Across Japan, corporations are grappling with a familiar tension: promising innovation initiatives that stall before they scale, not for lack of merit, but because rigid organisational structures move slower than the market opportunities in front of them. For a growing number of Japanese enterprises, the response has been to look outward, toward ecosystems agile enough to move at startup speed while stable enough to be trusted with real business problems.

Singapore has become one of the clearest answers to that need. It combines advanced technological capability with a predictable regulatory environment and an English-language business culture, an unusual combination that lets cross-border collaboration happen without the friction that typically slows first engagements between Japanese corporates and international startups.

This is the backdrop to the Boundary Spanning Service, a structured programme developed by Sony Acceleration Platform will connect Japanese corporations with Singapore-based startups capable of addressing specific innovation challenges. For founders in Singapore, understanding this shift, and how to engage with it, is becoming increasingly relevant.

It is also a shift worth watching closely, because it says something about where Singapore’s startup ecosystem sits in the region’s broader innovation map. Japanese corporates are not simply adding Singapore to a list of markets to monitor; they are treating it as a working partner capable of solving problems their own internal teams cannot move on fast enough. That is a different, and arguably more durable, kind of attention than a scouting trip or a pilot programme with no clear next step.

Why Japan is looking to Singapore

Japanese corporations rarely approach business matching service for passive technology scouting. As Sony Acceleration Platform has described in earlier conversations with e27, they typically arrive with clearly defined operational bottlenecks, seeking practical, market-ready solutions across digital transformation, automation, and niche technical capabilities that can be fast-tracked into commercial deployment.

Singapore’s appeal lies in the specifics: a stable and predictable regulatory environment, strong technological capability, and business conducted seamlessly in English. For Japanese corporates trying to manage risk in cross-border ventures, that combination offers a stable, efficient starting point. Singapore’s tech community, meanwhile, brings a level of global compliance and agility that makes it a natural counterpart for enterprises seeking fast, reliable co-creation. Boundary Spanning Service exists to structure that meeting point.

Rather than functioning as a one-off transactional exercise, it is designed around mutual trust and reciprocal value creation, with the Japanese corporate contributing operational resources, industry expertise, and market access, and the Singapore startup contributing agility, speed, and disruptive technology. The goal is to lower the practical cost of survival for good ideas, and to get them into real-world deployment before the window closes.

What this looks like for a Singapore founder

For founders, engagement begins with a streamlined application process supported by e27, designed to be low-friction rather than bureaucratic. From there, Sony Acceleration Platform shares and recommends qualifying startup profiles to the Japanese corporations it works with, meaning founders can expect a structured introduction rather than a cold pitch. What follows is worth entering with clear eyes. Japanese corporate decision-making tends to involve extensive internal consensus-building, a structural characteristic of how these organisations operate, not a signal of disinterest or a personal hurdle. That front-loaded alignment process takes time, patience, and a willingness to meet exacting standards around quality and operational stability.

The payoff for founders who can meet that bar is a relationship that, once trust is established, tends to be exceptionally stable and deeply committed over the long term, backed by the distribution networks and credibility of a major corporate partner. Founders considering Boundary Spanning Service should honestly assess whether they are ready for this type of long-game partnership, rather than a fast transactional deal.

In practice, that means being candid with yourself about a few things: whether your product is genuinely ready for enterprise-grade scrutiny, whether your team has the bandwidth to sustain a longer sales and alignment cycle, and whether you are prepared to treat quality and operational stability as a foundation rather than a formality. Founders who approach Boundary Spanning Service with that mindset tend to be the ones who get the most out of it.

Why Singapore, and why now

Singapore’s position as the starting point for this corridor is not incidental. It reflects a deliberate choice by Sony Acceleration Platform to begin Japan-Singapore collaboration where the ecosystem conditions- technological, regulatory, and cultural- are most conducive to building the kind of trust these partnerships require.

For Singapore founders, it is a rare opportunity to meet Japanese corporate organisations directly, a meaningful step in a process that, by Sony Acceleration Platform’s own account, rewards founders who invest early in the relationship rather than those looking for a quick win.

Founders who engage with Boundary Spanning Service now are positioning themselves at the front of a collaboration corridor that both governments and leading corporations have signalled a long-term commitment to developing. As that corridor matures, early movers are likely to be the best placed to benefit from it.

Register your interest

Singapore founders curious about what this collaboration could mean for their own business have immediate ways to engage: submit an application through the Boundary Spanning Service at https://tally.so/r/815k1k?source=article . The region’s startup ecosystem is evolving quickly, and this corridor between Japan and Singapore offers a concrete, structured way for founders to be part of what comes next.

The region is evolving quickly, and e27 and this collaboration offer the right place at the right moment to be part of what comes next. Register here to join the conversation.

—

Want updates like this delivered directly?

Join our WhatsApp channel and stay in the loop.

The e27 team produced this article.

We can share your story at e27 too! Engage the Southeast Asian tech ecosystem by bringing your story to the world.

You can reach out to us here to get started.

The post Why Sony Acceleration Platform is turning to Singapore for innovation, and what it means for your startup appeared first on e27.

For much of the past two years, crypto investors have been forced to separate signal from noise. Token prices recovered, regulators tightened their grip, and the industry’s loudest claims about remaking finance gave way to a quieter question: which parts of blockchain are actually useful?

Singapore-based digital asset firm Psalion is betting that the answer lies less in speculative trading and more in the plumbing beneath everyday business. The firm has launched its third and largest venture vehicle, a US$50 million fund targeting early-stage startups using blockchain infrastructure in real-world markets.

Also Read: Malaysia, Singapore investors rebalance portfolios without ditching crypto

The fund, Psalion VC Fund III, is structured as a Singapore-domiciled Variable Capital Company, or VCC, a flexible investment fund structure widely used by asset managers in the city-state. It is managed by Conduit Asset Management, which is licensed and regulated by the Monetary Authority of Singapore (MAS).

The choice of Singapore is not incidental. Over the past few years, the country has tried to draw a firm line between crypto speculation and regulated digital asset innovation. The MAS has tightened rules around retail crypto promotion while continuing to support institutional experiments in tokenisation, stablecoins and digital money through initiatives such as Project Guardian. For fund managers like Psalion, that regulatory posture offers both credibility and constraints.

“Crypto was the thesis. The traditional financial system was the antithesis. What we’re investing in is the synthesis, where web2 businesses run on web3 rails,” said Tim Enneking, Managing Partner of Psalion.

That framing captures where much of the venture market around blockchain has moved. The first wave of crypto investing was largely about native protocols, tokens and decentralised finance platforms. The next wave, if it materialises, may be less visible to end users: stablecoins for payments, tokenised real-world assets, blockchain-based trade finance, middleware for digital ownership, and consumer products where the underlying rails are decentralised but the interface feels familiar.

