Speaking in Doha on 28 October 2026 at the United Nations Trade and Development (UNCTAD) World Investment Forum 2026 panel, “Investing in AI for an Inclusive Digital Future”, Ant International CEO Peng Yang called for a new, inclusive “re-globalisation” project that harnesses AI and digital connectivity to give small and medium-sized enterprises (SMEs) greater access to global markets and opportunities.

“Globalisation 1.0 mostly benefited large enterprises, and a lot of SMEs very rarely got to participate in the global economy,” said Yang. “Achieving inclusion is critical to re-globalisation, and AI will accelerate it by democratising opportunities.”

The limitation of Globalisation 1.0 and how AI and other innovations offer a chance to improve

Yang said one of the greatest limitations of Globalisation 1.0 was its concentration of power among large enterprises. Small businesses, which make up over 90 per cent of businesses globally, have often been left behind due to the gap in technologies, talent and other resources. It is difficult for them to participate in the last round of globalisation, which involves handling multiple markets and the complexities of global trade.

According to the WTO’s 2024 World Trade Report, eight per cent of small firms and 15 per cent of medium-sized firms export directly, compared with 30 per cent of large firms.

The AI divide compounds the challenge: a 2025 survey by the Organisation for Economic Co-operation and Development (OECD) shows 52 per cent of large firms used AI, compared with just 17.4 per cent of small firms.

However, the rise of AI means we have the chance to level the field and close the gap if we democratise AI technology and its applications to give equal opportunities to SMEs.

Also Read: Why Indonesian fintech’s ASEAN expansion cost more than anyone priced in

When innovation meets inclusion, SMEs can also grow and thrive on the global stage

Yang said that in the past three years, Ant International has been serving over 1 million new small merchants from 54 new markets globally, empowering them to participate in global trade.

“In the last three years, we have seen tremendous growth for SMEs when innovation meets inclusion. When we offer them powerful full-stack AI-native solutions, cross-border transaction volume of SMEs across our own businesses has been growing by about 50 per cent every year, which is 10 times the growth rate of the total volume of global trade,” he said.

“This shows that if we can do things right, drive innovation, provide the AI technologies, and make it more accessible for the small guys, they can take the global stage to grow and thrive,” Yang added.

Ant International processes 25 million transactions daily across its global payment, global account, embedded finance and digitalisation services. 95 per cent of its 150 million-strong global merchant customer base are SMEs. In September 2026, Ant International announced sweeping full-stack AI-native solutions on payment, account, FX, treasury and growth for global businesses, many designed to tackle SMEs’ challenges.

New full-stack AI-native solutions empowering SMEs to drive truly inclusive re-globalisation

• Democratising the Falcon TST FX model for SMEs

Cross-border FX and money management is expensive for SMEs. Ant International has been developing its Falcon TST AI forecast model, which achieved SOTA (state-of-the-art performance) on the MASE metric on a top global benchmark, with a consistent forecast accuracy rate of over 93 per cent, and has been working with the world’s leading banks to keep increasing coverage. Today, it has been democratised as part of the full-stack AI-native solutions for SMEs to use so they can handle global transactions more efficiently with less FX risk.

• Proprietary agents WorldFirst Autopilot and Antom Autopilot automate operations for SMEs

Under WorldFirst and Antom – Ant International’s core Global Account and Global Payment services – enterprise-grade financial capabilities are engineered into SME-friendly applications.

WorldFirst combines cross-border collections, payments, foreign exchange and fund management in one global account. This enables smaller businesses to manage international operations, collect money and make payments with less FX risk and access working capital at lower cost, even without finance and treasury teams. Ant International’s global AI business agent for SMEs — WorldFirst Autopilot — further simplifies the operations by streamlining financial tasks and connecting businesses to partner services through natural language interactions.

WorldFirst helps over 2 million SMEs to manage their cross-border accounts, with more than US$500 billion in cumulative transaction volume across 210+ markets. Its network connects SMEs to 130+ e-commerce platforms and 600+ ecosystem partners, helping them navigate overseas markets and grow globally.

Antom Autopilot simplifies payment operations for merchants, especially SMEs, with AI support across onboarding, integration, analysis, risk management and reconciliation. By the end of 2026, 95 per cent of merchants are expected to complete onboarding with Antom within 1 hour. Its SmartDispute provides dispute analysis and appeal strategies, helping merchants achieve a 41.8 per cent win rate versus 24.5 per cent without the tool, while its fully automated AutoDispute helps merchants recover an average of 18 per cent of chargeback losses, reducing the additional resources SMEs need to devote to dispute management.

Also Read: Fintech, DeFi and applied AI define Southeast Asia’s new venture discipline

• Alipay+ empowers SMEs to access global consumers with AI-powered marketing solutions

For Global Payment, Alipay+, Ant International’s global wallet gateway, connects over 50 digital wallets to more than 150 million merchants globally, over 95 per cent of them SMEs. Beyond seamless cross-border payments, small merchants are targeting global consumers online and in-store with more effective and precise campaigns through Alipay+’s AI-powered marketing solutions.

In the Siam Square commercial area in Bangkok, Thailand, the Alipay+ marketing campaign combining pre-trip social-media push and on-site ad displays helped small merchants grow TPV by 66 per cent YoY, with an increase in the number of orders by 115 per cent YoY; for merchants in Gaya Street in Kota Kinabalu, Malaysia, it drove more than 120 per cent growth in TPV and 140 per cent in orders; for a local skincare brand in Bali, Indonesia, it delivered a growth of 137 per cent in TPV and 140 per cent in orders during the campaign.

New foundations of trust for inclusive AI in public-private sector collaboration

Yang also stressed that businesses, including SMEs, need confidence that AI can be used safely and responsibly, particularly as AI agents take on financial and operational tasks. Effective safeguards require collaboration across the technology and financial industries, alongside regulators and international institutions.

“Trust is everything in payments and financial services, and it becomes even more critical in the AI era. We need to work closely with international organisations and regulators to develop appropriate frameworks, while building safeguards that allow businesses to use AI agents safely. Collaboration across the private sector must also be complemented by international institutions and regulators to establish common standards and trusted infrastructure,” said Yang.

Ant International is advancing this approach through collaboration across the public and private sectors. In Singapore, it works with the Infocomm Media Development Authority on privacy-enhancing technologies and participates in the Monetary Authority of Singapore’s BuildFin.ai initiative on developing and scaling responsible AI solutions for financial services.

The company is also collaborating with Visa and Mastercard on interoperability for Know Your Agent (KYA), helping strengthen trust in AI agents acting on behalf of consumers and businesses.

It also takes a step forward by introducing the industry-first Antom AgentSafePay, built to protect merchants navigating agentic commerce, to verify multi-dimensional purchase intent in real time, and to offer a fund guarantee against losses from intent misinterpretation, malicious prompt injection or other risks within merchant-deployed agents.

Yang said that, with the cost of AI lowered and accessibility made easier for SMEs, both small merchants and individuals can be equipped with enterprise-grade innovative solutions, fully automate their operations and benefit from the next global economic revolution.

“We believe in a future where the new normal of Globalisation will leave no one behind. When innovation meets inclusion, everyone can pursue their opportunities like the big guys did in the past,” Yang said.

The ninth WIF, the UN’s flagship forum for the global investment community, brings together policymakers, investors and business leaders to explore how investment can support sustainable, resilient and inclusive growth amid geopolitical fragmentation and rapid technological change.

—

Enjoyed this read? Don’t miss out on the next insight. Join our WhatsApp channel for real-time drops.

