Here’s a simple way to think about FinSight Ventures: it is a late-stage fund that likes the places most investors avoid, and it has a one-line rule for deciding when to write a cheque: 35 per cent a year, in dollars, or it walks.

The US-based VC fund with operational hubs in Cyprus and India invests in AI, cloud, enterprise software and fintech, and it has done more than 60 late-stage deals worth over US$700 million. Its portfolio now runs to nearly 100 companies spread across South Korea, Central Asia, Latin America, the US, Europe (it holds a stake in buy-now-pay-later giant Klarna) and a long list in India, including CarDekho, Razorpay, Gupshup, MediBuddy, Bimaplan, Superfone and Visa2Fly. What that book still does not have, despite the firm describing itself as having “a strong strategic focus on Southeast Asia,” is a single Southeast Asian company.

Also read: SEA has the ideas; it needs the follow-through

We put that gap to Pavel Gurianov, FinSight’s Principal for Emerging Markets covering India, Indonesia and Latin America, along with tougher questions on fee structure, a Chinese-adjacent cap table, and what “liquidity” even means in a hyperinflating economy.

The missing middle in private markets

FinSight’s pitch is simple, once you see the gap it says it is filling. “In public markets you can pick stocks, hand money to a hedge or mutual fund, or buy the middle product — an index fund or ETF,” Gurianov said.

Private markets, he argued, never built that middle tier: you either do deal-by-deal co-investment, or you lock into a 10-12 year venture fund where, for the first four or five years, “you don’t know what you own.”

‘FinSight’s answer is a named, closed list of bets, disclosed before an investor commits. Its first such vehicle, the Generative AI Index Fund, launched last November as a US$50 million vehicle giving investors one ticket into a stated basket of AI leaders — Fin AI, Together AI, Crusoe, Glean.’ Gurianov puts the fund at 14 companies in total; FinSight is now readying a second, on Physical AI, with a dozen names, and the firm tells e27 a formal announcement is expected within days.

The terms on both are blunt: a five-year fund life, half the length of a standard venture vehicle, a 2 per cent management fee, 20 per cent carry, and no interim liquidity. “We invest, we hold, and we distribute as we exit each position,” he said.

The underwriting discipline, he insisted, is not glamour-driven. Every deal is priced against a price-to-earnings exit multiple, on the assumption that “any company, technology or not, eventually trades on earnings once it is public.”

The bar is that same 35 per cent annualised return in the middle scenario, not the bull case. Fail that test at today’s price, and FinSight walks rather than repricing its ambitions downward.

Physical AI’s underpriced bottleneck

The new fund is a wager that robotics and embodied AI are where generative AI was around 2023, except Gurianov thinks the market has mispriced which bottleneck matters. Not hardware, which he said is down 40-60 per cent in recent years on the back of overlap with electric-vehicle supply chains. Not talent, which is mobile. Data.

“The internet is digital by default; the physical world is not digitised,” Gurianov said. The same scaling law that powered large language models — more data in, exponentially better performance out — has now been proven for robotics, he argued, which shifts the real constraint to the supply of real-world data. He pointed to a crossover already underway: robot-hour costs converging with human labour-hour costs in several industries, tipping this year or next.

His pricing argument is the most quotable line of the interview: the top 30 private physical AI companies are worth just over US$400 billion today, roughly where the top 30 generative AI companies were valued in mid-2024, before they approached US$3 trillion. He is betting on a repricing, “probably not in 18 months,” since robots have to physically enter factories and homes rather than ride existing device penetration.

A Chinese co-investor, and the geopolitics FinSight says isn’t there

The sharpest question was about optics. In March, an Oman-led funding round valued Uzum at US$2.3 billion, with existing shareholders, FinSight among them, joined by Tencent and VR Capital Group. That is a Chinese name on the cap table of a payments company, at a moment of heightened US scrutiny of exactly that kind of exposure. Gurianov pushed back on the framing before answering it. “None of the investors you named is a state entity,” he said, describing them as private firms, some publicly listed.

Also read: Give physical AI a soul: Why your voice AI still feels like a bot

His broader case is that a diversified cap table mirrors a diversified economy: Uzbekistan, which he called the fastest-growing economy in Central Asia, is simultaneously attracting American, Middle Eastern and Chinese capital, alongside European and American companies opening local offices. On regulation, he noted Uzum’s primary compliance obligation runs to the Central Bank of Uzbekistan, where it holds financial and banking licences.

