Ask a Singapore founder who might buy their company and you usually get three names: the biggest local competitor, a supplier, maybe a customer. It is a sensible guess. On the data, it is also usually the wrong pond.

At Ayebot we keep a register of potential acquirers for Southeast Asian companies. As at 19 October 2026 it maps 2,488 of them: 1,662 trade buyers, 413 private equity firms, and family offices and other investors. Of the buyers whose headquarters we can identify, 82 per cent are based outside Singapore, across 32 countries. 214 are headquartered in Japan alone.

Three things in the data should change how founders plan.

• Your sector decides who is at the table. In building services, our register shows 194 trade buyers covering Singapore against 12 private equity firms. In food and beverage, 68 against five. Healthcare is the near tie: 83 trade buyers, 70 private equity firms. Education is the only sector where private equity leads, eight to seven. A sale process designed for a fund looks very different from one designed for a strategic buyer, so knowing the mix early is not trivia; it is strategy.
• Most buyers arrive through a regional mandate. Take healthcare in Malaysia and the Philippines: on our register, more buyers cover those markets through a Southeast Asia, Asia or global remit than by naming the country. A buyer search that only looks for firms naming your country can miss much of the room.
• Size is the sharpest filter. Buyers state the size of business they want. A company with SG$3m (US$2.3m) of revenue and one with SG$30m (US$23.4m) are, for most acquirers, different products. Screening on size before approaching anyone saves months and protects confidentiality.

Also Read: The frontier AI money is getting picky. Southeast Asia’s AI founders should price that in now

The table founders should look at first. Here is how the buyers covering Singapore split by sector on our register, as at 19 October 2026:

• Industrial and engineering: 238 acquirers, of which 184 trade buyers and 43 private equity firms.
• Building services and infrastructure: 222 acquirers; 194 trade, 12 private equity.
• Healthcare and medical: 203 acquirers; 83 trade, 70 private equity.
• Consumer brands and retail: 190 acquirers; 82 trade, 60 private equity.
• Technology and software: 147 acquirers; 83 trade, 36 private equity.
• Logistics and supply chain: 88 acquirers; 73 trade, eight private equity.
• Food and beverage: 81 acquirers; 68 trade, five private equity.
• Beauty, aesthetics and wellness: 69 acquirers; 49 trade, one private equity.

Two things jump out. First, in most sectors the natural buyer is another operating company, not a fund. That matters because a trade buyer usually pays for what your business adds to theirs: customers, licences, people, a foothold in a new market. A fund pays for the earnings it can grow and later sell on. The same company can be pitched very differently to each, and the founder who knows which one is across the table negotiates better.

Second, Japan shows up in almost every sector. 214 of the acquirers on our register are headquartered in Japan, including 44 covering Singapore industrial and engineering firms and 37 in building services. For many Singapore founders, Japanese corporates are the buyers they have never met but should. 343 of the acquirers on the register are listed companies, which brings its own rhythm: board approvals, disclosure rules and longer timelines.

Also Read: Visa fees are soaring, but founders want US revenue more than US investors

What AI changes, and what it does not. Finding these buyers used to take an analyst weeks of databases and annual reports. Today an engine can screen thousands of acquirers by sector, market and size in seconds; that is what we built Ayebot to do. What AI cannot do is decide which five deserve a confidential call, read the room in a negotiation, or tell a founder whether a buyer will look after the team. Those remain human jobs, and they are where most of the value of an exit is won or lost.

What to do with this if you might sell in the next few years:

• Count your buyers before you guess them. Sector, market, size.
• Find out whether your likely buyers are trade or financial; they value different things.
• Start preparing two to three years ahead: clean numbers, a team that runs without you, contracts that transfer.

—

The post Your buyer is probably not Singaporean: What 2,488 acquirers tell founders planning an exit appeared first on e27.

When Singapore arranged exclusive Southeast Asian dates for Taylor Swift’s Eras Tour in March 2024, six shows, only stop in the region, the diplomatic and economic discussion that followed was one of the more publicly visible moments of ASEAN economic competition in recent memory. Thai officials publicly questioned the arrangement. The Philippines registered concern. Indonesian commentary ranged from frustration to grudging admiration.

The reported terms, Singapore offering a grant of several million dollars per show in exchange for regional exclusivity, produced a public debate framed mostly as a competition story. What it actually revealed was a question about how ASEAN governments think about cultural and entertainment economic assets, and what the next decade of cultural-economic competition in the region might look like.