This is a sponsored article produced by the e27 team.

Featured Image Credit: Ant International

The post Ant International CEO: Achieving inclusion is critical to re-globalisation, and AI will accelerate it by democratising opportunities appeared first on e27.

Ask a deeptech founder in Southeast Asia what keeps them up at night and few will blame the science. The harder problem is finding someone willing to let an unproven technology loose on a real building, a real plantation or a real data centre and then pay for it.

That is the gap Singapore Management University (SMU) is now selling its flagship competition on. The 13th Lee Kuan Yew Global Business Plan Competition (LKYGBPC), run by SMU’s Institute of Innovation & Entrepreneurship (IIE), opened applications on Wednesday for university founders and recent graduates working on urban and sustainability problems.

Also Read: Thirty climate tech startups in Southeast Asia that would rather build than pledge

Fifty finalists will be flown to Singapore from 28 September to 1 October 2027 to compete for more than SGD2.5 million (US$2 million) in prizes. That pot is roughly where it stood for the 12th edition’s US$2M grand finals. What has changed is the story around it.

From trophy to pipeline

The organisers now frame the contest less as a podium and more as a front door. Selected finalists can move into SMU’s Urban SustaInnovator (USI) accelerator, launched in 2025, which promises market access, pilot projects and introductions across Asia. Behind that sits the US$10 million USI Fund SMU IIE launched in May to back ventures as they pilot and scale.

“Founders of breakthrough technologies can create value only when they reach the people, industries and communities they are designed to serve,” said Rony Palathinkal, who heads both LKYGBPC and USI at SMU IIE.

The competition covers 12 focus areas, from climate-tech, energy and urban mobility to food and land use, longevity and healthtech, and sustainable materials and circularity. It is open to startups with at least one founder still enrolled at university or within five years of graduating.

US$10 million is a modest cheque by venture standards, likely enough for a handful of seed-stage bets. But the fund signals an important admission: a pitch trophy, on its own, rarely moves a deeptech company forward.

What the alumni actually got

SMU’s evidence comes from three 12th-edition finalists, and it is worth reading closely.

Malaysian agritech Qarbotech turns agricultural waste into compounds that lift crop productivity — the kind of farm-waste play some think could mint the region’s next climate unicorn. It is no stranger to pitch stages, having also taken top honours at Tokyo’s SusHi Tech 2026 contest.

Co-founder Chor Chee Hoe credits the LKYGBPC and USI network with meetings with family offices that led to several pilots. “Right now, we are halfway through a 12-month pilot with a large sugar cane plantation in the Philippines and we’re launching in another Southeast Asian market,” he said. An SMU connection also found Qarbotech a distributor for Thailand and Pakistan.

Also Read: Your deeptech company does not have one capital problem

Singapore’s Standard Physics, formerly Inviscid AI, combines real-time IoT sensor data with computational fluid dynamics (the physics of how air and heat move) to build digital twins that help buildings and data centres cut energy use. Founded by Qiu Ziming and Kabir Jain, it has secured a pilot with The GEAR, the Singapore innovation hub of Japanese construction giant Kajima, and is discussing three more. It recently raised a US$2.3 million seed round led by Square Peg Capital.

“YC was good at getting us from zero to one, and then USI has been very helpful in getting us from one to 10,” Jain said, referring to the startup’s stint in Y Combinator’s Winter 2026 cohort.

The third, Cambridge-based Mimicrete, founded in 2021 by Dr Liz Zijing Li, is developing “self-healing” concrete that seals its own cracks, extending infrastructure lifespan and cutting maintenance and carbon. It is validating the material at The GEAR, joining a growing list of startups such as Neocrete betting on low-carbon concrete.

These are encouraging stories, but they are pilot stories. A 12-month trial is a hypothesis, not recurring revenue, and Southeast Asian corporates are far better at starting trials than signing purchase orders, a problem e27 has tracked in the “frozen middle” where 57 per cent of firms cannot scale their AI pilots. Attribution is murky too: Standard Physics had Y Combinator in its corner before USI.

Singapore’s sandbox sales pitch

The timing leans on Singapore’s ambition to be the region’s test bed for sustainability tech. Its SGD37 billion (US$29 billion) Research, Innovation and Enterprise (RIE) 2030 plan earmarks SGD10.8 billion (US$8.4 billion) for mission-oriented research, including urban solutions and sustainability. In February, two ministries added a SGD40 million (US$31 million) USS Translation Fund to help companies pilot and commercialise such solutions.

SMU’s pitch to founders is that it can carry them beyond the island. It runs overseas centres in Bangkok, Jakarta, Ho Chi Minh City and Shenzhen, and International Advisory Councils in eight countries. That network “enables us to forge connections that enable startups to cross borders,” said Professor Sun Sun Lim, SMU’s Vice President of Partnerships and Engagement.

The border-crossing is the real test. A sensor-laden pilot in a showcase Singapore building is a long way from a cash-strapped developer in Manila or a palm oil estate in Sumatra, where price tolerance is lower and procurement slower.

A crowded field for founders’ attention

LKYGBPC is not short of rivals. Accelerators such as Startupbootcamp now run dedicated sustainability cohorts in Singapore, Tokyo’s SusHi Tech hosts its own global pitch contest, and corporate innovation arms court foreign deeptech, as Deloitte’s recent demo day for Tokyo startups eyeing Southeast Asia showed. Closer to home, NUS and NTU run their own deeptech commercialisation programmes.

LKYGBPC’s edge is reach. Its 12th edition drew more than 1,500 deeptech entries from founders at 1,200 universities across 91 countries, including Stanford, Cambridge and the University of Tokyo, before ChemT and Polar Code took the top prizes. SMU says finalists from the 9th to 12th editions have collectively raised more than SGD1 billion (US$780 million).

Also Read: Startupbootcamp’s first Singapore sustainability cohort moves beyond generic climate tech

That figure deserves an asterisk. It is self-reported, cumulative across four editions and likely includes capital many of these teams would have raised anyway. Competitions are good at selecting winners; proving they made them is harder.

Shirley Wong, Entrepreneur-in-Residence at SMU IIE, Managing Partner of TNF Ventures and chair of the 13th edition’s advisory committee, frames urbanising Asia as both problem and opportunity. Applications close on 31 March 2027.

The more telling scorecard, though, will arrive before next year’s winners are crowned: how many of the last cohort’s pilots have turned into paying contracts.

The post LKYGBPC returns with a pitch to fix deeptech’s real bottleneck: customers appeared first on e27.

GoTo has now reported a net profit for three quarters in a row. Read past the headline, though, and the numbers show a company run by its payments arm, while its ride-hailing business absorbs the cost of a government-imposed fare squeeze.

GoTo reported a third-quarter net profit of Rp441 billion (about US$24.6 million), compared with a net loss of Rp255 billion (US$14.2 million) a year earlier.

Net revenue rose 17 per cent to Rp5.6 trillion (US$313 million). Group adjusted EBITDA jumped 84 per cent to Rp950 billion (US$53 million), and GoTo raised its full-year adjusted EBITDA guidance to Rp3.5-3.7 trillion (US$196-207 million), up from Rp3.2-3.4 trillion.

Also Read: GoTo’s first profit signals a fintech-driven future

CEO Hans Patuwo did not downplay the context. It was the first full quarter since the 8 per cent commission cap on GoRide took effect, and he said, “the impact has been significant.”

The 89 per cent growth is almost all GoPay

GoTo’s most striking number, an 89 per cent jump in core gross transaction value (GTV) to Rp194 trillion (US$10.8 billion), is mostly a payments figure. GoPay accounted for Rp187 trillion of that total, up 96 per cent.