Backing builders before the market turns

Psalion said the new fund will focus on pre-seed and seed-stage companies across infrastructure, middleware, trade finance, real-world assets, stablecoins and decentralised finance. It will also look at consumer applications where web3 technology changes how people own, trade or interact with assets without requiring them to behave like crypto-native users.

The timing is notable. Psalion’s previous two funds were also launched during weaker market cycles, and the firm appears to see the current environment as a feature rather than a drawback.

“Some of the best opportunities arise in down markets. Valuations are more reasonable, founders are more focused, and the builders who show up are in it for the long term,” Enneking said.

That view is common among venture investors, but it carries particular weight in crypto, where bull markets often inflate valuations before products have found real users. The collapse of several high-profile crypto businesses in 2022 and the subsequent regulatory clean-up changed the funding landscape. Startups now face harder questions around revenue, compliance and utility.

For Southeast Asia, those questions are especially relevant. The region has a large underbanked population, strong cross-border trade flows, high mobile adoption and fragmented financial infrastructure. These conditions have long made it attractive for fintech founders. Blockchain companies now need to prove they can solve similar problems without adding unnecessary complexity.

Also Read: How to use blockchain to fund and create a greener future

Stablecoins, for example, have gained attention as a tool for faster and cheaper cross-border settlement. Tokenised real-world assets have attracted banks and asset managers seeking more efficient ways to issue, trade and settle financial products. Trade finance remains a persistent pain point for small businesses across the region, where paperwork, trust gaps and slow settlement can limit access to working capital.

The opportunity is clear. The challenge is distribution, regulation and trust.

Beezie becomes an early test case

Alongside the fund launch, Psalion disclosed that it led a US$4 million round in Beezie, a commerce platform that blends digital ownership, discovery and liquidity into consumer transactions.

Beezie says it has processed more than US$170 million in gross merchandise value since launch and generated more than US$85 million in revenue year-to-date. It also claims to have attracted over 30,000 active users since January 2026. The company’s recent community raise on Echo, initially targeting US$250,000, reportedly sold out in 10 minutes and was capped at US$1 million within 24 hours.

The platform operates across collectibles, luxury and entertainment markets — categories where scarcity, authenticity and resale value matter. These are also areas where blockchain has often been pitched as useful, though consumer adoption has been uneven. Many users care about whether an item is authentic, tradable and liquid; fewer care about whether the backend uses blockchain.

That is precisely the point Psalion is making with the investment.

“Beezie is exactly the kind of company our thesis is built around — a real consumer product, real revenue, and blockchain quietly doing the heavy lifting underneath,” Enneking said. “This is what web2 meeting web3 actually looks like in practice.”

Beezie plans to use the capital for inventory acquisition, geographic expansion and growth across collectibles, luxury and entertainment. Founder and CEO Andrea Miele said the company wants to make transactions feel more engaging by combining ownership, anticipation, discovery and liquidity.

The broader question is whether that emotional layer can translate into durable commerce behaviour. Collectibles and luxury resale markets are already competitive, and user trust is hard won. Beezie will need to show that blockchain improves the experience rather than becoming a technical wrapper around familiar marketplace mechanics.

A crowded field for digital ownership

Beezie is entering a market with several established and adjacent rivals. Globally, platforms such as StockX, Whatnot, eBay, Sotheby’s and Christie’s already serve different parts of the collectibles, luxury and resale economy. In Southeast Asia, Singapore-founded Carousell and sneaker marketplace Novelship have built regional consumer bases around resale and authenticated goods. On the crypto-native side, platforms experimenting with tokenised collectibles and digital ownership have struggled to move beyond early adopters.

Psalion, meanwhile, is competing for deals with a global group of crypto and web3 investors, including Animoca Brands, Spartan Group, Hashed, Dragonfly, Pantera Capital and a16z crypto. The difference will not simply be cheque size, but whether the firm can help portfolio companies navigate compliance-heavy markets and reach users outside the crypto bubble.

Singapore’s role as a base for such funds may strengthen if institutional blockchain adoption continues to move from pilot projects to production systems. The city-state has already positioned itself as a hub for digital asset regulation, tokenisation trials and wealth management. But the region’s startup markets remain uneven. Indonesia, Vietnam, the Philippines, Thailand and Malaysia each present different regulatory and consumer realities.

Also Read: Singapore crypto adoption hits new high as 61 per cent now hold digital assets

For founders, that means a product that works in one market may need significant adaptation elsewhere. Payments, identity, gaming, commerce and financial services all carry local rules and behaviours. A regional web3 startup cannot rely only on the promise of decentralisation; it must solve very specific market problems.

Psalion’s new fund lands at a moment when blockchain has become less fashionable but potentially more useful. The speculative cycle has not disappeared, and neither have the risks. But investors are increasingly looking for companies that can hide technical complexity behind products people actually use.

If Psalion’s thesis proves right, the next major blockchain companies in Southeast Asia may not look like crypto companies at all. They may look like commerce platforms, trade tools, payment networks or financial infrastructure providers, with web3 working quietly in the background.

The post Singapore’s Psalion raises US$50M fund for Web3’s next practical phase appeared first on e27.

Across Southeast Asia, AI is reshaping how workers operate, but not in a uniform way. A new analysis reveals a striking cultural divergence: Filipino professionals are turning to AI primarily as a burnout shield, using it to offload repetitive, draining tasks and protect their mental bandwidth. Singaporean workers, by contrast, are deploying AI as a deep work enabler, a tool to carve out uninterrupted focus time and accelerate high-value cognitive output.

The difference is not merely stylistic. It reflects each country’s distinct labour market pressures and workplace cultures. The Philippines, home to one of the world’s largest business process outsourcing industries, faces chronic worker fatigue at scale. AI is emerging as a structural response to that exhaustion. Singapore, with its knowledge-economy positioning and productivity-first ethos, is using the same tools to raise the ceiling on individual performance.

For founders and operators across the region, the implications are significant. A one-size-fits-all AI adoption strategy will underperform. Localising how AI tools are introduced, framed, and incentivised — by country, by sector, by workforce profile — may determine whether AI delivers measurable gains or simply adds another layer of complexity to an already stretched team.

—–

REGIONAL

Psalion raises US$50M fund for Web3’s next practical phase: Singapore-based Psalion has closed a US$50M fund targeting what it calls Web3’s practical phase, focusing on real-world utility over speculation, one of the larger crypto-native fund closes in SEA this year.

Sunwah Innovations and Brinc launch Vietnam JV for university spinouts: The joint venture will channel university research in Vietnam into commercialised startups, targeting a gap between academic output and venture-ready companies across the country.