After 15 years inside Indonesian risk functions, I found the Singapore deal more interesting as a risk-management decision than as a diplomatic moment.

What Singapore actually did

The Singapore arrangement was, in risk-management terms, a specific kind of bet. Singapore committed a defined, disclosed cost, grant payment plus infrastructure, security, and hosting expenses, in exchange for asymmetric upside in tourism revenue, regional brand positioning, and economic activity concentrated in a single tour window.

Independent estimates suggested the six concerts produced inbound revenue several multiples of the grant cost. The bet was bounded, the cost was known, and the upside was probabilistically larger than the downside. That is what a risk-managed economic move looks like.

Also Read: Singapore tops a new Commonwealth AI readiness index, but Australia’s results show where the real gap lies

Why the approach worked

Three structural features made the Singapore deal defensible as strategy rather than as vanity expense.

The asymmetric payoff was real. The economic impact of major artist tour stops has been studied across markets for two decades. Cities that win exclusive or anchor stops capture meaningfully more revenue than the artist payment, especially when the catchment region is large.

The infrastructure was already in place. Singapore had invested in venue, transport, hospitality capacity, and event-hosting institutional capability over decades. The marginal cost of one more major event was substantially lower than for cities that would have needed to build infrastructure to host comparably.

The reputational asset compounded. Hosting the only Southeast Asian stops of one of the most discussed cultural events of the decade positioned Singapore globally as the regional event destination. That positioning is itself a long-term economic asset.

Lessons learned

Six principles emerge from the Singapore Eras Tour episode for other ASEAN cities and governments.

• Cultural events are infrastructure decisions, not entertainment decisions. The cities that win these deals consistently are the ones that have built event-hosting infrastructure, venues, transport, accommodation, public safety, payment systems, to a standard that makes them the default choice. The investment must be in place before the deal is on the table.
• Speed of decision-making is competitive advantage. Singapore’s ability to commit quickly, with the financial authority and institutional alignment to follow through, was as important as the financial terms. Cities that need months of inter-agency coordination lose to cities that can decide in days.
• Public-private deal-making capacity is real economic infrastructure. Cities that can structure, evaluate, and execute deals with global cultural actors capture asymmetric value. The capacity is institutional, not transactional. It is built over years.
• Risk-managed expenditure looks different from political expenditure. A grant payment to attract a tour is not the same kind of public spending as a subsidy or transfer. Treating it as if it is produces political backlash that misses the underlying economics.
• The “winner-takes-most” dynamic in cultural events is structural. Just as in financial centres and technology ecosystems, the city that captures one major event becomes more likely to capture the next. Network effects in global cultural production are real.
• Coordinated ASEAN positioning is missing infrastructure. The current model treats each ASEAN city as competing independently for global cultural events. A coordinated approach, where ASEAN cities collectively bid for multi-stop tours rather than competing to be the single exclusive stop, would produce better outcomes for the region as a whole.

Also Read: Nine in ten Singaporeans use AI. Only a third are comfortable with it

The macro stakes

ASEAN’s economic competition over the next decade will not only be about manufacturing FDI, financial services hubs, and digital infrastructure. It will also be about who captures the cultural and entertainment economic assets that generate disproportionate global attention, tourism revenue, and brand positioning.

The Singapore Eras Tour episode was an early signal of what that competition looks like. The Swift premium will be replicated, in different forms, many times over the coming decade. The question for ASEAN governments is whether they are building the institutional capability to capture it, or whether they will continue to react to it when their neighbours do.

—

The post The Swift premium: What Singapore’s exclusive Eras Tour deal taught ASEAN about cultural economics appeared first on e27.

There is no spreadsheet for this. You cannot plug in a dozen parameters, hit enter and receive a verdict on whether a startup matters. That was the opening salvo at a Huddle Global 2026 panel provocatively titled “The VC Truth: Build What Matters, Not What Gets Funded”, where three Indian venture capitalists spent the better part of an hour conceding, with refreshing candour, that the two are not always the same thing.

Moderated by Zerin Rahiman, Director at IndigoEdge, the session brought together Madhukar Bhatia, Managing Partner at Pentathlon Ventures; Anil Joshi, Managing Partner at Unicorn India Ventures; and Aditya Agarwal, Founder and Managing Partner at Tetration Ventures. What emerged was less a defence of the venture model than a frank audit of its blind spots, and a reminder to founders that the industry’s cheques follow a very particular arithmetic.