The on-demand business that made Gojek a household name grew GTV by just 1 per cent, to Rp16.9 trillion (US$944 million).

Profitability follows the same pattern. Fintech adjusted EBITDA rose 304 per cent to Rp550 billion (US$30.7 million) on net revenue that grew 49 per cent to Rp2.3 trillion. On-demand services (ODS) managed Rp347 billion (US$19.4 million), up 3 per cent. Fintech now produces nearly three-fifths of the group’s adjusted EBITDA.

That continues the trend from the first quarter, when the fintech unit was already doing the heavy lifting. Six months later, the gap between the two halves of the company has widened.

The commission cap bites where it was meant to

The cost of Indonesia’s new rules shows up in mobility. Mobility net revenue fell 5 per cent to Rp757 billion. Segment adjusted EBITDA dropped 16 per cent to Rp159 billion (US$8.9 million), and margin slipped 50 basis points to 2.5 per cent.

GoTo puts the cap’s direct cost at roughly Rp150 billion (US$8.4 million) for the quarter, in line with its Rp300 billion full-year estimate. GoRide went EBITDA-negative in July. The company says it is now trending towards break-even, but it has not said when it will get there.

Gojek’s response has been to chase volume rather than value. Completed mobility orders rose 10 per cent, driven by mass-market products with lower ticket prices, while mobility GTV did not grow at all. In plain terms, drivers are completing more trips that are each worth less.

The politics behind those numbers are now formal. A Presidential Regulation released this month locks in the 8 per cent cap on two-wheel transport and classifies driver-partners as micro-entrepreneurs, not employees. That gives drivers a bigger share of each fare but stops short of the labour protections that employment would bring.

GoTo points to its Partner Appreciation Program, which ranges from BPJS health and employment cover to scholarships and fully funded Umrah pilgrimages. These are welcome benefits, but they are discretionary and can be withdrawn, unlike rights under labour law.

Also Read: Grab’s US$1.49B Atome deal signals a deeper race for SEA’s credit economy

Delivery is the next uncertainty. GoTo is still waiting for implementing regulations from the Ministry of Communication and Digital Affairs, and it has widened its ODS guidance to Rp1.4-1.6 trillion to allow for them. Delivery performed better this quarter: adjusted EBITDA rose 20 per cent to Rp227 billion and margin improved to 2.1 per cent. Investors will be watching whether regulators extend similar pressure to food delivery.

A bigger loan book and a quiet admission

GoPay’s strength depends increasingly on lending. Loans outstanding rose 51 per cent to Rp12 trillion (US$670 million). Monthly transacting users grew 26 per cent to 30.4 million, while transactions grew 86 per cent to 2.8 billion.

The firm’s press release also contains a carefully worded admission. GoTo says it “pre-emptively tightened credit” for its highest-risk cohorts in the second quarter, which “limited the deterioration in credit quality” in the third. That means credit quality did get worse. As in previous quarters, GoTo has not disclosed non-performing loan ratios or delinquency figures. It argues that short loan tenors let it adjust underwriting quickly, but the market still has no data with which to check that claim.

The bottom line also needs some unpacking. Operating income was Rp351 billion, below reported net profit, so items below the operating line helped. Profit attributable to owners of the parent was higher still, at Rp557 billion (US$31.1 million). These figures are cleaner than the adjusted measures that drew scrutiny a year ago, but net profit is not yet driven mainly by operations.

Buying back stock at half the old floor

The capital-return story may tell investors the most. After the Indonesia Stock Exchange removed the Rp50 floor price on 28 September, GoTo spent Rp985 billion (US$55 million) buying back 36 billion shares, almost 3 per cent of its stock, by 29 October. That works out to an average of about Rp27 per share, roughly half the old floor and a fraction of the Rp338 IPO price.

GoTo will also cancel 32 billion treasury shares, about 2.7 per cent of shares outstanding, as approved at its 14 October extraordinary general meeting. It still has Rp2.5 trillion (US$140 million) available under a buyback authorisation that runs until June 2027.

Adjusted free cash flow of Rp1.2 trillion (US$67.5 million) means GoTo can afford the programme. Still, the fact that its stock is trading well below the old floor shows how little confidence the market currently has in the business, despite three profitable quarters.

The neighbours, and the merger nobody has ruled out

The competitive picture is not getting easier. Grab is moving deeper into credit through its Atome deal and its takeover of Superbank in Indonesia, putting it in direct competition with GoPay’s most profitable activity. ShopeePay remains a formidable rival in payments. On the streets of Jakarta, VinFast’s Green SM is challenging the gojek economy with e-scooters, at a time when the commission cap has made two-wheel rides less profitable for every platform.

Also Read: Taiwan kills Grab’s US$600M foodpanda deal, and Uber’s shadow is the reason

Grab faces its own questions. Its recent profit headline concealed cash burn in key areas, and the long-running prospect of a Grab-GoTo merger still hangs over both companies. A regulator that sets ride-hailing commissions is also likely to shape any deal between Indonesia’s two largest platforms.

Meanwhile, GoTo is investing in technology. Its 27-billion-parameter model that speaks Batak is a sign of ambition, though AI has yet to appear clearly in the margins.

What three quarters prove, and what they don’t

Three consecutive profitable quarters show that GoTo is no longer the cash-burning conglomerate investors once worried about. They do not yet show that Gojek can be a strong business under Indonesia’s new rules.

At the moment, GoTo is effectively a lending and payments company with a ride-hailing app on top. That can be a sound business model, but it carries more credit risk than the brand suggests. The next test is how GoTo handles a regulator that has already capped one of its revenue lines and may soon reach another.

The post Indonesia’s 8 per cent cap cost Gojek US$8.4M in a quarter. GoPay covered the bill appeared first on e27.

As Battery energy storage (BES) investment scales toward US$5 trillion, Hong Kong Science and Technology Parks Corporation (HKSTP) is positioning Hong Kong as Asia’s launchpad for materials, integration, green finance and recycling — a bridge Southeast Asian founders can use to finance, certify and scale.

A battery used to be something most people thought about only when a phone ran flat or a car wouldn’t start. That is no longer the case. BES is becoming part of the infrastructure that makes power grids more resilient, keeps renewable electricity available regardless of the weather, keeps electric vehicles moving and keeps increasingly power-hungry data centres running. Market projections cited by HKSTP indicate that investment in the battery and energy storage sector will reach US$5 trillion over the next decade.

Also Read: ASEAN battery push enters a harder phase: building factories, standards and markets

Look closely at this trillion-dollar industry and it is clear the global battery race has moved far beyond simply manufacturing cells. A complete BES ecosystem runs across four interconnected stages. It begins upstream with advanced materials research. It moves midstream into system integration and smart power electronics that bring the technology into community applications. It relies on international green finance to de-risk and fund project rollouts. And it ends downstream with the safe discharge, second-life reuse and recycling of spent batteries, which closes the sustainability loop.

In its latest BES Strategic Analysis Report, HKSTP describes Hong Kong as Asia’s leading hub with a complete, four-stage value chainand why that matters for the region.

Beyond the battery cell

At the upstream end of the chain, chemistry comes first. Research is advancing in solid-state batteries, water-based flow batteries and carbon-nanotube current collectors. Sodium-ion and other next-generation technologies are also being explored to improve cost, energy density, safety and durability.

