into the method. The report says organisational readiness, not technology or spend, is the strongest predictor of whether AI scales. However, Leaders were defined in the first place by combining deployment speed with “operating model maturity”. Finding that Leaders have mature operating models is therefore close to circular. The more interesting finding is the negative one: globally, company size, industry and reported AI spend did not predict Leader status.

Finally, Adobe sells AI tools to enterprises. It has an obvious interest in persuading companies that the obstacle to AI value is organisational, not technological. The company’s own earlier research found GenAI adoption rising in Asia while ROI remained elusive. This year’s report reads like the sequel.

Also Read: Why most enterprise AI in APAC is still stuck in the proof-of-concept room

None of that makes the diagnosis wrong. The pattern of rising spend, stalled pilots and nobody in charge matches what founders selling into Southeast Asian enterprises describe privately. Telcos and universities are now setting up joint efforts, such as Singtel’s RE:AI tie-up with the SIT-NVIDIA centre, specifically to close the deployment gap.

For B2B AI startups, the lesson is blunt. The buyer who signs the pilot is often not the person who can take it to production. Until someone with a budget owns the outcome, the region’s AI spending will keep rising faster than its returns.

The post Adobe study finds 57 per cent of SEA firms stuck in AI’s “Frozen Middle” appeared first on e27.

India’s MakeO has raised US$10 million in a round led by Singularity AMC, with unnamed funds and family offices also participating, eight months after the Mumbai-headquartered dental aligner company agreed to absorb Singapore’s Zenyum.

The cheque, disclosed in a statement on October 13, is modest by the standards of the clear aligner category’s go-go years. But it lands on a business that now claims something few consumer healthtech companies in the region can: more than SGD70 million (US$54.7 million) in annual recurring revenue (ARR) and sustained profitability on an EBITDA (earnings before interest, taxes, depreciation, and amortisation) basis.

Also Read: The most-funded healthtech startups in Southeast Asia: A decade in review

The money has three jobs. It will fund Zenyum’s dentist network expansion across Southeast and North Asia, strengthen the group’s technology, and expand manufacturing capacity. Zenyum also said it continues to evaluate acquisitions across India and other parts of Asia, a signal that the shopping basket is not yet full.

The combined group says it works with roughly 1,100 partner dentists across India, Singapore, Malaysia, Vietnam, Taiwan, Hong Kong, and Japan.

The deal behind the round

That dentist figure is, for now, largely the arithmetic of a merger rather than fresh growth. When the two firms announced their tie-up, Toothsi had about 500 partner doctors and more than 50 dental centres in India, while Zenyum had 600 doctor partners.

MakeO agreed to acquire Zenyum in February. Zenyum’s own framing at the time was more diplomatic, describing it as a “merger of equals” because both businesses had similar top lines and complementary geographies. Financial terms were never disclosed, and Zenyum CEO Julian Artopé declined to comment on valuation or the post-merger shareholding split.

The structure tells its own story. MakeO co-founder Arpi Mehta remains group CEO and Artopé continues to run Zenyum, with Mumbai as the group headquarters and Singapore retaining control of Southeast and North Asia operations. Crucially, Toothsi’s vertically integrated, US FDA-certified aligner platform now designs, manufactures and exports aligners for the whole business across Asia. Put simply: India is the factory, Singapore is the shopfront.

For Zenyum, it marks a notable turn. Founded in 2018 by Artopé and Frederik Krass, the startup raised a US$13.6 million Series A in 2019 and followed it with a US$40 million Series B led by L Catterton in 2021, backed by Sequoia Capital India (now Peak XV), RTP Global, Partech and others. Back then, Artopé was on a mission to help Asia smile more and Zenyum reported a fourfold revenue jump in 2020.

Five years later, a Series B-stage Singapore brand folding into an Indian peer is another data point for a regional exit market that increasingly favours trade sales over IPOs.

Also Read: ADA acquires India’s Algonomy in push beyond customer data into AI-led action

MakeO, for its part, has been on the acquisition trail. The Zenyum deal was its second acquisition in six months, and the company had earlier merged its Toothsi and Skinnsi brands under the MakeO umbrella in 2022. It raised a US$16 million bridge round in 2024 backed by Ashish Kacholia and 360 One Asset.

Reading the fine print

The headline numbers deserve a closer look. ARR is a natural metric for software subscriptions, but aligner treatment is largely a one-off purchase, topped up with consumables such as toothbrushes and retainers. The group did not explain how it calculates the figure, whether it is an annualised run rate, or how much comes from repeat customers.