Also Read: The new value test for investors: What founders are actually buying when they take your cheque

The panel agreed early that “what matters” shifts depending on who is asking, and when. Their advice was to stop thinking like an investor altogether. Start with the customer instead: what problem is being solved, who actually has it, how the solution differs from what is already on the market, and whether it can become viable and scale.

It is a mindset, not a gulf

Joshi argued that the distance between what matters and what gets funded is narrower than founders fear. The difference, he suggested, is largely one of mindset. A founder obsessed with how the product helps the customer tends to end up with a fundable business; one obsessed with impressing investors often ends up with neither.

That sounds tidy, so the moderator pushed harder: are there companies that genuinely matter but simply do not belong in a VC portfolio? Here the panellists got honest. One admitted to admiring certain companies from afar precisely because they do not fit his fund’s thesis, and occasionally writing a personal cheque instead. Another put it more bluntly. Sometimes the answer is as simple as a mandate: a fund that has made promises to its Limited Partners cannot wander off-thesis, however good a company looks. Founders who skip checking a fund’s focus before pitching are, in effect, interviewing the wrong investor.

There was also an unusual admission of fallibility. One investor said he would always rather be proven wrong, because a company he passed on that later succeeds is the industry’s best teacher. It is a rare sentiment on a stage where VCs are more used to dispensing advice than following it.

The taxi meter problem

The sharpest explanation of the gap came from the economics of the asset class itself. Venture capital, one panellist reminded the room, is considered the highest-risk form of investing, and so it carries the highest return expectations.

Take an IT services company. It runs like a taxi meter, generating revenue month after month, with that revenue steadily climbing. It is very likely to survive. But its growth is shackled to headcount, and it cannot compound the way venture funds need. Every venture bet must carry at least the possibility of returning 100 times the money; a services firm never will. It solves a real problem, it matters, and it is still the wrong fit. Investors, after all, fund predictability of a very specific kind.

Also Read: Growth at gunpoint: Why VCs share the blame for startup fraud

The same logic applies to founders who simply want to grow slowly and steadily. That is a perfectly legitimate way to build a company. It just is not a venture-backable one, and the panel did not pretend otherwise. Good ideas are everywhere; venture capital is not.

When a noble problem meets a broken supply chain

Sometimes a problem matters enormously and the business still buckles under diligence. One investor walked through a startup tackling stubble burning in Punjab, the seasonal crop-residue fires that choke northern India’s air every year. The company ran a biomass marketplace connecting farmers who generate agricultural residue with industrial buyers who could use it.

The investors were initially enthusiastic. But as they dug in, the real problem turned out not to be matchmaking at all. Residue has to be aggregated, transported and stored within a narrow seasonal window, and the marketplace had no answer to that physical bottleneck. It is a lesson agritech investors are learning the hard way: software cannot wish away logistics.

Backing the founder who has lived the problem

If there was one trait the panel prized above all others, it was depth. Investors want founders who have spent years inside a problem, through their studies or their careers, because time in a space builds both insight and the relationships needed to sell into it.

One example was a company in electronics manufacturing services (EMS), a segment widely written off as commoditised and hardly the classic investable startup. But its founders had long histories in the industry and were moving up the value chain, from pure contract manufacturing into designing products as well as making them. The fund backed them, and the bet has paid off well.

The sixty-something founder who outlasted a pandemic

The most striking story involved a healthtech investment made in 2019. The founder was in his sixties, far older than the typical VC-backed entrepreneur, and was building non-invasive screening technology for diseases linked to tobacco chewing and poor oral hygiene, a stubborn public health challenge in India.

The investor wrote the cheque. Within a month, COVID-19 arrived. For roughly a year and a half the investment looked all but dead; nobody was queuing for screenings, and no secondary buyer was going to touch it. Yet the founder’s sense of mission never wavered. Today, close to 70, he runs a company that turned profitable last year and is planning to set up its own manufacturing facility.

Mission does not have to mean social impact

The panellists drew one careful distinction. Mission, they argued, is not a synonym for social impact. Not every founder is solving an environmental or social problem, nor should they need to. What investors look for is conviction: founders who believe they are the only ones who can solve a particular problem, and who bring the experience and unique insight to back that belief.