Hong Kong’s universities give the city a solid foundation in this field. Luquos Energy is one example already in commercial use. The HKSTP company is developing a water-based sulphur flow battery, and its system has been deployed with EV-charging infrastructure on the Chinese mainland, where it has cut users’ electricity costs by close to 70 percent.

A storage system has to work with charging equipment, buildings or the electricity grid. Software has to decide when batteries should charge or discharge. Battery-management systems monitor safety and performance, while distributed-energy platforms coordinate different sources of electricity and storage.

This brings us to the second stage: integration and intelligence. Applications already range from mobile battery generators on construction sites to fast EV charging. In other words, some of the most valuable work may happen around the battery rather than inside it.

Where finance enters the chain

 Large-scale energy-storage installations require substantial upfront capital. Returns can stretch over many years, and revenues may depend on electricity tariffs, grid services and how often a battery is used. That makes finance part of the proposition. 

Also Read: Thirty climate tech startups in Southeast Asia that would rather build than pledge

The Hong Kong Taxonomy for Sustainable Finance covers activities including battery manufacturing and recycling, while energy storage supports greater use of renewable electricity. The framework can give banks and investors a clearer basis for assessing whether BES projects qualify for green and sustainable financing.

Building on this momentum, HKSTP recently signed a memorandum of understanding (MoU) with a “Green Accelerator” to inject capital into high-potential solutions and help startups cross the commercialisation “valley of death”.

This combination of green finance, research and international compliance has already drawn global attention. Three of the world’s leading battery and energy storage companies, BYD, CATL and Gotion, have established a presence in the HKSTP ecosystem. Their arrival signals that Hong Kong’s role is not to compete in mass manufacturing, but to serve as a global launchpad and a bridge across every stage of the BES value chain. For Southeast Asian founders, the relevant question is whether that bridge can be used for financing, compliance and market entry — not just for visibility.

What happens when batteries retire

Every battery eventually reaches the end of its first life. For Hong Kong, that is becoming less theoretical. The Environmental Protection Department estimates that retired EV batteries will rise from around 1,300 tonnes in 2026 to approximately 6,700 tonnes by 2030.

Hong Kong’s first large-scale EV-battery recycling facility, at EcoPark, is intended to turn retired batteries into recycled black mass. The government has also been working on a producer-responsibility scheme.

Some batteries still have useful capacity after they are no longer suitable for vehicles. They can be repurposed for stationary storage, giving them a second life before their materials are recovered. The BES report also identifies second-life applications and technology for safely discharging retired batteries as opportunities at the final stage of the chain.

Recycling can then feed materials back to the start of the cycle. That reduces reliance on virgin resources and creates incentives to design batteries with recovery in mind. A value chain that looks linear can become increasingly circular.

An Vibrant ecosystem around the technology

Battery storage is only one part of HKSTP’s green-tech ecosystem activity. More than 250 green-tech companies are now in its ecosystem. GreenTech Hub, launched in 2025, facilitated more than 360 business-matching sessions, and organised more than 100 green-tech events in its first year, while its public-private partner network grew to 20 organisations.

Several companies in the HKSTP ecosystem have also produced technologies that HKSTP describes as world-first or Hong Kong-first applications.

One Energy has developed a battery-swapping system that incorporates what HKSTP describes as the world’s first water-cooled fast-charging system, and has taken its electric-mobility business into Southeast Asia, Europe and Africa. Halo Energy, meanwhile, has partnered with XPeng to launch Hong Kong’s first megawatt-level ultra-fast charging station.

The technologies differ, but their commercial needs often overlap. The journey starts with talent driving breakthroughs at the research stage. It moves on to validating solutions in real-world settings, then to using financial capital and listing mechanisms to scale, and finally to connecting with international investors to go global. That path to commercialisation is notoriously difficult, and Hong Kong is positioning itself to support green-tech ventures at every turning point.

Also Read: GenZero backs PCG Global’s push to export China-tested renewable energy model

For Southeast Asian founders and investors, Hong Kong’s advantage is clear: green finance and bankability support, midstream integration and software expertise, and emerging second-life and recycling capacity. Hong Kong is not competing on mass manufacturing — it is building the research, finance, compliance and connectivity layer that regional BES projects need.

With HKSTP’s ecosystem, GreenTech Hub and international partner network, that support is already accessible. For Southeast Asian founders, the opportunity is straightforward: use Hong Kong as the launchpad to finance, certify and scale the next battery storage project in their own market.

The post Powering the green transition: Why the global battery race runs through Hong Kong appeared first on e27.

Indonesia’s sharia P2P lending sector looks different in 2026 than it did when Dana Syariah Indonesia and a cohort of sharia-compliant fintechs scaled between 2018 and 2022. The sector that aimed to bring Islamic finance principles into digital lending has consolidated significantly. Several platforms have been through restructuring. Dana Syariah Indonesia, whose OJK license was revoked in 2024, is among the more publicly documented examples of a sharia P2P platform whose operations did not survive the sector’s accountability phase.

After four years inside sharia risk policy at Indonesian banks and a further fifteen across the country’s financial functions, I have watched this sub-sector with more interest than most. The structural questions underneath the sharia P2P journey are not addressed by conventional fintech analysis — and the lessons matter for whoever builds the next chapter of Indonesia’s Islamic fintech infrastructure.

What sharia P2P aimed to do

The Indonesian sharia P2P sector that emerged after OJK and DSN-MUI established frameworks for digital Islamic financing had a clear thesis. Indonesia’s Muslim-majority population represented a credit access market that conventional banking and conventional P2P were not adequately serving with sharia-compliant products. The platforms that built early aimed to combine the operational efficiency of digital lending with the religious compliance of classical Islamic finance.

The implicit assumption was that the sharia overlay would be additive — fatwa review, DPS oversight, sharia-compliant contract structures — and that the underlying business model would work as conventional P2P did. The audience would respond to sharia compliance with stronger engagement and stickier behaviour.

Also Read: Youth brain drain: A hidden threat in Indonesia’s agriculture sector

What turned out to be harder

Three structural factors made sharia P2P harder than the original models priced in.

The sharia trust premium cuts both ways. Sharia products carry an implicit trust premium — borrowers and lenders engage on the assumption that the religious compliance of the product means the platform itself is operating to a higher governance standard. When the platform delivers on that assumption, the trust premium produces stickier engagement. When the platform fails to deliver, the trust premium reverses sharply. The sense of broken religious commitment compounds the sense of broken commercial commitment. Recovery is harder than for conventional platform failures.

DSN-MUI oversight capacity. The number of qualified sharia scholars available to review fintech products is finite. The volume of sharia-compliant fintech products being developed has grown faster than the scholarly review capacity. The result is review lag, occasional approvals that the underlying scholars might revisit with more time, and Sharia Supervisory Board arrangements at the institution level that sometimes operate more nominally than the regulatory framework imagines.

Property-backed sharia financing complexity. Several Indonesian sharia P2P platforms, including Dana Syariah Indonesia, concentrated on property-backed financing structures (murabahah, ijarah, musharakah mutanaqisah variants). Property-backed financing introduces collateral valuation risk, illiquidity, and concentration risk that sharia compliance does not mitigate. The product structures are religiously valid. The underlying credit risk profile is not transformed by the compliance overlay.

Lessons learned

Six principles emerge from the Indonesian sharia P2P cycle.

Religious compliance is not credit risk management. Sharia-compliant structures address the religious permissibility of the transaction. They do not address default risk, collateral risk, fraud risk, or operational risk. Platforms that treated sharia compliance as a substitute for conventional credit risk discipline mispriced their exposure systematically.