Betting on a dollarised Venezuela

If Uzum is a geopolitical question, Cashea — FinSight’s bet in Venezuela — is a plumbing one: how does capital move in and out of a market under currency controls and a history of hyperinflation that most global investors consider uninvestable? Gurianov’s answer starts by disputing the premise. The economy, he said, is “substantially dollarised,” and inflation has slowed well past its 2017-2018 peak, with growth resuming since 2020.

Cashea, which FinSight describes as Venezuela’s largest fintech, raised US$100 million across two rounds this year, a Series A and a Series B that FinSight led. It is the clearest sign yet of how far the firm is willing to lean into a market most peers won’t touch — and it isn’t a one-off.

FinSight’s Latin American book also includes Rappi, the super-app operating across Colombia, Brazil and Mexico, and Punto Pago, a smaller bet on building a Kaspi-style ecosystem across the Pan-Caribbean, starting in Panama.

On an eventual exit, Gurianov pointed to Kazakhstan’s Kaspi as the template — a company few international investors could name before it listed, now comparable in size to India’s Bajaj Finance and ahead of it on net income. For Cashea specifically, the case is simpler: the company is already profitable, which he treats as the precondition for any exit route, whether a foreign listing, M&A, or dividends.

On Rappi, long dogged by a cash-burn narrative, Gurianov said he ignores most unit-economics chatter and tracks exactly three metrics — EBITDA per order, EBITDA per customer, and company-level EBITDA — all of which he said are now positive.

India first, Southeast Asia second, and still absent

Pressed to rank India, Southeast Asia and Latin America by risk-adjusted return potential, Gurianov didn’t hedge: India first, Southeast Asia second, Latin America third. India wins on liquidity — the NSE lets companies list at US$100-300 million rather than requiring a billion-dollar valuation — despite expensive assets, crowded local capital, and a rupee that gives back 4-5 per cent a year against the dollar.

Southeast Asia, he said, is “the most important region of the next 10 to 20 years,” with roughly 700 million people and outsized trade flows, but its constraint is the same one holding back FinSight’s own entry: thin IPO history beyond Grab and Sea, and exchanges less developed than India’s. Hong Kong, in his view, is the variable that could change that calculus.

His answer to why Southeast Asia is still missing from FinSight’s book, despite the firm’s stated strategic focus, was disarmingly direct: “It is a fair observation. We are working on the region now and expect to add portfolio companies there shortly.”

Also read: Southeast Asia solved distribution: Now fintech has to scale on the balance sheet

That is more than a talking point. FinSight told e27 its team has just returned from Singapore and Indonesia, where it met founders on the ground, and that it hopes to announce a major Southeast Asian deal within months. Whether that lands before the next fundraising cycle, or before Hong Kong’s listing ambitions for the region firm up, is the trade FinSight’s own thesis says it should be watching.

For a firm that walks away from anything that doesn’t clear 35 per cent, the fact that it is still circling rather than committing tells you as much about Southeast Asia’s pricing as it does about FinSight’s patience.

The post Why FinSight thinks Physical AI is mispriced by a decade of data, not dollars appeared first on e27.

Databricks is putting more weight behind Singapore as large companies across Asia shift from experimenting with artificial intelligence to trying to run it safely inside their core operations.

The US data and AI company said it will invest more than US$350 million in Singapore over the next three years, expand into a new 32,000-square-foot regional headquarters, and grow its local workforce from about 250 people to more than 500.

The new office at IOI Central Boulevard Towers will quadruple Databricks’s current Singapore footprint and serve as its Asia Pacific and Japan hub.

Also Read: Why every warehouse in Singapore will run on AI safety monitoring within five years

The move comes at a moment when AI adoption in Southeast Asia is entering a more difficult phase. Over the past two years, banks, telcos, insurers, logistics firms and government agencies have tested generative AI through pilots, chatbots and internal productivity tools. The harder question now is whether these systems can be trusted with sensitive enterprise data, regulatory scrutiny and real business workflows.

That is the gap Databricks wants to occupy.

“Organisations across the region are moving quickly from AI pilots to production, but doing so successfully requires trusted data and context, strong governance, and control over models and costs,” said Simon Davies, SVP and GM of Databricks Asia Pacific and Japan.