That conviction matters because the journey is brutal. Some young founders start companies because entrepreneurship looks glamorous, one panellist observed, and the glamour evaporates the moment money runs short. Mission is what keeps a founder in the chair when it does.

Why India’s sports startups struggle to raise

An audience member brought the debate down to earth with a pointed question: sportstech startups in India rarely get funded, so is that because sport does not matter? The answer, once again, was market size. Indians largely consume sport on television rather than play it, which shrinks the addressable market investors model. Parents may spend generously on their children’s training, but VCs question whether that adds up to a large business within a decade. Founders usually understand their market better than investors do, the panel noted, and it is their job to make the case that it is big enough. The panel left the door ajar, though: AI’s arrival in sport could yet reshape that calculus.

Also Read: SEA’s venture capital shifts from mega-rounds to AI and SaaS

The session’s takeaway was uncomfortable but useful. Building what matters and building what gets funded can overlap, but only when founders understand the rules of the game they are entering. Sometimes what matters today only gets funded years later; one founder’s company took six or seven years to find its footing. And sometimes the smartest decision is not to take venture money at all.

The post Build what matters or what gets funded? India’s VCs say the real gap sits in the founder’s head appeared first on e27.

Jensen Huang has stepped down as president and chief executive of Nvidia. In a regulatory filing, the chipmaker said its board accepted his resignation on Wednesday, days after Huang admitted to an affair with Alice Huang, a former executive at Singapore-registered Megaspeed International whom US officials suspect of spying for China.

Chief financial officer Colette Kress will serve as interim CEO while keeping her finance role, and long-serving director Harvey Jones, a former Synopsys CEO, becomes Nvidia’s first-ever board chair. Huang has also agreed to testify before Congress by Thanksgiving. Investors have punished the uncertainty: Nvidia shares fell more than 11% in premarket trading on Thursday, extending a slide that has erased over US$300 billion in market value since Friday.

For Southeast Asia, the fallout lands close to home. Megaspeed says it is cooperating with US export-control inspections and has hired a US law firm to count the GPUs it holds, while Singapore police continue their own investigation. The episode puts the region’s GPU cloud operators, and the export rules that govern them, under a harsher spotlight just as Malaysia and Singapore court billions of dollars in AI data centre investment.

Editor.

---

REGIONAL

SoftBank, Grab commit US$1.85B to 5GW AI compute push in Malaysia:Petroleum Sarawak adds US$100M across two joint ventures in Sarawak and the peninsula. Cash arrives in milestone-linked stages, and neither venture has signed customers yet, leaving the five-gigawatt target a forecast for now.

Megaspeed says it is cooperating with US export-control inspections: The Singapore-registered Nvidia cloud partner, a former employer of Alice Huang, has hired a US law firm to count its GPUs and audit compliance. Singapore police have investigated the firm since October.

Sea’s Q3 revenue jumps 43.4% to US$8.6B as Shopee hits record highs: Net income rose 34% to US$502.5M. Shopee’s GMV grew 23.3% to US$39.7B, digital financial arm Monee lifted revenue 63.9% to US$1.6B, and Garena’s adjusted EBITDA climbed 45.9% as Free Fire bookings grew.

SEA AI funding more than doubles to US$4.2B in 9M 2026: Tracxn: Kling AI’s US$2.8B Series D accounts for 66% of the haul, and the ten largest rounds take 98%. Singapore-based firms captured all the capital, deal count fell 12% to 28, and no new unicorns emerged.

Airwallex pledges US$1.5B to Singapore, plans 1,000-plus local staff: The cross-border payments firm will invest the sum over five years and grow its local team from more than 750 staff. By e27’s reckoning, the pledge outstrips Singapore’s entire fintech funding haul.

Singapore’s SPIRE Asia targets US$500M for SEA farm climate resilience: Ravi Menon launched the blended-finance programme at COP31, citing Asia’s yearly adaptation need of up to US$400B against US$34B in actual flows. Temasek Trust will package smallholder projects into investment-ready portfolios.

AMRO says Thailand’s AI investment boom could revive long-term growth:Digital-sector investment applications in H1 2026 already exceeded all of 2025, but data centres import most high-value kit and hire few people. AMRO urges local supplier development and wider SME access to cloud and AI.