The sharia trust premium is a balance sheet item. The trust that comes with sharia branding is real capital — measurable in retention, repayment behaviour, and acquisition cost. It is also fragile in ways the platform’s operations need to actively protect. Platforms that treated sharia trust as a marketing message rather than as an operational discipline eroded it faster than they built it.

DPS authority must be real, not nominal. The Sharia Supervisory Boards at the institution level were designed to provide ongoing review of operational practice, not just product approval at launch. Platforms whose DPS reviewed only at the points the regulator required, rather than at the points the business actually needed scrutiny, accumulated compliance risk over time.

Also Read: The working capital trap: Lessons from Indonesia’s first D2C generation

Property collateral is not solvency infrastructure. Property-backed sharia financing carries the same valuation, illiquidity, and concentration risks as conventional property lending — sometimes more, because sharia structures often require longer holding periods before collateral can be enforced. Platforms that priced property as a low-risk asset class because the product was sharia-compliant mispriced one of their largest exposures.

Religious community contagion is faster than market contagion. When a sharia platform fails, the reputational damage spreads through religious community networks that operate independently of mainstream business media. The platform’s failure becomes a topic in religious discussion forums, mosque communities, and Islamic finance education networks. The recovery process must address that community dimension, not just the market dimension.

The next chapter requires institutional infrastructure. Sharia P2P will return — the underlying demand for sharia-compliant credit access is durable. The next generation will need stronger institutional underpinnings: more robust DSN-MUI oversight, more independent DPS arrangements, more sophisticated underwriting that integrates sharia compliance with conventional credit discipline.

The macro stakes

Indonesia is the largest Islamic finance market in the world by population. The sharia fintech opportunity is real and not going away. The first generation of sharia P2P platforms — Dana Syariah Indonesia among them — built something Indonesia’s Muslim-majority population had not had before: digital access to sharia-compliant credit products at scale.

Some of that infrastructure survives. Much of it does not. The lessons from the first generation are now the foundation for whatever the next chapter looks like. For founders, investors, and regulators thinking about sharia fintech in 2026 and beyond, the question is no longer whether the opportunity exists. It is whether the institutional infrastructure — scholarly review capacity, DPS authority, integrated risk management, community trust stewardship — has caught up with the scale the first generation tried to operate at.

—

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 sharia trust premium: Lessons from Indonesia’s first sharia P2P generation appeared first on e27.

The October 28 crypto session clearly shows how two forces can shape prices at the same time. Bitcoin gained 0.59 per cent to US$80,576.90 over 24 hours. The total crypto market cap rose 0.64 per cent to US$2.7 trillion. My view is that the headline numbers hide an important split. Bitcoin moved with traditional markets ahead of the Federal Reserve decision.

The wider crypto space also received a specific lift from Ripple’s institutional partnerships. That split matters more than the small daily gain because it shows Bitcoin trading as a macro instrument while select altcoins begin to respond to their own fundamental news.

The primary force behind Bitcoin was a market-wide beta lift. The largest digital asset rose in sync with equities and gold as traders positioned themselves before the Fed announcement. Bitcoin showed a 91 per cent correlation with the S&P 500 and a 92 per cent correlation with gold. The Nasdaq jumped 1.2 per cent as U.S. stocks rallied into the decision. The SPY correlation stood at 0.9074 and the gold correlation at 0.92271. This behaviour points to a rate-sensitive asset responding to interest rate expectations and liquidity conditions.

The data showed no clear coin-specific trigger. Spot volume fell 15.6 per cent and failed to confirm a strong bullish breakout. Derivatives activity stayed muted, with funding rates positive but low. Twenty-four-hour liquidations decreased to US$32.84 million. I read this as portfolio adjustment rather than aggressive long positioning in Bitcoin. That makes the gain fragile. A hawkish Fed surprise could erase it quickly.

The wider market found a different primary force. Ripple announced two developments on October 27 that changed sentiment for XRP and its new stablecoin. BNY Mellon added XRP and Ripple’s RLUSD stablecoin to its custody platform. Robinhood entered a strategic partnership to use RLUSD for institutional trade settlement and treasury operations. These steps directly address institutional custody and liquidity needs. They validate Ripple’s utility in traditional finance and lift sentiment across assets in its orbit.

BNB also performed well. BNB Standard Corporation trades on the Nasdaq Capital Market under the ticker symbol BNC and provides public market equity exposure to the BNB cryptocurrency. The Altcoin Season Index rose 5.56 per cent to 57, which signals a nascent rotation of capital from Bitcoin into select altcoins. Bitcoin dominance stood at 60 per cent. This is a meaningful shift. While Bitcoin mirrored the S&P 500 and gold, another development pulled capital toward assets with a clear institutional story.

Also Read: 3 scenarios for October 28: Which one breaks Bitcoin’s US$80,000 resistance

A supportive macro backdrop helped both stories. The S&P 500 gained 0.6 per cent, and the Nasdaq gained 1.2 per cent. The wider crypto space showed a 91 per cent correlation with the S&P 500 and a 93 per cent correlation with gold. Those links suggest a shared liquidity move rather than isolated crypto strength.

My perspective is that this combination is healthy. Bitcoin has become a short-term macro proxy. That role brings institutional legitimacy and liquidity. It also imports volatility from traditional markets. Altcoins now have room to trade on their own news flow. The Ripple partnerships involve named institutions with real custody and settlement functions. That kind of development can withstand a Fed-induced volatility spike better than a pure-beta trade.

The near-term Bitcoin picture is balanced but conditional. Bitcoin is testing the lower bound of its recent range. Key support sits between US$79,380 and US$80,000. Resistance rests near US$81,800 to US$82,500. The 50-day exponential moving average sits near US$81,873. Significant sell-side liquidity rests in the US$82,300 to US$82,500 zone.

A daily close above US$81,800 would signal bullish momentum. A break below US$79,380 would open a path toward US$77,500. The Fed policy statement and press conference later today, October 28, will decide the next directional move. After the announcement, Bitcoin needs to reclaim and hold above its 50-day exponential moving average. Without that step, the gain remains a macro bounce rather than independent strength.

Also Read: Bitcoin’s macro reset: Reclaim US$80,000 before the Fed speaks?

The wider crypto valuation faces its own technical test. The market cap is testing the 78.6 per cent Fibonacci retracement level at US$2.73 trillion, which now acts as near-term support. Holding above that level could set the stage for a retest of the US$2.81 trillion level, the 50 per cent Fibonacci mark.

A hawkish Fed tone or a break below support could reverse gains and lead to a test of the swing low at US$2.68 trillion. The Fed decision arrives on October 28 at 18:00 UTC. The key question is whether momentum from Ripple’s institutional breakthroughs can withstand volatility from the Fed policy signal. I will watch sustained on-chain activity for RLUSD and XRP. I will also watch whether other major custodians follow BNY Mellon’s lead. Bitcoin dominance at 60 per cent remains another useful gauge of rotation.

My overall view is cautiously optimistic, but the reasons differ from the headline numbers. Bitcoin’s gain reflects a macro-led bounce with the broader market, not independent strength. That makes it vulnerable to a single-event risk. The Ripple development provides a sector-specific trigger that could sustain altcoin outperformance even if Bitcoin stumbles. The 92 to 93 per cent gold correlation across both stories suggests crypto currently trades as a rate-sensitive asset class. That link brings liquidity and institutional legitimacy. It also brings imported volatility.

The smart approach for the next session is to watch two questions at the same time. Can Bitcoin reclaim and hold above US$81,873 after the Fed announcement? Does the Ripple story generate sustained on-chain activity for RLUSD and XRP? If both answers are yes, the market can build on this session. If Bitcoin breaks below US$79,380 while altcoin flows remain strong, the split between macro beta and fundamental triggers will define the week.