The profitability claim also comes with an asterisk. EBITDA strips out depreciation and amortisation, which are precisely the costs that matter for a company expanding manufacturing capacity and carrying acquisition-related intangibles on its books. Being EBITDA-positive is a real achievement in a sector that has burned cash freely, as Carsome’s run of profitable quarters in used cars has shown is possible elsewhere. But it is not the same as net profit.

Then there is the size of the round. At under a fifth of the stated ARR, US$10 million looks like targeted growth capital rather than a war chest, which makes the continued talk of acquisitions intriguing. Either deals will be small, or more capital will follow. The valuation, and the identity of the participating family offices, remain undisclosed.

A crowded mouth

The rival landscape is formidable. Align Technology’s Invisalign remains the global category leader and sells through orthodontists and dentists across Asia. Swiss implant giant Straumann pushes its ClearCorrect brand, while China’s Angelalign has scaled aggressively in its home market and beyond.

The cautionary tale is SmileDirectClub, the US direct-to-consumer pioneer that collapsed into bankruptcy and shut down in late 2023, partly because cutting dentists out of the loop invited regulatory and clinical pushback. Zenyum and Toothsi have both leaned the other way, routing patients through in-person dentist consultations, which is why the partner network figure sits at the centre of this announcement.

The wider funding climate adds pressure. Southeast Asian healthtech endured a funding drought in 2024, and the investors still writing cheques increasingly want businesses built around healthcare workflows rather than flashy apps. Consolidation is picking up too, with Zeya Health acquiring ConnectLah to build booking rails for clinics.

What comes next

The clearest clue lies in what MakeO said in February. It flagged selective M&A and partnerships with local dental chains and clinical practices in markets including India, Singapore, Taiwan and Hong Kong.

Also Read: Vietnam’s healthtech boom has a talent problem nobody is talking about

That would mark a strategic shift. Owning clinics, rather than partnering with them, turns an asset-light network into an asset-heavy one, with more control over patient experience but more capital locked in bricks, chairs and staff. It would also put the group in direct competition with some of the very dentists it relies on for referrals.

For now, MakeO has a profitable (on its own definition) pan-Asian platform, an Indian manufacturing base and fresh capital. Whether investors keep smiling will depend on how much of that S$70 million turns into cash that stays.

The post MakeO secures US$10M to scale Zenyum’s dentist network across Asia appeared first on e27.

A founder I know recently received two term sheets in the same week.

The first was from a regional VC firm with a recognisable brand. The cheque was for US$1.5 million. The terms were reasonable. The partner had introduced himself at a conference six months earlier and had been pleasant in every interaction since.

The second was from an operator who had built and exited a company in the same space ten years earlier. The cheque was for US$400,000. The terms were similar. The operator had spent two hours on Zoom with the founder going through the customer acquisition strategy line by line, and had ended the call by listing six potential customer introductions he could make in the first month.

The founder took the smaller cheque.

When I asked why, his answer was direct. The bigger cheque came with a logo and a partner who would attend his board meetings. The smaller cheque came with someone who could shorten his customer acquisition cycle by six months. Six months of runway was worth more than the US$1.1 million difference.

That decision is becoming more common in 2026. It reflects a quiet but significant shift in how founders are evaluating capital.

The cheque is becoming the easiest part

For most of the past two decades, the cheque was the product. Investors raised funds, deployed capital, and provided light-touch support around the edges. The relationship was simple. The founder got the money. The investor got equity. Everything else was a bonus.

That model worked when capital was scarce and the value of the cheque was higher than anything an investor could provide around it.

Capital is no longer scarce for a growing share of early-stage founders. Solo GPs, family offices, corporate venture arms, angel syndicates, and a hundred new fund structures have flooded the market with capital options. The cheque has become the easiest part of the deal to get.

What is harder to get is everything around the cheque. Customer introductions that close in the first month. Engineering hires who would not have taken the call otherwise. Distribution partnerships that take three years to build from cold. GTM expertise from someone who has scaled the same playbook five times. Cross-border access into markets the founder cannot enter alone.

These are the assets that still have scarcity value. The investors who can deliver them are increasingly the ones founders are choosing. The investors who can only deliver capital are increasingly the ones founders are walking away from.

Also Read: Entering Taiwan: A practical resource map for foreign founders

The six things founders are actually buying

Ask a founder raising in 2026 what they want from an investor and the cheque will rarely be the first answer. Here are the six things that come up most often.