MDEC ropes in seven lenders for SME e-invoice financing exchange: CIMB, AFFIN, Alliance Bank, Funding Societies, CapBay and Bettr will tap authenticated Peppol e-invoice data to assess borrowers with thin credit histories. The planned pilot also tracks loans and blocks double financing.

Vietnam’s Easy AI, AI Leaders tie up to embed AI agents in businesses: The partners will combine Easy AI’s Revenue AI OS, which serves more than 300 businesses, with AI Leaders’ consulting and executive training, and build department-specific agents measured against revenue and cost targets.

GCash’s loan book grows 43% as Mynt posts record Q3 revenue: Q3 revenue hit US$400M, up 16.6%. Loans reached US$1.3B, growing four times faster thanusers, while the bad-loan ratio held at 4.8%. Q4 will absorb most one-off IPO costs.

Philippine Series A value falls 98% as Series B jumps 795%: Gobi Partners data shows fintech took 78% of deal value while AI-native startups got under 1%. Mynt’s US$960M IPO mostly paid existing shareholders, leaving little fresh money for early-stage founders.

Lunio parent Horizon raises US$12.5M from Vertex to move beyond beds: The Bangkok firm will stretch into ergonomics, recovery devices and LED beauty tech across Thailand, Taiwan, Vietnam and Malaysia. It claims 4.3M followers but discloses no revenue, profit or customer numbers.

SeaX bets US$1.65M on Skribe’s AI heart patch for cancer patients: More than 800 patients have enrolled in its NOECHO study, including at one Southeast Asian site, with SeaX helping recruit over 100. FDA programme selection is not clearance, and no accuracy data is public yet.

AquaBloom raises seed funding for seaweed-based crop biostimulants: Spiral Ventures led the undisclosed round. The Jakarta startup claims 20–30% yield gains across 40-plus field trials on almost 30 crops, but has not detailed who ran them or over how many seasons.

Stablecoins must behave like bank money to win APAC firms: survey: Only 10% of 150 respondents use stablecoins today, and 96% say regulatory compliance would build confidence. Morph, which sells stablecoin payments, commissioned the survey, and samples run thin per market.

INTERVIEWS & FEATURES

ShopBack’s Justin Lee on letting Meta’s Muse in as Amazon shut it out: ShopBack was one of eight launch connectors when Meta switched on its Muse agent in Singapore. In the US, Amazon blocked it within 12 days, arguing agents bypass the personalisation built into its store.

Kerala bets on working capital and its elders to build bigger startups: At Huddle Global 2026, Chief Minister V. D. Satheesan unveiled plans for a state debt consortium so founders need not sell equity to grow. Kerala now counts over 9,200 startups, up from 2,200 in 2019.

Beyond GCash and Maya: 30 fintech startups rewiring Philippine money: The list profiles newcomers taking on a gap long filled by padala, paluwagan savings circles, five-six lenders and sari-sari store credit, because banks shunned small customers.

How Hostinger helps SEA solo founders go from idea to paying customers: The partner piece tackles the weekend founder’s checklist of domain, hosting, site builder, payments, mailing list and customer data, setting out how Hostinger’s tools cover it without a developer.

INTERNATIONAL

Ex-Nvidia CEO Huang agrees to testify before Congress by Thanksgiving: He relented after senior Republican lawmakers threatened a subpoena. Committee chair Brian Mast expects video and a transcript of the closed-door session, saying Huang has serious questions to answer despite Nvidia’s importance to the US.

WhatsApp chief Kunal Shah plans bill payments, loans and smarter AI: The CRED founder, five months into the job, wants voice notes, groups and stickers simpler to use. Its 3 billion-plus users send over 10 billion voice messages daily, and Shah promises messages stay end-to-end encrypted.

SoftBank’s US$3B-plus turbine write-off may cap AI’s natural gas rush: The group reportedly put deposits on more than 13GW of turbines, nearly 50% above its needs, then posted a US$7.1B quarterly loss. Analysts say grid links plus large batteries may suit data centres better.

Hot US inflation print drags Bitcoin below US$73,000: October headline CPI rose 0.4% month on month against a 0.3% forecast, pushing Bitcoin to a near three-month low. Anndy Lian weighs what a bigger Fed move would mean for crypto.

Europe’s tokenisation push hinges on adoption, not a 2028 blueprint: The author argues the Eurosystem’s Pontes strategy underestimates the cost of persuading businesses to connect, since rebuilding applications and changing custody arrangements erodes the commercial value of a technically capable network.