—

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 Fed decides today: Here are the 2 levels every Bitcoin trader should watch appeared first on e27.

When Meta switched on Muse in the US last month, it let its new personal AI agent wander into Amazon.com unannounced. Twelve days later, Amazon showed it the door, complaining that the bot did not identify itself and that nobody had asked permission.

In Singapore, Meta has brought the hosts along.

Also Read: Zuckerberg’s Muse Charm doesn’t care about your ad budget

Muse, which Meta pitches as an agent that does the work rather than merely answering questions, went live in the city-state today on iOS, Android and muse.ai. It arrives with eight local launch partners, or “connectors” in Meta-speak: Shopee, Grab, Singapore Airlines, Agoda, ShopBack, Klook, Trip.com and Lazada. Together they span shopping, groceries, dining and travel, which covers most of what a Singaporean does on a phone between meetings.

What Muse can actually do here

Connectors let the agent act inside services people already use. Beyond the eight local partners, Singapore users can plug Muse into Gmail, Google Calendar and Instagram to manage email, juggle schedules and draw on saved posts.

For instance, ask Muse to pick up a pack of newborn nappies and the ingredients for fried rice, and it asks clarifying questions, builds a GrabMart basket, checks the total and delivery time, then waits for your approval. Ask for the cheapest return flight to Seoul, and ShopBack compares fares across booking sites after cashback. Klook’s pitch involves friends in town for December’s BTS concert, with Muse suggesting and booking activities in the same conversation.

The fine print matters, though. Meta says connector features will roll out progressively, with some experiences arriving only after launch, and it does not say which ones. Day one may be thinner than the partner list suggests.

Muse is free for most uses, with paid tiers for heavier users. In the US, those tiers reportedly cost US$20 and US$100 a month; Meta has not disclosed Singapore pricing. A version for AI glasses is “coming soon”.

Why the platforms said yes

The partner quotes are mostly what you would expect. Grab’s chief product officer Philipp Kandal talks about linking conversational AI with “real-world action”. Agoda’s Matteo Frigerio claims his company is the first digital travel platform globally to connect its own AI travel assistant to Muse.

The most candid line comes from ShopBack CPO Justin Lee. Agents get things done, he argues, but the web was built for people to read rather than for agents to act on, hence the need for connectors. It is the same problem startups such as Singapore-founded Lightsage are raising money to decode.

Also Read: Agentic commerce’s dirty secret: the data powering AI purchases is often wrong

Read that again from ShopBack’s side. A cashback business lives in the gap between shopper and merchant. An agent that can compare prices on its own threatens to route around such middlemen entirely. Becoming a launch connector is partly insurance: if the agent is going to shop, better to be the tool it shops with.

Shopee and Lazada face a sharper version of the same trade. Amazon blocked Muse partly because, it said, an agent shopping for a customer bypasses the personalisation built into its store. For SEA marketplaces whose economics lean on in-app discovery and sponsored listings, handing the browsing to Meta’s agent is no small concession. Shopee has shown a willingness to open its storefront to outside discovery, most recently by teaming up with Pinterest in Indonesia. Muse goes a step further: the customer may never scroll the app at all.

Neither Meta nor its partners have said whether money changes hands, who keeps the customer data, or how the agent decides which marketplace wins a query. Those answers will decide whether this is a partnership or a slow handover of the shop window, and they sit at the heart of the agentic commerce trust gap nobody has yet fixed.

The safety pitch, and the asterisks

Meta says most of its engineering effort went into safety. Each Muse runs on Muse Secure VM, a dedicated cloud computer that stores credentials for connected services while keeping passwords and payment methods out of the agent’s view. Users choose what Muse can touch, and it checks in before sensitive actions such as sending an email or making a purchase.

That is the right architecture on paper. In practice, Meta delayed the US launch to work on security after mixed internal testing, and The Information reported that the agent sent emails without approval during those tests. Amazon also alleged Muse appeared to capture and store customer credentials, which Meta disputes.

Singapore is a demanding test bed. Local firms lead APAC in deploying AI agents, but also in rolling them back, and payments players are already debating “Know Your Agent” checks alongside KYC. An agent that buys the wrong nappies is an annoyance. One that books the wrong non-refundable flight is a customer-service crisis.

A crowded race for your errands

Meta is hardly alone. OpenAI has Operator, Google has rolled out Gemini Spark to work in the background, Perplexity has its Comet shopping agent, and Amazon is building its own buying assistant into Alexa. Muse’s edge is reach: Meta already owns WhatsApp, Instagram and Facebook, and a week after its US debut Muse was the top free app on Apple’s US App Store, ahead of ChatGPT.

Also Read: When AI agents start acting on our behalf, security gets more complicated

Southeast Asia is also where Meta is leaning hardest into chat commerce, having appointed Dhruv Vohra to lead its regional business as AI reshapes online retail. Launching here with signed partners rather than uninvited browsing suggests Meta learned something from its Amazon spat.

“Muse is your personal agent, built to help make everyday life easier,” said Nicole Tan, Managing Director of Meta Singapore.

Whether easier for users also means better for Shopee, Grab and the rest is the question the press release leaves out. For now, Muse has been invited in. Staying welcome is the harder part.

The post Meta’s Muse lands in Singapore with a guest list Amazon never got appeared first on e27.

For years, the anti-phishing playbook has been drilled into office workers like a fire-safety briefing: check the sender, hover before you click, never type your password into a page you reached from an email. The latest data from Check Point Research suggests attackers have read the same briefing. Increasingly, they are not after your password at all. They want you to open a terminal and paste in their code yourself.

The Nasdaq-listed cybersecurity company’s Brand Phishing Ranking for Q3 2026, released on Wednesday, shows Microsoft remained the most impersonated brand, accounting for 33 per cent of all brand phishing attempts tracked in the quarter, up from 23 per cent in Q2. Apple climbed two places to second with 8 per cent, level with Google in third. LinkedIn slipped two places to fourth at 6 per cent, and Amazon rounded out the top five at 4 per cent.

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

Those five names account for more than half of all impersonation attempts, and Microsoft’s share alone is roughly four times that of any other brand. The logic is not hard to follow. A single Microsoft 365 login unlocks email, cloud storage, Teams chats and often the company’s single sign-on, so one stolen credential works as a skeleton key to the whole office.

What the release does not explain is why Microsoft’s share jumped ten percentage points in a single quarter, or how many attempts the ranking is built on. Without that denominator, the percentages are best read as a direction of travel rather than a census.

From stealing passwords to borrowing your hands

The more interesting part of the report sits below the league table. In Q3, Check Point says, attackers followed users into AI and developer tools, and several campaigns stopped bothering with credential theft altogether.

One impersonated Claude, Anthropic’s AI assistant, with a fake payment-failure email claiming a EUR500 (about US$580) charge had been declined. The message warned that API access and Claude seats would be paused unless the recipient updated their payment details, a lure aimed squarely at whoever holds both the company card and the API keys.

Another used a lookalike website for Google Antigravity, Google’s AI-powered coding environment, which displayed what looked like a routine Linux installation command. The command actually decoded a Base64-encoded instruction (Base64 turns text into an unreadable string of characters, which conveniently makes malicious code look like harmless gibberish) that downloaded and ran a shell script from an external server. The result was remote code execution: the attacker could run commands on the victim’s machine at will.