Why Singapore matters

Singapore has long positioned itself as a regional command centre for enterprise technology, helped by its concentration of banks, multinational headquarters, government-backed digital infrastructure and deep pool of technical talent. Its National AI Strategy has also made AI a policy priority, with an emphasis on responsible deployment rather than unfettered experimentation.

For global software companies, that combination is useful. Singapore is small enough to test with government and regulated industries, but connected enough to influence technology buying decisions across Southeast Asia, India, Japan, Australia and the wider Asia Pacific region.

Databricks’s investment reflects that role. The company said the expanded headquarters will include training and collaboration facilities for customers, partners and learners to develop data and AI skills, test use cases, and move projects into production.

That focus on skills is not incidental. Across Southeast Asia, many companies still struggle with fragmented data, legacy systems and a shortage of engineers who understand both data infrastructure and AI deployment. The excitement around generative AI has often run ahead of the readiness of internal systems.

A chatbot is relatively easy to launch. A governed AI agent that can retrieve the right internal information, take action, respect permissions, and operate within budget is much harder.

From data lakehouse to AI agents

Databricks built its business around the “lakehouse” idea, which combines elements of data lakes and data warehouses so companies can store, manage and analyse large volumes of data in one place. That foundation has become more important as businesses try to build AI tools on top of their own data rather than rely only on public models.

Also Read: AI agents could help Southeast Asian firms untangle cross-border payment costs

The company is now pushing a set of products aimed at what it sees as the next phase of enterprise AI. Lakebase, its serverless Postgres database, is designed to provide a fast and secure operational database layer for AI agents. Genie acts as an AI coworker that helps users query business data and get answers grounded in enterprise context. Unity Gateway provides governance, model routing and cost controls across different AI models, tools and agents.

Put simply, Databricks is betting that enterprises will not rely on a single AI model or vendor. Instead, they will need systems that let them choose between models, control access to data, monitor usage, and avoid runaway computing costs.

That message is likely to resonate in sectors such as financial services and telecommunications, where Southeast Asia has some of its most aggressive AI adopters but also some of its strictest compliance requirements. A regional bank, for example, may want AI systems to help with fraud detection, customer service or wealth management, but it must also ensure that customer data is protected, outputs are explainable, and regulators can audit what happened.

Customers in regulated sectors

Databricks said its customer base in the region now includes iFAST Corporation, Singapore Customs and Singtel. They join other organisations using its platform, including Airwallex, CPF Board, GovTech Singapore, LG Electronics, Standard Chartered and Toyota.

The mix is notable because it spans both private and public sector users. In Singapore, government agencies have been active in adopting data platforms and AI tools, but public-sector deployments typically require a higher bar for governance, security and accountability. Winning such customers can help enterprise software companies build credibility in other regulated markets in the region.

Singapore Customs, for instance, operates in an area where data quality, cross-border coordination and risk detection are central. Telcos such as Singtel sit on vast network and customer datasets, which can support everything from service optimisation to fraud prevention. Financial platforms such as iFAST need to balance personalisation and automation with compliance.

These are not the low-stakes use cases that defined the first wave of generative AI trials. They are closer to the infrastructure layer of the economy.

A crowded enterprise AI race

Databricks is not alone in chasing this opportunity. Its closest global rival is Snowflake, which has been expanding from cloud data warehousing into AI and application development. The large cloud providers are also formidable competitors: Microsoft is bundling Fabric, Azure AI and OpenAI services into its enterprise stack; Google Cloud combines BigQuery with Vertex AI; and AWS offers a broad set of data and machine learning tools through services such as Redshift, Bedrock and SageMaker.

In Southeast Asia, this rivalry is intensified by the fact that many large enterprises already buy from multiple cloud vendors. Rather than replacing existing systems outright, Databricks will often need to fit into hybrid environments where CIOs are trying to avoid lock-in while still moving fast on AI. Its pitch around openness, governance and model choice is aimed squarely at that concern.

The company’s global scale gives it a strong starting point. Databricks says more than 20,000 organisations worldwide use its platform, including 70 per cent of the Fortune 500. But Southeast Asia is not a simple copy of the US or Europe. Markets differ sharply in cloud maturity, data regulation, talent availability and AI readiness.

Also Read: Why Singapore’s biggest startup opportunity isn’t AI; it’s building ASEAN’s operating system

That makes Singapore a logical base, but not the whole story. The bigger test will be whether Databricks can use its expanded presence there to support customers across more complex regional markets, from Indonesia and Thailand to Vietnam, Malaysia and the Philippines.