Bitcoin is bruised near US$75,000 but the crypto sector holds firm: Trading at about US$74,980 on 10 November after its lowest levels since August, Bitcoin has withstood a slide that would have broken weaker assets, Anndy Lian argues, despite alarming headlines.

CYBERSECURITY

Cyberattacks on SPH Media expose personal data of 9% of its database: A single threat actor hit five systems between 4 and 8 November, exposing mainly email addresses. The publisher of The Straits Times and Tech in Asia says attackers accessed no payment card or banking data.

Cyberattacks could make open-source AI less open, Hugging Face warns: Policy chief Irene Solaiman cites the OpenCode API-key theft and an OpenAI agent swarm breach. Tighter identity checks could shut out hobbyists; for now, Hugging Face rotates admin tokens more often instead.

Singapore’s CSA funds 13 firms with US$4.3M to secure AI agents: Grants average about US$330,000 per company, less than many seed rounds. Projects span MCP traffic inspection, quantum red-teaming and a virtual CISO for SMEs, while CyberCall 2026 runs until 9 January 2027.

Pasting client data into ChatGPT is a cross-border transfer under PDPA: Vasileios Stergiou warns the US CLOUD Act lets authorities compel US AI providers to hand over data wherever it sits, and urges regulated firms to move workloads onto locally run open-weight models.

Unpatched email servers are a governance failure before a security one: After a July flaw disclosure, confirmed compromises of unpatched servers rose from 155 to 274 within 48 hours of CERT Polska’s indicators. IBM puts the average Southeast Asian breach at US$4.12M.

SEMICONDUCTOR

TSMC approves up to US$1B in stakes in OpenAI and Anthropic: The board cleared up to US$500M each, with the Anthropic purchase tied to its IPO price, plus another US$500M for an unnamed chip-ecosystem vehicle. It also approved about US$40.93B in capital appropriations.

Staera raises US$4.6M to replace copper with on-chip light for AI: The SMART spin-off, led by Openspace Capital with SGInnovate joining, targets under 0.2 pJ/bit against about 5 pJ/bit for co-packaged optics. Those are design goals, with commercial release pencilled in for 2030.

AI

OpenAI’s quarterly revenue reportedly tops US$10B for the first time: Third-quarter sales jumped almost 50% from US$6.7B, Bloomberg reported, and the adjusted operating margin narrowed to -109%. Rival Anthropic still leads, reporting US$17B in Q3 revenue ahead of its IPO.

New York’s RAISE Act puts OpenAI and Anthropic under state oversight: From 1 January, frontier labs with more than US$500M in annual revenue must publish safety frameworks and report critical incidents within 72 hours. Fines reach US$1M for a first violation and US$3M thereafter.

AI reskilling gap leaves 96,500 Filipino workers a year without a seat: Gobi Partners says 150,000 need reskilling annually but capacity tops out at 53,500. Women hold 58% of BPO jobs while emerging machine-learning roles skew 68% male, and TESDA earmarks just US$8.7M for AI micro-credentials.

Sionic AI brings its top-ranked Thai OCR model to Southeast Asia: The Seoul firm’s PepperOCR-VL runs on a single GPU, suiting on-premise bank deployments, and it built an agent platform with Siam AI. A Singapore insurance chatbot trial nears completion, with no paying customer named.

THOUGHT LEADERSHIP

Venture debt is reaching smaller startups. Check the terms first: SEA recorded 23 debt deals worth US$1.36B in H1 2026, up from 17 a year earlier, while the average deal shrank from about US$99M to US$59M, pulling smaller companies into the market.

What Singapore’s regulated institutions ask fintech partners first: MAS’s AI risk management guidelines, issued on 7 October, phase in over 2027 and 2028, while its new Future of Finance Institute will extend the MindForge toolkit to AI risk assessments through SafeFin.ai.

3 ways the EDGE Grant changes how Singapore SMEs buy software: For years, a vendor’s place on the PSG pre-approved list largely decided SME software purchases. The author sets out what sensible owners should weigh now that the EDGE Grant has changed that.

Most APAC bank AI strategies are the 2015 playbook at lower cost: Ahead of the Singapore Fintech Festival, the author argues that beneath the agentic and generative AI branding on every stand, banks are mostly running old strategies with a lower cost base rather than rethinking them.