A fake Apple support article coaxed Mac users into running terminal commands with the promise of better performance. And a spoofed Docusign page told visitors a reader update was needed before they could view a document. Clicking Continue downloaded FleetDeck, a legitimate remote monitoring and management tool that IT teams use to administer devices, and which an intruder can use just as easily.

None of these campaigns exploits a software flaw. The victim does the installing. That is the point: security software trained to flag malware has little reason to stop a genuine remote-access tool, and a developer who pastes an install command from a web page is doing something the industry has spent years normalising.

Also Read: Southeast Asian SMEs remain soft targets as ransomware groups refine extortion tactics

“Brand phishing is entering a new phase,” said Omer Dembinsky, Data Research Manager at Check Point Research. Attackers, he added, are “moving beyond password theft to convince victims to run malicious commands or install remote access software themselves.”

Why this lands differently in Southeast Asia

The shift matters for a region whose startups run lean. In a 15-person company in Jakarta or Ho Chi Minh City, there is often no finance team between a billing alert and the person who pays it, and the engineer installing a new AI coding tool may also be the only one who would notice it misbehaving. Speed and informality, the traits founders prize, are exactly what these lures exploit.

Regional data points the same way. A Mimecast study published in July found 65 per cent of IT and security decision-makers surveyed in Singapore and Australia expected an AI-enabled attack on their organisation within 12 months, while 60 per cent said they were not fully prepared for AI-driven threats that exploit human weaknesses. Lean SMEs, often without a dedicated security hire, sit at the soft end of that gap.

Policymakers are busy, but mostly on the consumer front. Singapore has tightened scam rules for messaging, social media and e-commerce platforms, wants Google’s AI to flag scammers before victims transfer money, and its police worked with Meta to disrupt 3.7 million scam-linked assets on Facebook and Instagram, after authorities earlier knocked out 30,000 iMessage accounts tied to scams.

Those efforts target the channels where ordinary people lose their savings. A fake API bill in a developer’s inbox, or a poisoned install command on a cloned product page, falls largely outside them and lands on companies’ own defences instead. Add the rise of deepfake fraud, and the weak point looks less like the software than the moment a busy person decides to trust a familiar logo.

The fix, according to the firm that sells it

Check Point’s prescription is to stop phishing before it reaches inboxes, use AI-powered detection for brand impersonation and credential harvesting, automate investigation and response, monitor for unapproved remote management tools, and scan the wider web for lookalike domains and cloned login pages so they can be taken down. All sensible, and all, unsurprisingly, categories Check Point sells into.

Also Read: Ransomware in Singapore is becoming a human problem, not just a malware one

For smaller teams without an enterprise security budget, some of the most effective controls cost nothing. Billing problems can be handled only by logging into the vendor’s console directly, never through an emailed link. Developer tools can be installed only from documentation reached by typing the address yourself. And a short list of approved remote-access software makes anything else stand out.

Generative AI, Dembinsky warned, is making convincing emails and cloned websites cheap to produce at scale. The uncomfortable corollary is that the next breach at a regional startup may not need to break anything. It will just need someone to copy, paste and press Enter.

The post Phishers now want you to type the malware in yourself appeared first on e27.

Open 10 startup websites. Cover the logos. Shuffle the pages. Nothing breaks. “Revolutionising how finance works.” “Trusted by industry leaders.” Nobody asks which leaders.

The usual diagnosis is lazy copywriting. It lets the founder off. The real cause is worse: she did everything right.

Imagine a Series A software company. Watch one sentence travel through its second year.

The founder writes it in an afternoon: We connect community banks to their old core systems, so they never have to replace them.

Marketing arrives, because the product needs customers: We help financial institutions modernise.

PR arrives, because the company needs news: The fintech infrastructure company reshaping banking.

Investor relations arrives, because the Series B needs a story: A platform for the global financial industry.

An agency polishes the lot: Powering the future of finance.

Every edit was defensible. Every hire was rational. Every version sounded more professional than the last.

Now look at what went missing. Community. Old. Never have to replace them. Those three phrases were the strategy: a narrow customer, an unglamorous problem, a niche larger rivals might ignore. Nobody deleted them on purpose. Each department simply preferred a bigger sentence.

Call it compound dilution. Nobody lies. The judgement leaks anyway.

No one is paid to decide which customers the company will refuse, which markets it will skip, which promises it will not make. That job belongs to the CEO, who is approving budgets and sitting through board meetings.

Also Read: Male founders are losing confidence. Women are doubling down on AI

The company has acquired professional management. It has lost the person who decides what it is not.

And the damage does not stay in the copy. In this imagined firm, sales reads “financial industry” and starts calling insurers. Engineers, being obliging, build what the insurers ask for. The roadmap bends toward the market investor relations described, not the market that pays. Each choice carries a business case. The sum is a company that traded a market it knew how to win for one where it has no particular edge, without a single meeting at which anyone decided to.

The company stops describing its business and starts building the business it has described.

Silicon Valley has circled this problem since Paul Graham published Founder Mode in September 2024. Drawing on Airbnb’s Brian Chesky, he attacked the advice to hire good executives and leave them alone. The argument soon became a quarrel over micromanagement, and both camps missed the odder risk. A company needs no incompetent executive to lose its way. Three excellent ones, each optimising a different version of the firm, will do it.

A tighter finding sits in a 2025 Harvard Business Review study by Scott Cook, Intuit’s co-founder, and Nitin Nohria, a former Harvard Business School dean. They profiled leaders at Amazon, Danaher, RELX and Toyota. The ones who impressed them did not simply make every decision. They designed how work gets done and how decisions get made.

A CEO who delegates the design of judgement has delegated the company.

That does not make micromanagement the answer. A founder who approves every press release becomes the bottleneck, and specialists earn their salaries by challenging her assumptions. But challenging a strategy and quietly replacing it are different acts.

Who decides whether a new customer segment strengthens the business or distracts it? Whether a bigger market is worth a weaker position? Whether the investor narrative matches the product customers actually buy?

These are not communications decisions. They decide where the next dollar goes.

Also Read: Why con artists get the meeting that honest founders can’t

Investors meet the consequences across a table. Sequoia’s pitch guidance opens by asking for the company in one declarative sentence, then moves to the customer’s problem, why now, the competition, the business model. The sentence is the compressed record of choices already made.

A founder who stumbles in the first minute may still have a good company. Brilliant engineers present badly. Smooth talkers build nothing. But when a polished answer collapses at the first question about the customer, a careful investor should wonder who decided what it says.

The temptation is real for Asian start-ups that raise from Singapore, sell to Japanese corporates and recruit distributors in Indonesia within a single quarter. Each audience deserves its own emphasis. None deserves its own company.

The founder’s job is not to write every sentence. It is to make sure each one traces back to a decision she is prepared to defend. Otherwise professionalisation becomes abdication: more people responsible for communicating the strategy, fewer for keeping it intact.

Picture the next fundraising meeting. The marketing deck says the company sells to everyone. The IR deck says the market is enormous. The agency has made both look convincing.

The investor closes the laptop and turns to the founder.

Which one is the company actually building?

—

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 Every department you add makes your startup harder to understand appeared first on e27.

Ask a typical AI assistant a question and it answers. Amit Suxena thinks that is already yesterday’s AI.

“You don’t need AI where I ask a question and get an answer. I think that’s already passed,” says the GVP and Head of NetSuite Asia, who oversees Southeast Asia, the Indian subcontinent and Greater China for Oracle’s cloud ERP business. “I want AI to give me suggestions before I ask.”