Singapore’s Tikva targets solid cancer barrier with US$8M Series A: Tikva is using its Series A raise to advance therapies that address tumour microenvironments, a persistent obstacle in treating solid cancers, positioning Singapore as a node in precision oncology.

Why investors back Vietnamese startups more aggressively than Thai peers: Structural factors, including Vietnam’s younger demographic, higher risk appetite, and stronger technical talent density, help explain why Vietnam consistently attracts more aggressive VC bets than Thailand.

Malaysia and Singapore investors rebalance portfolios without ditching crypto: Institutional and retail investors across both markets are adjusting allocations amid macro uncertainty, trimming risk exposure while retaining crypto positions — suggesting a maturing, not retreating, digital asset stance.

The 27 SEA biotech firms betting on cells, fermentation, and code: A deep-dive list of 27 Southeast Asian biotech companies working across synthetic biology, cultured meat, and computational biology signals the sector’s growing depth beyond Singapore.

Triple-A says own digital assets hit by unauthorised access: Singapore-based crypto payments firm Triple-A confirmed a security breach affecting its own holdings — a significant incident for a firm that processes crypto transactions for major enterprise clients.

SEA’s regulatory patchwork demands a local-first approach: Fragmented licensing regimes across Southeast Asia’s fintech and tech sectors are forcing companies to build compliance infrastructure country by country rather than regionally — raising costs and slowing expansion.

Vietnam’s tech talent market is broken and hiring habits won’t fix it: Most companies are sourcing Vietnamese tech talent through outdated channels that prioritise credentials over capability, exacerbating a structural mismatch between employer needs and available skills.

---

INTERVIEWS & FEATURES

What BEYOND Expo 2026 revealed about Asia’s hardware edge: Observations from BEYOND Expo 2026 point to a regional shift; Asian hardware founders are moving faster from prototype to production than Western counterparts, with supply chain proximity as a structural advantage.

How do you finance a first nuclear reactor for a data centre: A detailed breakdown of emerging deal structures for nuclear-powered data centres shows how project finance, offtake agreements, and government backing are converging to make first-of-kind builds viable.

---

INTERNATIONAL

Anthropic’s Dario Amodei flags risk from Chinese AI, not open-weight models: Amodei clarified he does not oppose open-weight AI development but warned that Chinese AI advancement poses a more material geopolitical risk — a framing with direct implications for SEA’s AI policy debates.

Satya Nadella warns against over-reliance on a single AI system: Microsoft’s CEO cautioned that companies trusting one AI for everything risk structural fragility, a message directed at enterprise AI buyers accelerating consolidation of their vendor stacks.

Lyft and Baidu begin robotaxi testing in London: The partnership marks Baidu’s first Western robotaxi deployment, testing Apollo Go technology in a heavily regulated market, a move that signals Baidu’s intent to internationalise its autonomous driving stack beyond China.

Waymo reportedly weighing a break with Uber: A reported split between Waymo and Uber would reshape the autonomous vehicle partnership landscape, with implications for how robotaxi services are distributed globally, including in markets where Uber dominates ride-hailing.

Microchip Technology acquires Israeli AI chipmaker Hailo: The acquisition gives Microchip a dedicated edge AI inference chip portfolio, strengthening its position in embedded AI hardware used in automotive, industrial, and smart device markets across Asia.

DeepSeek pauses fundraising amid viral post scrutiny: DeepSeek’s decision to halt funding discussions follows a wave of social media attention that drew regulatory and investor scrutiny, a rare instance of viral exposure creating friction rather than momentum for a high-profile AI lab.

Apple sued over US$1.8M App Store crypto scam: A lawsuit alleges Apple failed to remove a fraudulent crypto app despite user complaints, raising platform liability questions relevant to SEA’s growing base of mobile-first crypto users.

Swiggy names Nandita Sinha to head Instamart: The appointment puts an experienced consumer leader in charge of Swiggy’s quick-commerce arm as it battles Blinkit and Zepto in India’s intensifying 10-minute delivery war.

---

CYBERSECURITY

Microsoft launches its first cyber model and agentic security system: The new agentic cybersecurity system uses AI agents to detect and respond to threats autonomously, a significant product move that could reset enterprise security procurement benchmarks across Asia.

Hugging Face CEO calls for transparency after OpenAI hack: Following what he described as an unprecedented breach at OpenAI, Hugging Face’s CEO argues the AI industry must adopt radical transparency on security practices, or risk systemic trust collapse.

Singapore’s agentic AI ambitions hinge on code trustworthiness: Deploying agentic AI systems at scale in Singapore requires a level of software auditability and verification that current frameworks do not yet adequately address, according to a new analysis.

---

SEMICONDUCTOR

Quantum sovereignty in Asia: computing, AI, and emerging ventures: A detailed look at Asia’s quantum computing landscape maps how governments and startups across the region are racing to establish independent quantum capabilities as a strategic national asset.

Are brain waves the next unlock for physical AI?: Researchers exploring brain-computer interfaces as an AI input layer argue that neural signals could give physical AI systems a more precise and low-latency human control mechanism than current interfaces allow.

---

AI

AI voice is scaling across APAC, but listening hasn’t kept pace: Voice AI deployment across Asia-Pacific is outrunning the comprehension capabilities of underlying models, particularly in multilingual and dialect-heavy markets — creating accuracy gaps that affect trust and adoption.

Indonesia’s AI hiring gap is real, just not 28x: Claims of a 28-fold AI talent shortage in Indonesiaare overstated, but the underlying deficit is genuine — driven by a mismatch between university output and the practical skills employers need to deploy AI systems.

AI is making SEA’s startups faster not richer, yet: AI tools are compressing execution timelinesfor SEA startups, cutting weeks off product cycles and reducing headcount needs, but revenue impact remains limited as monetisation models lag behind adoption.

Enterprise AI adoption in SEA accelerates despite internal friction: Southeast Asian enterprises are pushing ahead with digital transformation despite unresolved questions around data governance, integration costs, and workforce readiness, a paradox driven by competitive pressure rather than strategic clarity.

Asia’s AI race will be won by whoever keeps the lights on: Energy infrastructure, not capital or talent, is emerging as the decisive constraint in Asia’s AI buildout, with power availability and grid reliability determining which markets can scale compute at speed.

How AI and blockchain could make commerce decisions more accountable: Combining AI decision-making with blockchain audit trails could create verifiable, tamper-resistant records of automated commercial choices, an approach gaining traction among compliance-focused fintechs in SEA.

AI repriced SEA’s marketing agencies; it didn’t replace them: AI tools have fundamentally altered agency pricing models across Southeast Asia — compressing margins on execution work while shifting value toward strategy, creativity, and client relationships.