Its US$350 million commitment suggests the company expects enterprise AI spending in Asia to deepen, not fade, after the initial hype cycle. The bet is that companies will move from asking what generative AI can do to asking how they can run it reliably, securely and affordably.

For Southeast Asian enterprises, that second question is where the real work begins.

The post Databricks doubles down on Singapore with US$350M AI investment plan appeared first on e27.

Hang Ten Systems has raised an additional US$53 million in seed funding, just five weeks after closing its first seed round, as large enterprises search for ways to move artificial intelligence from experimentation into core operations.

The new round was led by Xora, a Temasek-backed fund focused on AI and deep tech. Mayfield, which led Hang Ten’s earlier seed financing, and Aramco Ventures also joined the round. The latest cheque brings the Menlo Park-based company’s total funding to US$85 million.

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

For a company still at seed stage, the number is striking. But it also reflects a broader shift in enterprise AI: the bottleneck is no longer whether large language models can write code, summarise documents, or automate parts of office work. The harder problem is making those capabilities reliable, secure, and useful inside complex organisations that run on legacy software, strict governance, and fragmented data.

Hang Ten is trying to position itself in that gap. The company provides advisory, transformation, and applied-AI services to large enterprises, using what it describes as an AI-native delivery model. In practical terms, that means combining agentic code generation, a reusable library of technical “skills”, and domain expertise to build, change, and operate enterprise software with smaller teams and shorter timelines.

Its work so far spans software development, finance analytics, enterprise migrations, human resources, and other internal business functions.

From thesis to signed contracts

According to the company, the two seed rounds closed five weeks apart. During that period, Hang Ten signed several multi-million-dollar contracts with global enterprises, covering both advisory and software development work. Some projects are already underway, while a few have been completed.

The engagements range from short advisory assignments to multi-year development programmes across multiple industries. The new capital will be used to expand delivery capacity, hire across engineering and consulting, and continue building the company’s platform, infrastructure, and agentic skills library.

“Enterprise technology has long been one of the largest cost lines in essential industries — slow to change, and slower still to translate AI into real economic value,” said Phil Inagaki, Managing Partner and Chief Investment Officer at Xora. “The bottleneck is implementation: deploying it securely, economically and at speed.”

That point is especially relevant in Southeast Asia, where large banks, telcos, manufacturers, energy firms, and government-linked companies have been testing generative AI over the past two years. Many have run pilots in customer service, software engineering, document processing, and knowledge management. Fewer have rebuilt workflows around AI in ways that materially reduce cost or improve productivity.

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

Singapore has been one of the region’s most aggressive markets in pushing enterprise AI adoption, backed by national AI strategies, public-sector digitalisation, and a deep base of regional headquarters. But across the wider region, adoption remains uneven. Enterprises in Indonesia, Vietnam, Thailand, Malaysia, and the Philippines often face a different mix of constraints: fragmented legacy systems, talent shortages, language localisation, regulatory ambiguity, and the need to prove near-term return on investment.

That is where firms such as Hang Ten see an opening. Rather than selling a horizontal AI tool and leaving customers to figure out implementation, the company is offering a services-led model that embeds AI into delivery itself.

A familiar problem with a new toolkit

Hang Ten’s proposition is not entirely new. Enterprises have long hired systems integrators and consulting firms to modernise software, migrate data, implement enterprise resource planning systems, and redesign operations. What has changed is the toolset.

Agentic AI systems can break down tasks, write or modify code, call software tools, and perform multi-step work under human supervision. In theory, this allows engineering and consulting teams to do more with fewer people. In practice, enterprise use requires guardrails: security controls, audit trails, quality assurance, compliance checks, and a clear understanding of where human experts still need to intervene.

That distinction matters because large companies rarely lack access to AI tools. What they lack is the ability to combine those tools with production-grade workflows. A chatbot demo is easy. Refactoring a core finance system, migrating enterprise software, or automating HR processes without breaking compliance is much harder.

Hang Ten says its model is built for that second category. Its reusable skills library appears to be a core part of the pitch: instead of solving each customer problem from scratch, the company can reuse patterns, code, and process knowledge across engagements.

The company is already working with Aramco on applications supporting operations across several functions. Aramco Ventures CEO Mahdi Aladel said the appeal lies in the breadth of possible use cases inside a large organisation.

“The more closely we look at their approach, the more functions and use cases we find across our own operations where it would fit,” he said.