AI agents that review every payment could upend subscriptions: Subscriptions win because renewing requires no fresh decision. John Millar asks what happens to that predictable revenue once AI agents scrutinise each recurring charge on a customer’s behalf.

A WNJ Ventures investor on using AI to fix VC’s decision bottlenecks: Wing Lee maps a typical venture investment process and shows where AI eases the bottlenecks and inefficiencies that can lead investors into untimely or wrong investment decisions.

The DEFY framework shows where agentic AI earns its marketing keep: Prof. Aleks Farseev starts from a gap in BCG’s survey of 300 CMOs: 96% claim AI has changed their function, yet few can say what their agents did on their own.

When buyers ask AI who to buy from, SMEs must make the shortlist: AI assistants now give buyers a few named suppliers rather than ten links. Donghyun Kim sets out what that changes for small businesses in Southeast Asia and the plain work that improves their chances.

The future of AI is distributed, not centralised, says David Seng: After a year of using AI to write Python programs on an Ubuntu server and relearning old Linux commands, the author argues computing is circling back to a client/server model.

AI agents can ship working code yet lose the decision behind it: In a hypothetical workout app, one agent builds local storage for account-free recording, and a later agent adds synchronisation. Tijn van Daelen shows how the original rationale gets lost between handovers.

The most expensive AI mistakes don’t look like mistakes: Norm Bond recounts an enterprise using AI to turn years of internal documentation into a tidy wiki on a legacy product’s engineering setup, meant as new engineers’ single reliable source.

AI is a tool, and treating it as a strategy misses the point: Using the image of a Swiss army knife that does everything, Larry Chua argues businesses need the right tool for a specific job rather than an all-purpose AI programme.

One worker, two hours of AI drama a week: The real production costs: Xiaohu Hou says model usage, generation and related tools made up about 60% of direct production costs on a client-accepted project, with the author’s labour accounting for the rest and no team involved.

A self-taught builder fears AI will pull the ladder up behind him: Faheem Aizat Kamsan, who built OnTheRice, credits AI with letting him research, connect information and build systems from outside the industry, and warns that the same technology could shut the door on others.

Four hundred business cards and zero customers: The trade show trap: David Kim describes founders leaving shows with a dozen promises, sample requests and praise, but no conversations with actual users, because booths draw importers, distributors and consultants rather than end customers.

Five ways Huang’s exit could hurt the AI industry he championed: TechCrunch’s Julie Bort argues open-weight models, agent safety efforts and Nvidia’s financing of data centre build-outs, including a US$500B private-equity pledge, have lost their loudest advocate just as lenders sour on AI debt.

The post Jensen Huang resigns from Nvidia as alleged spy affair wipes US$300B appeared first on e27.

In 2019, Ubisoft invited me to Paris. As I strolled through Station F, the sprawling startup campus that French billionaire Xavier Niel opened in 2017, I ran into a name I never expected to see nearly 8,000 km from home: Kerala Ventures. A venture capital firm, founded by French partners, named after the state I hail from.

Curious, I emailed them to ask why. The answer was disarmingly simple. The name was deliberate, they said, because they believed the next decade belonged to Kerala.

Seven years on, more than half of that decade is spent. So when Vikram Gupta, Founder and Managing Partner of Mumbai-headquartered IvyCap Ventures, took the stage on day two of Huddle Global 2026 in Kovalam with a keynote titled “Kerala’s Moment – From Talent to Scale”, the question hung in the salty air: has the bet paid off, or is Kerala still a promise?

Also Read: Why Kerala’s grandparents may be its smartest startup bet

Gupta’s verdict is that the fundamentals are in place: a deep pool of talent working both inside the state and across the world, visible momentum among young founders, state-backed capital, and a growing roster of VC funds supported by the government. What is missing is the plumbing that carries a company from early traction to genuine scale.

That diagnosis lands at a convenient moment. A day earlier, Chief Minister V.D. Satheesan opened the eighth edition of Huddle with a scale-up mission that shifts the target from launching 10,000 startups to building more than 100 that each earn over INR 100 crore (about US$11 million) in annual revenue. Gupta, who runs one of India’s largest homegrown VC firms and invests through the alumni networks of the IITs and IIMs, effectively laid out what it would take to get there.

Here are the five lessons from his keynote.