• Distribution. Can this investor get my product in front of customers I cannot reach myself. A US enterprise SaaS founder will pay a meaningful premium for an investor who can introduce them to Fortune 500 CIOs. A Southeast Asian consumer brand will pay the same premium for an investor who can open doors at major regional retailers.
• Customer introductions in the first ninety days. Not eventually. Not someday. Specific names, with introductions made within the first three months of the cheque clearing. Every customer an investor delivers in the first quarter is worth more than every customer the founder would have closed without that introduction.
• Hiring leverage. Senior engineering, product, and go-to-market hires are scarce everywhere. An investor who can pull a senior hire from their network in the first six months is delivering value worth multiples of the cheque size. Most founders underestimate how much faster the right hire arrives when the investor has personal relationships with the candidates.
• GTM expertise from someone who has done it. Strategy decks from a consultant are widely available. Operating advice from someone who built and scaled the same model is rare. Founders are increasingly choosing investors based on whether they have run the exact playbook the company needs to execute, not on whether they advise in general.
• Cross-border access. For founders whose markets sit in multiple countries, the value of an investor who can open regulatory, distribution, and partnership doors in the second and third market is enormous. A Singapore founder expanding into Indonesia, Vietnam, and the Philippines needs three different sets of local relationships. The right investor brings two of those sets pre-built.
• Strategic advice when things break. Companies do not run smoothly. Pivots happen. Growth slows. Customers churn. Co-founders leave. The investors who matter are the ones who pick up the phone when the company is in trouble, not the ones who only show up when the metrics look clean.

The investors who can deliver all six are rare. The investors who can deliver four or five are valuable. The investors who can only deliver capital are increasingly being filtered out by founders who have other options.

What this means for investors

The implication for investors is straightforward. The brand of the firm is no longer enough. The value delivered around the cheque is becoming the actual product. Investors who have not built a deliberate value-add capability are about to find themselves losing the deals they used to win on brand alone.

The investors who will keep winning are the ones who have built genuine assets around their capital. A real customer network. A real hiring pipeline. A real distribution rolodex. A real operating playbook in a specific category. Not slogans. Not platform teams. Not marketing claims about being founder-friendly. Actual delivered value that the founder can name and quantify.

Also Read: The 90-second pitch that helps foreign founders crack Tokyo’s networking scene

The brand premium that used to justify a US$1.5 million cheque over a US$400,000 cheque is shrinking. In some categories, it has already disappeared.

What this means for founders raising in 2026 and 2027

Three actions for founders evaluating capital this year.

Before accepting any cheque, write down what you actually need beyond the money. Be specific. Three named customers you cannot reach. Two senior hires you cannot recruit. One distribution channel you cannot build alone. Then evaluate every investor against that list. The right investor is the one who can deliver the most items on that list, not the one with the biggest cheque.

Ask every investor for three specific examples of value they delivered to a recent portfolio company. Not a general claim. A specific story. Which founder. What was the problem. What did the investor do. What was the outcome. The investors who can answer this question crisply are usually the ones who deliver. The ones who answer vaguely usually do not.

Be willing to take less money to get the right partner. The compounding value of an investor who delivers around the cheque can be ten times the differential in cheque size over three years. A founder who optimises for cheque size in 2026 is optimising for the wrong variable. The variable that matters is what the cheque comes with.

The investor-founder relationship is being rewritten. The investors who recognise that the cheque has become the easiest part of the deal will keep winning. The founders who recognise that the cheque is no longer the most valuable part of an investor’s offer will spend the next decade building with the right capital partners around them.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic.

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27.

Join us on WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected.

The post The new value test for investors: What founders are actually buying when they take your cheque appeared first on e27.

Southeast Asian companies are spending more on AI and getting little out of it. Adobe’s 2026 State of AI Readiness Report, which surveyed 461 senior executives in the region, finds 73% raising AI budgets year on year, while 57% remain stuck in AI’s Frozen Middle, piloting without the coordination to scale.

Only 15 of every 100 AI ideas reach production. IT evaluation kills 36% of proposals and security review another 18%, gatekeeping that partly reflects the risk of unchecked agent privileges. The deeper problem is ownership: just 12% of firms have an AI owner with budget authority, against 89% among the 27 Leaders, and only 8% measure ROI comprehensively. The findings reinforce e27’s argument that the region has a scaling problem, not an adoption one.