His example is deliberately mundane. An AI notices that a distributor’s average inventory cycle is 22 days, works out that a tweak could bring it down to 17, and tells the owner what those five days are worth in cash. The owner, happy with 22 days because customers tolerate 30, would never have thought to ask.

Also Read: Language was never the problem: Inside SEA’s real AI adoption gap

That pitch, proactive and boring in the best sense, sits at the heart of how NetSuite is trying to sell AI to the region’s growing businesses. It also builds on the argument Suxena and NetSuite founder Evan Goldberg made to e27 at SuiteWorld in August: context, not automation, is the real prize.

Thailand, the surprise package

Asked which markets have surprised him, Suxena does not reach for the obvious. Singapore leads, he concedes, largely because labour is expensive and AI lets a small team do more. China has been among the fastest adopters, though he says widespread adoption there has been “up and down”.

The real surprise is Thailand. “Thailand actually stands out as a country where the founders and organisations have been ahead,” he says.

His explanation is supply chains. Thailand hosts a dense cluster of multinationals, many of them Japanese manufacturers, and the SMEs that supply them are pushed to adopt the latest tools to keep up. The country is also courting AI more deliberately, from an OpenAI-backed startup accelerator to a US$80 billion semiconductor push.

Malaysia and Indonesia, he says, still have catching up to do. Indonesia is the bigger prize: a sprawling archipelago of SMEs and mid-market firms operating under different regional rules. Vietnam, meanwhile, still leans heavily on traditional workforces to deliver output. “Big potential,” he says, diplomatically.

Fear, not cost, held SMEs back

So what has slowed the laggards? Suxena’s answer is fear of implementation more than data privacy. The small business owner’s reflex, he says, is: “I have a small business. Can I actually implement AI? It’s too complicated.”

NetSuite’s response is speed. Suxena points to its accelerated implementation programme, which he says can get a company live in a matter of weeks, along with prebuilt industry dashboards and AI that answers questions in plain language. “Suddenly the fear is taken away,” he says.

The deeper problem, he argues, is fragmentation. SMEs buy best-of-breed tools for each job: one for supply chain, another for finance, a third for CRM. Bolt a third-party AI onto that patchwork and it works blind.

He sketches a leather manufacturer shipping to a US customer who expects delivery in 18 days. An AI optimising logistics alone would rush the order. It would not know the customer pays late, or that another buyer who pays reliably deserves priority, even at an extra four per cent discount. “Everything that AI does is out of context,” he says.

It is a self-serving argument for a vendor selling an all-in-one suite. It is also one e27 keeps hearing: Southeast Asia’s SMEs have a record problem, not an AI problem, and even in Singapore, messy data keeps AI pilots stuck.

No one asks AI what business to be in

Practical AI has become every vendor’s favourite phrase, so what have customers asked for that was not practical? Suxena insists the requests have been grounded. “Nobody has asked us what business should I get into, or should I leave this business,” he says.

Instead, they ask about cash flow. One recurring question is what a new market or product line would do to the books. Another is whether a marketing agency’s sales forecast matched reality; NetSuite’s dashboards now compare the prediction with the outcome.

Also Read: Who referees the AI agents? Enterprises aren’t sure, new report shows

Currency volatility is the region’s chronic headache. Suxena is careful not to overclaim. NetSuite is not a currency trading platform, and “it’s impossible for anyone to predict where currencies are going,” he says. “The only certainty is that the uncertainty is there to stay.”

What the software offers instead is historical correlation: what a firm bought from country A, how long it took to arrive, and how currencies moved across markets C and D. The newer twist is that the AI surfaces those patterns before anyone asks.

What ASEAN can borrow from China

Suxena, who stresses he is not a China expert, sees one transferable lesson in how Chinese firms expand. Many run a two-tier ERP model: a heavy enterprise system at headquarters, and a lighter, multi-currency, multi-jurisdiction system such as NetSuite for overseas subsidiaries.

The point, he says, is that Chinese companies were not constrained by their domestic software when they went abroad. For Southeast Asian founders eyeing regional expansion, the lesson is to pick tools for the markets you are entering, not the one you started in.

Compliance and the hallucination problem

India’s DPDP Act, Indonesia’s PDP law and China’s data export rules all restrict where business data can travel. Suxena’s answer on compliance is reassuring in tone but thin on detail. NetSuite has a product team tracking regulatory changes in each country, he says, and customers are “also expected” to check what they are using is compliant.

His stronger card is that NetSuite’s AI is built in and runs on Oracle’s own cloud infrastructure, so analysis stays inside the customer’s context rather than being handed to someone else. Founders connecting external models through NetSuite’s newer AI connectors may want a sharper answer than that, given how Singapore firms fear data sovereignty failures but remain underprepared.

On hallucinations, particularly agents that go rogue with payments, his mantra is human control. Users decide which tasks an agent can execute alone and which need a yes or no.

He offers a telling example. An AI’s idea of a “bad paying customer” may differ from the owner’s, so before automated payment reminders go out, the owner can review the list. “Being in control is the mantra,” he says. That is sensible, but it also places the burden of catching errors back on the customer.

Rivals, and the car with a free engine

NetSuite is hardly alone. Zoho, Kingdee and a crowd of regional accounting and ERP players are adding AI to cheaper suites, while SAP Business One and Microsoft Dynamics 365 Business Central chase the same mid-market. Suxena declined to discuss specific competitors.

Instead, he set out three tests. First, does the AI see the whole business or just one function? Second, does the vendor own its infrastructure as well as its software? Third, how fast can it be deployed? “You want something which works for you today,” he says, “not something which takes a year to implement.”

The infrastructure test is the shakiest. Microsoft, for one, pairs Dynamics with its own Azure cloud, so Oracle is not unique there.

On price, Suxena is emphatic that AI is not a back door to higher licence fees. Everything Goldberg announced in the keynote a day earlier, he says, is included in the existing subscription. He borrows Goldberg’s line: “When you buy a car, you don’t expect to pay separately for the engine.” NetSuite is also offering a startup version that, he says, costs nothing at the outset, betting customers will scale with it, as some have from under US$100,000 in revenue to multi-billion-dollar businesses.

The junior accountant question

Finance back-office work is a major employer in the Philippines and India. What happens to the junior accountant when AI handles reconciliation and exceptions?

Suxena reaches back to NetSuite’s early days as one of the first cloud companies, when customers feared the cloud would hamper them. Those early adopters, he says, are now far ahead and probably did not cut staff. Junior accountants, he argues, will shed repetitive data-entry work and reskill into more proactive roles.

It is the industry’s standard reassurance, and it sidesteps the harder question of who pays for that reskilling. e27 has argued that agentic AI should worry the people it replaces, and the evidence from shared-services hubs is not yet in.

How the boss uses it

Suxena’s own favourite use is telling. He asked NetSuite’s dashboard which country in his region was most productive and profitable, and where he should add headcount. The AI weighed staff numbers, sales, renewals (“extremely important” for a software company), local regulations and more.

Also Read: SMEs adopt AI four times slower than big business, the fix starts with the PC

The answer surprised him. He would not name the country. “I would have never thought that particular country is the one where I should be investing,” he says.

What has he stopped doing? Digging through forecasts line by line. He now simply asks what the six-month forecast looks like and what risks it carries.

For Southeast Asia’s founders, that may be the most practical lesson of all: the value is less in AI that answers, and more in AI that notices what you never thought to ask, provided the data underneath is in order.

Quotes have been lightly edited for clarity and length.

The post Thailand, not just Singapore, is Asia’s quiet AI surprise, says NetSuite’s Amit Suxena appeared first on e27.