Philippines use AI to avoid burnout; Singapore for deep work: A comparative study of AI adoption behaviours across the two markets reveals that workforce culture, industry structure, and labour market pressures shape how and why employees actually use AI tools at work.

---

THOUGHT LEADERSHIP

How the next-generation neobank should be built for the agentic economy: Agentic AI will fundamentally change banking UX, shifting from app-based interaction to autonomous financial agents that act on behalf of users, requiring neobanks to redesign core infrastructure rather than just the interface.

Bitcoin lost US$65,500 support three times before the Fed spoke: BTC’s repeated failure to hold US$65,500 signals weakening buyer conviction at a key technical level — and the pattern emerged before any Federal Reserve guidance, amplifying concern about near-term price direction.

Volatility is here to stay; here is what founders should do: Persistent market volatility is no longer a cyclical event but a structural condition, and founders who treat it as temporary will be outmanoeuvred by those who build resilience into their operating models from day one.

Is ether quietly stealing Bitcoin’s throne as institutional favourite: On-chain data and institutional flow metrics suggest Ethereum is gaining ground over Bitcoin as the preferred vehicle for sophisticated investors, driven by yield potential, programmability, and growing ETF momentum.

SEA’s fintech apps don’t have a literacy problem — they have a fear problem: Low fintech adoption in underserved SEA markets is less about users not understanding the products and more about a deep-seated distrust of digital financial systems rooted in past fraud and institutional failure.

The 3Cs+1 framework for founders navigating geopolitical fragmentation: A practical strategic framework helps founders map their exposure to geopolitical risk across customers, capital, code, and compliance — and decide where to localise, hedge, or exit.

The real workforce challenge: bridging the credential-capability gap: Credential inflation is masking a widening gap between what workers are certified to do and what employers actually need, a structural problem that AI adoption is making more urgent, not less.

How Singapore SMEs should choose a payment provider: A practical checklist for SME payment provider selection covers fee structures, settlement timelines, FX handling, and integration depth, useful for founders navigating Singapore’s crowded but inconsistent payments landscape.

The future of clinical trials: Decentralised clinical trial models using remote monitoring, real-world data, and AI-driven patient matching are accelerating drug development timelines and expanding access beyond traditional trial sites in Asia.

The evolution of trust: from early adoption to institutional maturity: Digital asset markets are moving through a structural trust transition, from speculative retail participation to regulated institutional frameworks, with SEA markets at varying stages of that journey.

The post How the Philippines and Singapore use AI differently at work appeared first on e27.

A few months ago, I was invited to a private university to deliver a keynote on C-Suite skills for 2026 and beyond. The audience was a room full of MBA students. Smart, ambitious, mid-career. People who had paid real money to think seriously about their next decade.

After the talk, one of them raised his hand and asked the question I knew was coming. The question he probably could not ask his own boss. The question that no senior leader at his company wanted to hear out loud.

“Will AI replace senior leaders too?”

The room went quiet. Not because the question was clever. Because it was honest.

Every executive in the world is privately asking themselves the same thing. Almost none of them say it in public. Saying it would look weak. It would unsettle their teams. It would invite the same question right back at them.

But the student had no such constraint. He had paid to learn. He wanted the truth.

Here is what I told him.

I have been studying AI since 2017. Even back then I was telling people that this technology was going to replace a lot of jobs. Including the roles people assumed were untouchable. Doctors. Lawyers. Consultants. Senior managers. Anyone whose work was mainly about applying expertise was going to feel the wave.

But the question most people ask is the wrong one.

AI is a tool. And a tool only performs at its best when someone behind it takes responsibility for the outcome.

So the real question is not “will AI replace you.”

The real question is “are you the one taking responsibility for what AI produces?”

Also Read: The role of thought leadership in scaling beyond your first market

That single shift changes everything. The moment you stop seeing yourself as the person executing a task that AI can do, and start seeing yourself as the person who owns the outcome AI helps produce, you stop being a candidate for replacement. You become the operator of the tool.

This is the part most senior leaders refuse to confront. They built their careers on being the best executor in the room. The most experienced. The most certified. The most credentialed. AI is taking that ground away. And instead of moving up to the layer AI cannot reach, many of them are doubling down on the layer it is colonising.

The student in front of me understood this immediately. He asked the follow up question that most senior leaders never get around to.

“How do I become that person?”

I told him three things.

First, take ownership of outcomes, not tasks. A task is what you do. An outcome is what your work produces in the world. AI can do tasks. AI cannot want an outcome. The person who owns the outcome cannot be automated out of the loop because the outcome is the reason the loop exists.

Second, build judgement in places AI cannot reach. AI is excellent at pattern recognition. It is much weaker at situations where the patterns are new. Reading a customer who has never bought anything like your product before. Navigating a partnership that has never been structured this way. Making a call when the data is incomplete and the stakes are high. These are the moments where human judgement compounds. Spend your career deliberately collecting these moments.

Third, become responsible for results that involve other humans. AI is improving rapidly at producing content. It is not improving at being trusted by a CFO who needs to make a hard decision. Or being chosen by a customer who has options. Or being followed by a team through a difficult quarter. Trust, choice, and followership are the human-only layer. Senior leaders who operate in that layer remain irreplaceable.

This is what I have been doing personally since 2019. I mentor startups. I invest in founders. I help them accelerate beyond their home country. I connect them with local partners in new markets. I help them generate revenue and lift their valuations toward exits.

Can AI do parts of this work? A lot of it, actually. AI can draft introductions, analyse markets, summarise a business plan, generate first-draft strategy.

But who takes responsibility when the call goes wrong?

Who reads the market that does not show up in any dataset?

Who carries the relationships that took years to build?

That part still belongs to a human. And that human is the one who deserves to be in the senior seat.

Also Read: 7 leadership skills every manager needs in a monitored workplace

This is also why I built Future 500 to operate the way it does. Every founder we back gets direct access to people who have actually built businesses from zero. Not because AI cannot do the analysis. It can. But because the founder needs someone willing to take responsibility for the outcome alongside them. That layer of accountability is what AI cannot replicate. It is the layer where senior leadership lives.

When I evaluate a founder for investment, I am looking for exactly this signal. Does this person take responsibility for outcomes, or do they deflect to circumstances? Do they own the numbers, good or bad? Do they propose the next move before being asked?

The founders who answer those questions clearly are the ones I back. The ones who cannot, get a polite pass. AI did not create this filter. It just made the filter sharper, because the candidates who pass it are now even more rare.

If you are a senior leader reading this and feeling shaky about your future, the question to ask yourself is not “will AI replace me?”

The question is “am I the one taking responsibility for what my work produces?”