Leaders’ most effective move was funding employees already making AI work, echoing research on workers sprinting ahead of bosses. Caveats apply: the sample covers only firms with at least US$100M in revenue, the Leader cohort is tiny, and Adobe, an AI vendor, has an interest in blaming organisation over technology. The report reads like a sequel to Adobe’s earlier GenAI ROI findings.

Efforts such as Singtel’s RE:AI tie-up with SIT-NVIDIA aim to close that deployment gap.

—

REGIONAL

Malaysia’s Budget 2027 earmarks US$110.4M for startups: National fund of funds Jelawang Capital and KWAP’s Dana Perintis will deploy the money, Prime Minister Anwar Ibrahim told parliament, though e27’s analysis shows the bigger cheques go elsewhere in the budget.

TCGX closes US$600M Asia biotech fund, opens Hong Kong office: The US healthcare investor’s first Asia-focused vehicle was oversubscribed by international institutional investors and will back drug developers across stages. TCGX is also opening in Shanghai, but has not named LPs, cheque sizes or portfolio companies.

BII commits US$1.6B to SEA climate finance over five years: The UK development financier will put up about half and mobilise the rest privately, largely through British Climate Partners, targeting renewables, transmission, EV infrastructure and water in Vietnam, Indonesia, the Philippines, Malaysia and Thailand.

Zenyum parent MakeO raises US$10M to grow dentist network in Asia: Singularity AMC led the round. The group, with 1,100 partner dentists across seven markets, has passed US$54.7M in ARR, is profitable on an EBITDA basis and is weighing more acquisitions in India and Asia.

NEU Battery Materials raises US$7.6M to scale LFP recycling: Shift4Good and SGInnovate co-led the round after NEU commissioned a 1,300-tonne-per-year Singapore factory. Both backed its 2023 seed, and NEU’s emissions claims still lack an independent life-cycle assessment.

Vietnamese-founded Drylab raises US$3M for autonomous AI lab agents: Vertex Ventures Southeast Asia & India led the seed round, with PatSnap joining. The Singapore-based startup’s agents run entire computational research workflows across 800-plus tools, though it has yet to name customers.

UOB and AWS team up on hybrid cloud and AI banking across ASEAN: The MoU creates a dedicated Singapore cloud environment, AWS’s first hybrid-cloud bank tie-up in ASEAN, designed to meet data residency rules while testing conversational banking and personalised services across UOB’s five markets.

VinFast takes 36% of Vietnam’s car market in first nine months: The Vingroup EV maker’s 177,697 domestic deliveries already exceed its full-year 2025 total, and September marked its 25th straight month on top. Toyota trailed in second place with 53,978 units.

INTERVIEWS AND FEATURES

17 SEA startups that can actually prove their social impact: In a funding winter, founders stretch the social impact label to cover anything with a farmer or a recycling bin in the deck. e27 picks 17 startups taking the harder route of proving it.

Singapore workers won flexibility, then lost their evenings: Jobstreet’s study finds 52% of employees report higher stress since 2019, against 27% citing better balance. Leaner teams spread the workload, yet only 12% plan to hire to ease the pressure.

Vietnamese shoppers ask AI first, then check Google before buying: Ho Chi Minh City-based Q&Me finds 62% of buyers use AI, including generative tools, to research purchases, but Google search remains the final checkpoint before most shoppers commit.

INTERNATIONAL

Moonshot execs seek settlement with Recurrent.ai ahead of HK IPO: In Hong Kong arbitration since 2024, Recurrent.ai accuses the Kimi maker’s CEO and CTO of misleading investors and moving or deleting AI data, seeking 10.5% of Moonshot’s original shares, potentially worth over US$700M. Yang denies wrongdoing.

Firmus pulls US$5B Australian IPO, will turn to private capital: Investors balked at a roughly US$31B valuation, three times its August round, for an operator running 42MW of a 1GW target. Firmus is building AI factories in Singapore, Indonesia and Malaysia, with OpenAI anchoring two Malaysian sites.

IMF tips Vietnam to lead SEA growth at 8.4% as global GDP slows: The fund projects 3.1% world growth in 2026, with Malaysia at 5.2% and Singapore at 5%. SEA gains from the AI cycle mainly through trade and manufacturing, not by building AI infrastructure.

CYBERSECURITY

TikTok CEO Chew Shou Zi warns of AI agents’ cybersecurity risks: Speaking at DBS Asia Future Forward in Singapore, Chew said users may grant AI agents broad permissions, including access to credit cards, without grasping the consequences, and urged firms to stop staff sharing trade secrets with public chatbots.