Siemens Gamesa Renewable Energy is also among Hang Ten’s customers. Its CEO, Vinod Philip, said the company began working with Hang Ten two months ago on a short list of outcomes, before widening the scope after early delivery.

A crowded field, but a large prize

Hang Ten will not have the market to itself. The company is entering a field dominated by global consulting and IT services giants such as Accenture, Deloitte, IBM Consulting, Capgemini, Cognizant, Infosys, TCS, Wipro, and HCLTech, all of which are investing heavily in generative AI delivery. In Southeast Asia, these firms already have deep relationships with banks, telecom operators, manufacturers, and public-sector agencies.

There are also software engineering specialists such as EPAM and Globant, as well as a growing crop of AI-native startups offering coding agents, workflow automation, and enterprise AI deployment tools. Hang Ten’s challenge will be to show that its AI-native services model is not just faster in early projects, but repeatable across sectors, regions, and heavily regulated environments.

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

Its investor base may help. Xora’s link to Temasek gives Hang Ten a potential bridge into Singapore and regional enterprise networks. Aramco Ventures offers exposure to large-scale industrial deployment, particularly in energy and infrastructure. Mayfield brings early-stage company-building experience, while individual backers include Intel CEO Lip-Bu Tan, Micron chairman and CEO Sanjay Mehrotra, and Yahoo co-founder Jerry Yang. Yang has also joined Hang Ten’s board.

For Southeast Asian enterprises, the significance is less about another US startup raising a large seed round and more about what the financing signals. AI implementation is becoming a board-level concern, not an innovation lab experiment. Companies are under pressure to extract productivity gains, but they cannot afford failed deployments in critical systems.

Hang Ten founder and CEO Dr Vishal Sikka (who is former CEO of Indian software giant Infosys) framed the opportunity around that tension between ambition and dependability.

“AI can be the most powerful amplifier we have ever built. But in the enterprise, AI’s power must go hand-in-hand with reliability,” he said.
That is likely to be the defining test for Hang Ten. The market is full of AI promises. The enterprises it is targeting will care less about novelty than whether projects ship, systems hold up, and costs come down.

If Hang Ten can prove that its model delivers those outcomes consistently, its unusually large seed round may look less like hype and more like early infrastructure for a new generation of enterprise services.

The post Temasek-backed Xora leads Hang Ten’s US$53M seed round appeared first on e27.

A customer in Jakarta spends 20 minutes explaining a billing dispute to a chatbot, gets bounced to a human agent, and has to start the story over from scratch. Multiply that across the millions of AI-mediated conversations happening daily across the region, and you get a sense of the trust deficit quietly building beneath Asia Pacific’s AI customer service boom.

New research from customer engagement infrastructure firm Twilio’s 2026 Customer Insights Series puts hard numbers to a problem consumers have long felt anecdotally: AI agents across APAC suffer from what the company calls “AI amnesia”, a tendency to forget who a customer is or what they said moments earlier.

Also Read: Twilio on why AI companies must rethink customer engagement to succeed in Asia Pacific

Seven in 10 APAC consumers say they have abandoned an AI-powered customer service interaction mid-conversation because the system failed to recognise them or lacked context from a prior exchange. Five per cent didn’t just abandon the chat; they walked away from the brand entirely.

For a region where AI adoption in customer service is accelerating on multiple fronts at once, from Singapore’s banking sector to Indonesia’s e-commerce giants to the Philippines’ business process outsourcing industry, that is an expensive gap between ambition and execution.

A confidence gap between brands and their customers

The disconnect starts with perception. Over four in five APAC brands (84 per cent) believe their AI agents are doing a good job recognising returning customers and recalling history. Yet 70 per cent of consumers say they routinely have to start from scratch every time they open a new AI conversation, a gulf between self-assessment and lived experience that should worry any product team that has taken its own dashboards at face value.

The problem doesn’t resolve itself when a human steps in, either. Nearly two-thirds of APAC consumers (65 per cent) report having to repeat themselves or fill in gaps after being handed off from a bot to a human agent, which suggests the underlying issue is systemic rather than a chatbot-specific flaw.

Robert Woolfrey, Twilio’s Asia Pacific and Japan vice-president, put it simply: AI is only as useful as the context fed into it. Brands that connect their customer data properly, he argued, can move past basic chatbots into assistants that actually remember individuals.