If you are, AI will multiply you. It will sharpen your judgement, accelerate your output, and free you to focus on the layer of work that only humans can do.

If you are not, you were already replaceable. AI just sped up the timeline.

The MBA student walked out of that hall with a clearer answer than most senior leaders will ever give themselves.

The question is whether you are willing to ask it now, while you still have time to act on the answer.

—

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 future of leadership: Owning outcomes in the age of AI appeared first on e27.

Hyperscalers are preparing to spend roughly US$700 billion in 2026 to make artificial intelligence more capable and cheaper to run. That sounds like a straightforward growth story. More models, more compute, more demand, more data centres.

The risk is that the spending is buying the very forces that weaken part of the demand case.

AI build-outs are usually debated through the depreciation question: will the chips become obsolete too fast to earn back their cost? That is a real issue, but it is not the only one. A more important pressure may come from below. Once a model is good enough to complete a defined task, the buyer no longer needs the newest frontier model for that job. And once the cost of running that level of performance keeps falling, more of the work can move onto hardware the buyer already owns.

The problem is not that AI demand disappears. It probably does not. Cheaper intelligence creates more use. The problem is price. The share of work that can be done by a good-enough model on local hardware sets a ceiling on what a data centre can charge for that work.

That ceiling is falling.

The four largest US technology platforms have guided toward extraordinary capital spending. Amazon has pointed to about US$200 billion, Microsoft to about US$190 billion for the calendar year, Alphabet to US$175 billion to US$185 billion, and Meta to US$115 billion to US$135 billion. McKinsey’s widely cited estimate puts worldwide data-centre capital needs near US$6.7 trillion by 2030, with about US$5.2 trillion AI-specific. Morgan Stanley’s estimate is lower, near US$3 trillion through 2029, but still leaves a financing gap it puts around US$1.5 trillion.

These sums create fixed obligations: debt, power contracts, long-dated capacity deals, and depreciation schedules. Those obligations must be serviced regardless of what price inference commands.

The spending buys two things at once. It buys capability, so models can do more. It also buys efficiency, so a fixed level of AI performance becomes cheaper. Both are desirable. Both also undercut scarcity pricing.

Also Read: How AI and blockchain could make commerce decisions more accountable

Capability has a bound for each task. A tax return, a legal draft, a customer-support exchange, or a code review requires a model good enough to finish the job to a competent standard. Once that threshold is crossed, a better model does not make the completed task more complete. The buyer then moves from “best available model” to “cheapest model that clears the bar.”

Efficiency then moves the same work toward commodity pricing. Epoch AI has found the price of reaching a fixed performance milestone falling between ninefold and several-hundredfold a year in some cases. Andreessen Horowitz has tracked inference at constant capability falling at roughly tenfold a year. Gartner expects inference on a trillion-parameter model to cost more than 90 per cent less in 2030 than in 2025, with on-device and edge inference among the drivers.

That matters because the edge is no longer theoretical. Gartner puts AI PCs near 31 per cent of the market in 2025. Counterpoint puts penetration closer to two-fifths. Canalys expects more than 200 million AI PCs shipped annually by 2028, and IDC expects neural processing units to be near-universal in new PCs by then.

Capability is moving with the hardware. A 120-billion-parameter open model now runs on a single desktop appliance at roughly 32 tokens per second. A 70-billion-parameter model runs on a mini-PC costing about US$1,500 at 12 to 15 tokens per second. Apple’s unified-memory Macs and Nvidia and AMD desktop systems put comparable capability on millions of desks. The best open-weight models now trail the leading closed systems by low single-digit points on neutral capability indexes, with several available under permissive licenses.

This does not mean the edge replaces the frontier. It does not. Frontier training remains capital-bound. The largest facilities have minimum efficient scale that owned devices cannot touch. Workloads needing very long context, low latency at high concurrency, proprietary cloud data, always-on orchestration, or the newest reasoning models still belong in centralised infrastructure.

But that is exactly the distinction. The defensible data-centre market is the work that structurally needs the data centre. The fragile slice is ordinary inference that once lived in the cloud only because a capable model could not run anywhere else.

Also Read: The barrier to AI adoption was never budget, it was knowing where to start

For that slice, the buyer has a standing choice: rent compute from a data centre, or run the workload on owned hardware. The amortised cost of the owned option becomes the maximum sustainable cloud price before work defects. A chip can be fully utilised and still earn a thinning margin if the price it commands is set by what the same job costs on a machine the customer already bought.

This is why the depreciation debate misses part of the issue. Obsolescence asks whether a chip’s useful life is shorter than the accounting schedule. Edge substitution asks whether the work the chip serves is still scarce enough to command the assumed price. Those are different risks.

The price signals are already mixed, which is what should make the issue interesting rather than dismissible. Hourly rates for prior-generation accelerators fell sharply from around US$8 in early 2024 into a US$1.50 to US$3.50 band by late 2025, as newer chips arrived and new providers entered the market. That looked like commodity pressure. Then rates reversed: one-year rental contracts rose about 40 per cent from an October 2025 low of US$1.70 to US$2.35 by March 2026, with on-demand capacity sold out. The rebound shows that demand is still strong enough to absorb capacity. It does not prove the ceiling has vanished.

McKinsey’s own estimate captures the uncertainty. It expects roughly 60 per cent to 65 per cent of US and European AI workloads to sit on hyperscaler infrastructure by 2030. That still leaves a third or more elsewhere, and the cloud-versus-edge split is a live swing factor.

The strongest objection is simple: total demand may outrun the whole problem. Agentic workflows can use five to thirty times more tokens than a chatbot exchange. Older chips can flow down into inference rather than strand. One Bernstein estimate says a five-year-old accelerator can still earn about US$0.93 an hour against US$0.28 of cash cost, implying a contribution margin above 70 per cent. If AI-generated productivity gains keep arriving, demand expansion may swamp pricing pressure.

That objection is serious. The likely outcome is not collapse. It is segmentation. Frontier work remains centralised and expensive. Commodity work gets cheaper. Some of that commodity work stays in data centres because management, integration, security, and convenience matter. Some moves to owned hardware because the economics become too obvious to ignore.

The watch points are clear. If open-weight models keep closing the gap, the local option strengthens. If memory shortages keep edge hardware expensive, the ceiling falls more slowly. If frontier capability re-widens, centralised infrastructure keeps more pricing power. If commodity accelerator rental rates soften while AI PC penetration rises, the pressure is binding.

The AI build-out may still work. But the risk is not just that racks sit idle. The sharper risk is that the racks are busy carrying work whose price is being set somewhere else.

—

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 US$5 trillion AI data-centre buildout unleashes the paradox that limits its returns appeared first on e27.