1. The diaspora is an asset class Kerala has not built yet

Gupta’s most distinctive argument is about money that already flows into the state. According to the RBI’s latest remittances survey, Kerala received 19.7 per cent of India’s inward remittances in 2023-24, second only to Maharashtra. Almost all of it goes into family maintenance and savings.

His pitch is to channel a slice of that into a professionally managed endowment, a long-term corpus that pulls the ecosystem onto a single platform. The Gulf diaspora, he argued, brings more than cash; it brings networks and access to markets in the UAE, Saudi Arabia and beyond. Gupta knows the model: he set up India’s first endowment fund at IIT Delhi in 2019.

The caveat is obvious. Remittances are household income, not idle risk capital, and endowments live or die on governance and patience. Southeast Asia has learned the hard way that the weak link is often the LP base, not the VC model itself.

2. The real valley of death sits between years three and five

Across India, Gupta said, there is now some momentum at the earliest stages. The trouble begins between years three and five, where too many Kerala startups simply get lost. Funding is only part of it. He pointed to gaps in CSR and grant money, the struggle to land a first anchor customer, and a mismatch between the timelines of capital and technology.

Deeptech illustrates the problem. Investment in the space rose, dipped and is now recovering, but early-stage grants remain scarce and translation capital thin. As founders across the region have found, a deeptech company rarely has just one capital problem.

3. Patient, domestic institutional capital is strategic, not optional

The Indian VC market is crowded, yet capital remains scarce at specific stages, and Gupta expects that to persist. His remedy is domestic institutional money, from pension pools to the Centre’s INR 1 lakh crore (about US$11 billion) Research, Development and Innovation (RDI) scheme.

Also Read: Build what matters or what gets funded? India’s VCs say the real gap sits in the founder’s head

He described RDI capital as money with a horizon of up to 20 years at very low interest, which has begun moving into a handful of funds and research-led pockets after a slow start. Strategic and corporate investors, he added, are also showing up earlier in the funding cycle. For founders, the message was blunt: learn how this capital works and go after it.

4. Build clusters around research, not just more incubators

The US and China, Gupta noted, built their innovation hubs around universities and research institutions. If India wants to win in AI, climate, food and health, it needs targeted hubs, and Kerala is positioned to host several. Japan is running a similar experiment, with Kyoto rather than Tokyo emerging as its real deeptech bet.

The mechanics matter more than the buildings. Gupta wants research-to-venture pipelines where academia, translation funding and hands-on support converge, because deep tech needs scale-up teams, not just scientists. Corporates have a role too: customer access, he said, is the biggest hurdle for most startups, echoing the argument that customers, not capital, are deeptech’s real bottleneck.

Southeast Asian readers will recognise the template in Singapore, where National GRIP shepherds deep tech ventures from lab to market. IvyCap’s own twist is to recycle part of its profits into endowments at institutions such as the IITs, putting money back into research.

5. AI has rewritten the cost of competing, so prove capital efficiency

AI, Gupta said, is changing not only how products are built but the cost of building them and the nature of competition. When it is cheap to build, it is cheap to copy, and investors want to see defensibility and a credible business model.

His sharpest note was on capital efficiency. Too many impact-driven companies, he argued, fail to show investors how their capital will actually be used. Hype will not close that gap; unit economics will.

A first cheque, and a founder’s homecoming

Gupta closed by handing the stage to Aparna, founder of NeuMe NeuroTech, which he described as the first investment from the endowment he has been championing. Her story was personal: a family member survived a stroke, and physical therapy got her walking again, but not back to herself. “The body survives, but the mind remains stranded,” Aparna said.

Also Read: India’s IPO boom is rewriting the exit playbook for global investors

She cited roughly 1.8 million stroke cases a year in India, with four in ten survivors struggling with memory, attention and other cognitive deficits that standard rehabilitation, which measures progress in physical milestones, largely misses. Her startup pairs multimodal brain-computer interface hardware with adaptive, game-like cognitive software designed to keep patients engaged rather than frustrated, and positions it as a tool for clinicians and neurologists, not a replacement. She launched it officially at Huddle, in her home state.

One cheque does not make an asset class, and the endowment’s size, governance and return expectations remain unanswered. But if diaspora money begins backing companies like this at scale, the French partners who named their fund after Kerala may yet be proved right. On their own clock, the state has a little over three years left to show it.

The post Kerala’s moment: Five lessons from IvyCap’s Vikram Gupta on turning talent into scale appeared first on e27.