Igloo exports cyber insurance to Turkey in first move beyond APAC: Through Aktif Bank’s Sigortayeri, the Singapore insurtech will sell cover for cyberbullying and fraud to the N Kolay app’s eight million users for about US$43. Mitsubishi Corporation brokered the deal.

Microsoft’s Brad Smith wants breach alerts for rogue AI agents: Calling AI a weapon and a tool that threatens global cybersecurity, Smith argued the UK, not the US or EU, should lead safety rules, and said voluntary pledges only bind responsible companies.

SEMICONDUCTOR

TSMC posts record US$46.7B Q3 revenue on AI chip demand: Sales rose 50% year on year and topped TSMC’s own guidance of US$44.6B to US$45.8B, as well as analyst forecasts. Full earnings, with margins and outlook, land on 15 October.

Sumitomo Bakelite adds chip packaging lines in Singapore and Suzhou: The Japanese firm expects to roughly double underfill material capacity for AI and satellite chips, with Singapore output from early 2029. It also completes its US$190M purchase of Kyocera’s chemical business this month.

FedEx report finds AI boom pushing chipmakers across Asia Pacific: The AI race now hinges on where chips are made, how far they travel and whether supply chains survive the next geopolitical shock, putting logistics resilience on the agenda for SEA manufacturers.

a16z sends infrastructure team to South Korea to court chip firms: Partner Raghu Raghuram will meet Korean AI hardware firms and officials weeks after the US$1.1B Machine Age Fund launch, eyeing semiconductor and robotics partnerships as the country cements its role as AI’s memory backbone.

AI

Anthropic test agent sent false murder tip to Philadelphia police: The city’s spam filter caught the July tip, and Anthropic took over two months to spot it. Police called the delay unacceptable, and Anthropic is now cutting its internal evaluations off from the live internet.

Satya Nadella calls for an ’emergency brake’ on AI models: Microsoft’s CEO wants every meaningful model action logged with tamper-proof evidence and an authorised person able to pause a model mid-task. Firms should assume a model is compromised from the start, he argued.

Singapore SMEs save time with AI, but messy data stalls scale-up: Small businesses cite time savings as AI’s main payoff, but systems that are not ready stop pilots from running more of the business, mirroring the scaling gap larger firms face.

India’s Clueso raises US$10M Series A for AI business video: Dharana Capital led the round, lifting total funding to US$12M. Clueso, used by Atlassian and Zendesk among 1,000-plus companies, will spend the money on enterprise sales and stronger agentic editing tools.

THOUGHT LEADERSHIP

US-China trade enters managed rivalry; SEA firms must adapt: Shawn Balakrishnan argues Trump’s May visit to Beijing produced a transactional framework, including new Boards of Trade and Investment, that Southeast Asian businesses must now plan around.

Deeptech founders should split their funding needs by job: Jun Yan argues one round size hides money doing different jobs, from funding an unproven product to buying equipment, carrying inventory or financing receivables, and each calls for a different kind of capital.

Human work is priced by the hour, machine thought by the token: Sayan Ghosh notes that only machine intelligence prices are falling fast, while timesheets and pricing models still assume human hours, forcing organisations to rethink how they value knowledge work.

Widespread AI tool use does not make a firm AI-ready: Elvis Lau argues that staff using ChatGPT and Copilot daily, plus a handful of production pilots, can still leave enterprises short of genuine AI readiness at the organisational level.

SEA’s SMEs should chase AI value, not just automation: Arnold Warner argues the real question is where AI creates measurable value without disproportionate complexity, cost or risk, a distinction that matters for the region’s small businesses.

Stop building new apps; build inside the ones users already open: Citing Sensor Tower, Manit Parikh notes users open around ten apps a day and global downloads grew just 0.8% last year, so founders should build inside existing platforms.

AI video is fast to generate and slow to make usable: Drawing on his work with AI-generated cinematic video, Jay Chua explains why prompt-to-scene speed is no guarantee of producing a usable video, where most of the effort now sits.

SEA retailers need observability to protect AI customer service: With Singapore offering a 400% tax deduction on AI spend for smaller retailers, Simon Rizkalla argues that real-time monitoring of AI systems decides whether the customer experience holds up.

The post Ecosystem Roundup: Frozen Middle: 57% of SEA firms can’t scale their AI pilots appeared first on e27.