For over a decade, Sony Acceleration Platform has quietly supported more than 1,000 business development cases through Sony’s internal innovation arm. Its newest initiative, the Boundary Spanning Service, extends that experience across borders, connecting Japanese corporations seeking practical innovation partners with Singapore’s fast-moving startup ecosystem.

Developed in collaboration with e27, Boundary Spanning Service reflects a broader shift in how Japanese enterprises are approaching open innovation: not as a scouting exercise, but as a long-term, trust-based collaboration. e27 spoke with Sony Acceleration Platform team about what is driving this initiative, and what it means for founders on both sides of the corridor.

e27 sat down with the Sony Acceleration Platform leadership to discuss the vision behind Boundary Spanning Service and what it means for Singapore founders.

The vision and context

Could you share a little about what led to the creation of the Boundary Spanning Service – what need or opportunity did Sony Acceleration Platform observe that this initiative is designed to address?

The creation of the Boundary Spanning Service was driven by a strategic need to address a critical structural issue in business development: the tendency for promising projects to ‘run out of time’ before they can fully scale.

Drawing on over 12 years of experience supporting more than 1,000 business development cases within the Sony Acceleration Platform, we observed that one of the reasons why new businesses often fail is not lack of viability, but due to rigid corporate constraints.

To bridge this gap, Boundary Spanning Service was established to act as a ‘Boundary Spanner’. Sony’s own history is defined by absorbing external knowledge to drive growth; thus, facilitating open innovation is a natural extension of our DNA. Through the Boundary Spanning Service, we would like to provide a structured framework that connects Japanese corporations facing specific innovation bottlenecks with Singapore’s agile and highly capable startup ecosystem. By crossing organisational and geographic borders, Boundary Spanning Service lowers survival costs and accelerates real-world deployment, ensuring valuable business developments are not prematurely terminated.

From Sony Acceleration Platform’s perspective, what makes Singapore a meaningful starting point for this initiative? We’d love to understand what Sony Acceleration Platform sees in the Singapore ecosystem that feels relevant to Japanese corporates.

Singapore is uniquely positioned as our starting point because of its exceptional alignment with the strategic needs of Japanese enterprises. The Singapore ecosystem offers a rare combination of advanced technological capability, a highly robust and predictable regulatory environment, and seamless English-language business operations.

For Japanese corporations looking to mitigate risk and operational uncertainty in cross-border ventures, these factors provide an incredibly stable and efficient platform. Singapore’s startup community is not only highly innovative but also operates with global compliance and agility, making it the ideal counterpart for Japanese enterprises seeking reliable, fast-paced co-creation.

Also Read: e27 expands AI-powered business matchmaking with Sony Acceleration Platform collaboration

When Japanese corporations come to the Boundary Spanning Service, what kinds of challenges or aspirations are they typically bringing with them? Are there particular themes or sectors you’ve seen emerge?

Japanese corporations are not coming to Boundary Spanning Service for abstract technology scouting or passive trend-watching. Instead, they bring highly defined, practical operational bottlenecks and specific innovation challenges. They are actively seeking practical, market-ready solutions that can be integrated into their existing value chains.

While the technical fields vary, the underlying theme is the need for rapid digital transformation, advanced automation, and niche technological capabilities that can be fast-tracked for real-world, commercial deployment. They come with a genuine urgency to solve immediate business bottlenecks by leveraging the agility of external partners.

The nature of Japan-Singapore collaboration

How would you describe the spirit of collaboration that Boundary Spanning Service is designed to enable? What does a meaningful, productive engagement between a Japanese corporate and a Singapore startup tend to look like in practice?

Boundary Spanning Service is designed to move far away from low-value, one-off “transactional procurement”. A truly productive engagement must be grounded in mutual trust and reciprocal value creation.

In practice, this means establishing a collaborative framework where both parties act as equal partners. The Japanese corporate provides deep operational resources, industry expertise, and market access, while the Singapore startup provides the agility, speed, and disruptive technology needed to overcome the bottleneck. We foster an environment where cultural and organisational differences are structurally bridged, transforming potential friction into collaborative synergy.

Could you share a sense of what success looks like for both the Japanese corporate and the Singapore startup that comes through the Boundary Spanning Service? Even in broad terms, what outcomes feel meaningful to Sony Acceleration Platform?

Meaningful success is achieved when the collaboration translates into a viable, long-term business outcome. For the Singapore startup, this means successfully scaling their operations and entering the Japanese market backed by the massive distribution networks and credibility of a major corporate partner. For the Japanese corporate, it means successfully resolving a critical business bottleneck while absorbing the entrepreneurial agility and speed of the startup.

For Singapore founders

For a Singapore startup founder who is curious about Boundary Spanning Service, what qualities or characteristics tend to make for a strong and rewarding collaboration with Japanese corporate entities? What do you find matters most?

The most critical asset a founder can bring is a commitment to mutual alignment and long-term planning. Japanese corporate partners place an immense premium on quality standards, operational stability, and meticulous planning.

Instead of viewing these requirements as bureaucratic delays, successful founders recognise them as the very foundation required to achieve sustainable, enterprise-grade scalability. A willingness to understand these operational values, combined with transparent communication and professional patience, is what truly secures a rewarding, high-yield partnership.

Also Read: Global expansion is no longer about reducing information costs, it’s about reducing trust costs

What would a founder’s journey through the Boundary Spanning Service look like from start to finish – from the initial application through to the first conversation with a Japanese corporate? What should they expect in terms of timing and engagement?

We have structured the journey to be as seamless and high-probability as possible. Through our joint effort with e27, startups can apply through a highly streamlined, low-friction application process.

From late September, Sony Acceleration Platform will start sharing and recommending the startup’s profile to the participating Japanese companies who are interested in connecting with Singapore companies.

Is there anything you’d like to say to Singapore founders who may be open to this kind of collaboration but are perhaps unfamiliar with how Japanese corporates typically work or what they value in a working relationship?

It is important to understand that the thoroughness of Japanese corporate decision-making—which often involves extensive internal consensus-building—is a structural characteristic, not a personal hurdle.

While this front-loaded alignment process takes time and structured engagement, the payoff is unparalleled. Once a Japanese corporate commits to a partner and establishes mutual trust, that relationship translates into exceptionally stable, deeply committed, and highly scalable long-term support. We encourage founders to focus on building a robust foundation of quality and trust from day one.

Looking ahead

Looking ahead, what does Sony Acceleration Platform hope the relationship between Japanese corporate innovation and Singapore’s startup ecosystem will look like over time? What would feel meaningful to you personally?

Our ultimate vision is to see cross-border co-creation transition from a “special project” into a standard, daily operational model. Personally, it would be deeply rewarding to see this corridor become the default pathway for global-scale business development.