The root cause, per Twilio, isn’t the models themselves but fragmented data, a customer’s history scattered across a CRM here, a support ticketing tool there, a separate loyalty app elsewhere. No agent, human or artificial, can assemble a coherent picture from pieces that were never designed to talk to each other.

For Southeast Asian businesses running newer AI agents alongside legacy systems that predate the current AI wave by a decade or more, that fragmentation problem is arguably sharper than in more digitally mature markets. It is also precisely the gap products like the next generation of AI-native CRM are being built to close.

Consumers want disclosure, not just competence

Memory isn’t the only trust issue. The report also surfaces a widening expectations gap around transparency. Seven in 10 consumers believe AI agents should identify themselves clearly at the start of every conversation. Only 22 per cent of APAC businesses currently do this consistently, a gap wide enough to suggest many brands still treat AI disclosure as optional rather than a baseline expectation, despite mounting evidence that the AI trust gap is now a commercial liability, not just an ethical nicety.

Consumers want more than a heads-up, too. Fifty-eight per cent want the option to switch to a human agent on request, 57 per cent want assurance that AI-driven actions require human sign-off before execution, and 51 per cent want visibility into exactly what data an AI system can access. This is a more demanding, more specific version of trust than “does the bot work” — it’s “can I see and control what the bot knows about me.”

Also Read: The AI trust gap: Why SEA startups need proof before they scale

Businesses aren’t entirely ignoring the ask. Some are prioritising explainability features (47 per cent), shorter paths to a human agent (44 per cent), and giving customers more say over whether their data feeds AI model training (39 per cent). But across nearly every trust metric Twilio measured, the gap between what consumers expect and what businesses actually deliver remains wide.

Gemma Calvert, a professor of consumer neuroscience cited in the report, framed the stakes bluntly: when AI is deployed before it can handle the full range of real customer needs, consumers don’t blame the technology in isolation; the failure becomes part of how they experience the brand itself. It’s a useful corrective to the instinct, still common among product teams racing to ship, that a forgetful chatbot is a minor UX bug rather than a brand-trust problem with compounding costs.

The road to AI-to-AI service

Despite the friction, APAC consumers aren’t turning away from AI-driven service — quite the opposite. Sixty-eight per cent say the bots they’ve dealt with have genuinely improved over the past year, and a notable 65 per cent are already comfortable with AI agents from different companies communicating directly with one another to resolve issues, without a human shuttling information back and forth. That comfort level matters: it is the precondition for the kind of agent-to-agent commerce that vendors building payment agents are betting will become the region’s next infrastructure layer.

Everyday task delegation to AI is already mainstream: 86 per cent of consumers are comfortable letting AI schedule appointments, 85 per cent trust it with dinner reservations, another 85 per cent with processing returns, and 82 per cent with picking concert seats. None of these are trivial — each requires a degree of confidence in AI judgement that would have looked premature just a few years ago, and echoes the broader question of whether it’s possible to install real judgement into AI agents at all.

Also Read: The app worked, the product didn’t: Can we install judgement into AI agents?

Businesses are positioning for a future where delegation deepens further. APAC leaders expect AI agents to handle 65 per cent of all customer service interactions by 2027, up sharply from 52 per cent today. And 91 per cent of business leaders say they are actively building AI workflows for proactive support, reaching out before a problem is even reported, rather than waiting for a complaint to land.

What it means for Southeast Asia’s builders

For the region’s fast-growing customer engagement, fintech and e-commerce platforms, the Twilio findings amount to a fairly unambiguous signal: the race to deploy AI agents is outpacing the race to make those agents actually remember customers and disclose themselves honestly. That gap won’t close on its own, and it certainly won’t close by adding another point solution on top of an already fragmented stack.

As AI-to-AI communication moves from novelty toward norm, the startups and enterprises that invest early in unifying fragmented customer data, and in being upfront about when and how AI is being used, stand to capture a trust dividend their competitors are currently leaking away, interaction by frustrating interaction. In a region where word-of-mouth and app-store reviews can make or break a consumer brand overnight, fixing AI’s memory problem may prove just as commercially important as building the AI in the first place.

The post “AI amnesia” is quietly costing Southeast Asian brands their customers appeared first on e27.

The first generation of ASEAN’s unicorn founders is moving on. Some have stepped back from operational roles. Some have moved into investing. A few have taken government appointments and, more recently, moved back out of them. The question I keep coming back to, fifteen years into a risk career that has run alongside this generation’s rise, is what the next cohort of founders is supposed to learn from all of it.