We highly encourage forward-thinking founders to leverage the Boundary Spanning Service to position themselves at the very forefront of this evolving bilateral corridor.

—

Want updates like this delivered directly? Join our WhatsApp channel and stay in the loop.

The post Bridging boundaries: Sony’s blueprint for Japan-Singapore innovation collaboration appeared first on e27.

Southeast Asia’s startup story is usually told through ride-hailing, fintech, and e-commerce. But a quieter, stranger, and potentially more consequential sector is taking shape in the region’s labs, farms, and hospitals: biotech.

Founders are using fermentation to turn waste into biomaterials, growing seafood and fat from cells, building AI tools for cancer care and heart scans, engineering crops for tougher climates, and developing diagnostics designed specifically for Asian populations. It is not an easy sector to build in; biotech demands patient capital, deep technical talent, complex regulatory navigation, and long commercialisation cycles.

Also Read: Asia’s biotech boom: Innovation, investment, and a new era of discovery

Yet the sheer breadth of companies now emerging from Singapore, Malaysia, Indonesia, Thailand, Vietnam, and the Philippines suggests the region’s life-sciences ecosystem has moved well past the experimental stage, and is beginning to challenge assumptions about where deeptech gets built.

Below is a list of 27 biotech firms that are redrawing the region innovation map:

RWDC Industries (Singapore)

RWDC ferments used cooking oil into PHA, a fully biodegradable biopolymer that’s meant to replace single-use plastic in straws, cutlery and packaging.

Protenga (Singapore/Malaysia)

Protenga runs “Smart Insect Farms” that turn organic waste into black soldier fly protein for aquaculture, animal feed and pet food.

Engine Biosciences (Singapore/US)

Engine Biosciences combines AI with wet-lab biology to map gene interactions and speed up cancer drug discovery.

Us2.ai (Singapore)

Us2.ai uses AI to automate the reading of echocardiograms, cutting a process that takes cardiologists many minutes down to under two.
Founders: James Hare, Dr Carolyn Lam, Dr Yoran Hummel and Paul Seekings

AMILI (Singapore)

AMILI runs Southeast Asia’s first gut microbiome bank, building an Asia-specific database to power diagnostics and personalised nutrition.

KYAN Technologies (Singapore)

KYAN applies “small data AI” to match cancer patients with the most effective drug-dose combinations, developed with NUS and UCLA.

ImpacFat (Singapore)

ImpacFat cultivates omega-3-rich fish fat from stem cells, aimed at alt-meat, cosmetics and supplements.

Qarbotech (Malaysia)

Qarbotech makes QarboGrow, a carbon-quantum-dot photosynthesis enhancer that boosts crop yields without genetic modification.

NLYTech Biotech (Malaysia)

NLYTech develops biodegradable, plastic-free packaging materials made from natural ingredients as an alternative to single-use plastics.

Vulcan Augmetics (Vietnam)

Vulcan builds affordable, modular robotic prosthetics designed to click together like building blocks.

Teora (Singapore)

Teora develops biologics that manage disease in agriculture and aquaculture without relying on chemical pesticides.

KINNVA (Singapore)

KINNVA is a synthetic-biology company using fermentation to turn waste streams into biochemicals for food, feed and cosmetics.

Sinhke (Vietnam)

Sinhke builds AI-powered hardware that measures shrimp larvae health before farmers commit to large-scale cultivation.

Virdalis (Singapore)

Virdalis is building a cultivation and data platform around duckweed, the world’s fastest-growing flowering plant, as a soy alternative for animal feed protein.

Allozymes (Singapore)

Allozymes runs an ultra-high-throughput microfluidics platform that screens millions of enzyme variants a day, effectively an “enzyme discovery engine” for pharma, food and chemical industries.

FathomX (Singapore)

FathomX is an NUS/NUHS spin-off building AI to improve the accuracy of mammograms, particularly for dense breast tissue common among Asian women.

REVIVO BioSystems (Singapore)

An A*STAR spin-off building “organ-on-a-chip” 4D human skin models to replace animal testing for cosmetics and pharma compounds.

Advanx Health (Malaysia)

Malaysia’s first consumer DNA-testing company, offering genetic reports on health risk, nutrition and fitness traits.

KosmodeHealth (Singapore)

An NUS Food Science spin-off extracting proteins and fibres from food-processing waste to formulate functional foods and biomedical bio-ink.

Imagene Labs (Singapore)

A subsidiary of Asia Genomics offering saliva-based DNA testing for personalised nutrition, skincare and fitness products across Asia.

Meatiply (Singapore)

Meatiply is a multi-cell-type cultivated meat company that produced Asia’s first cultivated smoked duck breast.

Singrow (Singapore)

An agri-biotech firm that cross-bred and gene-edited the world’s first tropical-climate strawberry, grown in an indoor vertical farm.

Dendrotonics (Philippines)

Develops biodiversity-restoration technology to make degraded land ecologically and commercially productive again.

QuikPath (Singapore)

Built a self-administered RT-PCR Covid-19 screening technology designed to scale rapid, accurate infection monitoring.

ETBio (Singapore)

Harnesses microalgae to build next-generation air filtration solutions.

Ternion Biosciences (Singapore)

Provides high-throughput cardiac safety screening assays used in preclinical drug development.

CloudSeq (Singapore)

CloudSeq runs an NRF-backed cloud platform built to handle big data for healthcare and agricultural genomics.

The post The 27 SEA biotech firms betting on cells, fermentation, and code appeared first on e27.

Something strange is happening in Vietnam’s engineering hiring market right now. Companies report talent shortages. Candidates report rejection rates higher than ever. Both are telling the truth.

The gap between them isn’t a supply problem. It’s a definition problem, a fundamental mismatch between what companies say they need, what their hiring processes actually select for, and what the engineers who will matter most in the next three years actually look like.

I’ve spent a decade placing engineers across Southeast Asia. I’ve seen this kind of misalignment before. I haven’t seen it this structural.

The shift nobody fully prepared for

Eighteen months ago, the hottest debate in engineering circles was whether “Vibe Coding”, the practice of generating code entirely through natural language prompts to AI, was a legitimate workflow or a shortcut for junior developers. That debate is now over.

It turns out both sides were right, and both sides missed the point.

Yes, Vibe Coding accelerated output. A GitClear analysis of over 211 million lines of changed code found that AI-assisted workflows boosted boilerplate writing by 25–50 per cent. Yes, it also introduced a technical debt crisis: refactoring rates dropped from 25 per cent in 2021 to under 10 per cent in 2024, while duplicated code quadrupled. Over 40 per cent of junior developers admitted to deploying AI-generated code they hadn’t read.