Nadiem Makarim’s trajectory is the most prominent example of the full cycle. He founded Gojek in 2010, built it into one of the region’s first true super-apps, took the role of Indonesia’s Minister of Education and Culture in 2019, served until 2024, and has since returned to private life under public scrutiny over decisions made during his ministerial tenure. The specific facts of any ongoing matter are for the courts to determine. The structural patterns underneath the trajectory are for the rest of us to learn from.

What the first generation built

The unicorn cohort that emerged from Indonesia, Malaysia, Vietnam, and the Philippines between 2010 and 2018 did something previous ASEAN business generations had not. They built consumer technology platforms that absorbed enormous segments of daily economic life — payments, transport, food delivery, e-commerce, financial services. The largest became regional infrastructure, not just products. That scale changed what ASEAN governments expected from technology founders. Public consultations included them. Regulatory drafts circulated through them. The relationship between unicorn founders and the state moved from arm’s-length to participatory inside a decade.

The pivot to public service

It is in that context that the founder-to-public-servant pivot became a recognisable category. Nadiem’s appointment as Minister of Education and Culture was the most prominent example. Sandiaga Uno’s earlier trajectory from private equity to multiple ministerial roles foreshadowed it. Patrick Walujo’s transition from Northstar to GoTo’s chief executive seat in 2023 represents the inverse — a public-facing technology role assumed by someone with an investment background, in a moment where the line between the two was getting blurrier.

Also Read: Why Beyond Border thinks visas are now part of the founder playbook

Each version solved real problems for the institutions concerned. Founders brought operational discipline, technology fluency, and direct user-side experience that career officials often lacked. The institutions, in return, gave founders access to systems and constraints that pure private-sector roles cannot teach.

The governance tensions that emerged

Three structural tensions are now visible enough to name.

Dual-role separation. A founder who moves into public office still owns a meaningful stake in the company they built. The mechanisms for managing that separation — blind trusts, board recusals, family-arm-length arrangements — exist on paper. They are tested only when specific decisions cross the boundary between private interest and public mandate. Most of the legal scrutiny that follows founder-to-minister transitions globally — not just in ASEAN — comes from the friction at that boundary.

Decision velocity asymmetry. Founders are trained to decide quickly with imperfect information. Senior public roles require slower, more documented, more procedurally cautious decision-making. The instinct that produced unicorn-scale results in a startup is the instinct that produces governance friction in a ministry. The transition between the two cultures is real and underestimated.

Reputational concentration. In the private sector, a founder’s reputation is concentrated in their company. In public office, it is concentrated in their portfolio’s outcomes — slower, more contested, more politically interpreted. The reputational risk in public service is structurally larger than the founder’s private experience would suggest.

What the next generation should absorb

Three lessons are worth carrying forward.

Decide the role boundary before the offer arrives. The most defensible transitions I have observed are the ones where the founder has thought through what their company ownership, family economic interests, and public stewardship obligations would look like in combination — before the appointment is on the table. The founders who decide that boundary under public pressure decide it badly.

Also Read: How to turn your founder’s opinions into media-ready narratives

Build the second line before the first line leaves. The institutions that survive the founder pivot best are the ones where the founder had already built a deep operational team capable of running without them. The transitions that get into trouble are the ones where the founder’s exit reveals how much was running on the founder’s personal capital.

Treat public office as a different risk category. The skills, networks, and instincts that build a unicorn are not the same skills, networks, and instincts that protect against the scrutiny of public office. The transition requires its own preparation — legal, governance, communications — that founders typically underinvest in because the private-sector playbook has gotten them this far.

The macro stakes

The relationship between technology founders and the state in ASEAN is not going back to where it was in 2010. The companies are too big, the economic stakes are too high, and the regulatory complexity is too dense for governments not to want founders close to policy decisions. That proximity is not the problem. The lack of standard governance infrastructure around it is.

Makarim’s specific trajectory will be litigated by people closer to the facts than I am. The broader pattern it sits inside — the founder-to-public-servant pivot, the governance gaps that come with it, and the ecosystem’s collective underpreparation — is something every next-generation ASEAN founder will need to think about earlier than they currently do. The right time to plan for that transition is before it becomes available, not after.

—

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 founder-to-minister pivot isn’t the problem, ASEAN’s missing governance infrastructure is appeared first on e27.