By early 2026, Andrej Karpathy, one of the people who popularised the Vibe Coding concept, had already moved on. He began describing what comes next: Agentic Engineering. Not writing code, not prompting AI to write code, but orchestrating autonomous AI agents: setting specifications, auditing outputs, managing feedback loops, and owning architectural decisions.

The role of a software engineer is shifting from executor to decision-maker.

That shift has a direct consequence for anyone responsible for hiring.

The talent gap is real, but it’s not the gap most people think

Vietnam’s numbers are striking. The country now has 530,000–600,000 software engineers in a workforce of over 1.2 million ICT professionals. Universities produce 55,000–60,000 IT graduates per year. On paper, this looks like abundance.

In practice, demand for new technology positions exceeds 500,000 roles annually, a structural mismatch approaching 10x. And that gap is widening, not closing, for two compounding reasons.

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

• First, traditional industries entered the race. Banking, retail, and manufacturing have accelerated digital transformation, competing directly with tech companies for the same engineering talent. The pool didn’t grow; the number of teams fishing in it multiplied.
• Second, semiconductors arrived. Both Ho Chi Minh City and Hanoi are now running significant semiconductor and chip-design initiatives; HCMC alone is targeting 3,000 specialist engineers, backed by programmes at ĐHQG and SHTP. Vietnam’s national target is 50,000 semiconductor engineers at university level by 2030, supported by government scholarships worth 1,300 billion VND (US$49.4 million) annually for approximately 30,000 learners.

The consequence? The best STEM graduates are no longer choosing software development as their default. The pool of candidates that tech companies relied on for the past decade is being redirected upstream.

The result: average time-to-fill for a Senior offshore engineer has stretched to 45–60 days. Retention for engineers over two years has fallen from 78 per cent in 2023 to roughly 65 per cent in 2026. Senior IT salaries have risen 50–70 per cent compared to 2024, with significant variation depending on specialisation and language ability.

The two mistakes companies are making right now

Mistake one: Hiring for the old role

Most job descriptions I see in 2026 are still optimised to find Task-Based Coders, engineers who execute well-defined tickets, follow established patterns, and stay in their lane. The interview process tests syntax, algorithms, and framework knowledge.

But the engineers who will deliver the most value in an Agentic Engineering environment are evaluated on completely different dimensions: system design judgment, the ability to audit AI-generated outputs, risk assessment, and the capacity to make independent technical decisions under pressure. These skills don’t show up on a LeetCode score.

The irony is that the very engineers companies need most are often screening out of traditional hiring pipelines, because they’ve spent recent years developing meta-skills rather than memorising framework internals.

Mistake two: Trusting language credentials over language capability

Vietnam’s tech talent market has a well-documented phenomenon I call the Paper Certificate Trap.

Language certifications, TOEIC 850+, JLPT N2, TOPIK 5, are treated as proxies for communication ability. In practice, they measure test-taking performance under controlled conditions. I have interviewed engineers with near-perfect TOEIC scores who go completely silent the moment a client asks a follow-up question in a technical meeting.

This matters because language ability is one of the strongest economic multipliers in Vietnam’s engineering market. Engineers with professional English (B2–C1), Japanese (N3–N1), or Korean (TOPIK 4–6) command salaries 30–50 per cent higher than peers with equivalent technical experience but limited to Vietnamese. A Senior AI/ML engineer with strong English can realistically earn US$3,800–US$6,000+ per month, a meaningful difference driven entirely by the ability to negotiate architecture directly with international clients.

Companies that can’t reliably identify genuine bilingual capability are paying a premium for a credential that doesn’t reflect reality, while missing engineers who have real cross-cultural communication skills but modest exam scores.

Also Read: Great talent is what happens after AI creates the first draft

What actually works

Move to skills-based sourcing

Replace credential screening with competency screening. Define the actual decisions and judgment calls the role requires, then design your process to surface those directly.

For senior roles in an Agentic Engineering environment, the relevant competencies are: Can this person write a system specification and defend it? Can they review a diff they didn’t write and identify the architectural implications? Can they set up a feedback loop between AI agents and quality gates?

None of these appears on a CV. All of them can be assessed in a structured technical conversation.

Implement live communication audits

For any role requiring cross-timezone collaboration or direct client contact, add a real-time communication component early in your process, not a written English test, but an actual technical conversation under mild pressure.

A 20-minute session where a candidate explains a system they’ve built, fields two or three unexpected questions, and works through a hypothetical trade-off out loud will reveal more than any certification score. Done well, this eliminates the majority of candidates who present strong paper credentials but lack genuine communication fluency, before you’ve invested weeks in technical rounds.

Match your hiring model to your actual risk profile

Not all talent gaps require the same solution, and the mid-2026 environment punishes generic approaches.

Early-stage teams prioritise flexibility over headcount permanence; access to senior expertise without long-term fixed cost is often more valuable than a full-time hire at a moment when product direction is still shifting. Growth-stage companies typically benefit from a hybrid structure: a stable core for culture and IP continuity, with flexible capacity to absorb demand spikes. Vietnam’s tech market has consistent biannual attrition cycles, with July historically the highest-churn month as mid-year reviews conclude and bonuses are paid, predictable volatility that hiring plans rarely account for.

Larger enterprises and foreign-invested companies face a different constraint: the gap between standing up a dedicated engineering function and actually integrating it. Whatever structure is chosen, the critical implementation principle is the same. Cultural and operational integration (shared tooling, CI/CD pipelines, daily standups) must begin from Day 1, not at the point of handover. Teams that delay this until a later phase consistently experience attrition at precisely the moment continuity matters most.

Also Read: Great talent is what happens after AI creates the first draft

The bigger picture

The scarcity that talent leaders are experiencing in 2026 is not a temporary supply shortage. It reflects a structural reclassification of what engineering capability means and a transition period where the market hasn’t yet developed reliable signals for identifying the new kind of engineer.

The companies that hire well in this environment will be the ones that invest in building those signals themselves: clearer definitions of what “decision-ready” means for their specific context, better processes for detecting genuine bilingual capability, and hiring models flexible enough to absorb the volatility of a market where the best people have more options than ever.

The companies that don’t will spend the next 18 months paying premium salaries for engineers who looked right on paper, watching their technical debt compound quietly in the background.

I’ve seen both outcomes. The difference is almost always made before the offer is signed.

—

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 Vietnam’s tech talent market is broken and most companies are still hiring the wrong way appeared first on e27.

Thai founders sometimes ask why a Vietnamese startup with a comparable product or level of traction can appear to raise a larger funding round.

The answer is rarely that one startup is inherently better than the other.

Venture capital reflects the company being financed, but it also reflects the market surrounding it. Investors consider the size and growth of the domestic economy, the availability of follow-on capital and the likelihood of eventually selling their shares.

On these measures, Vietnam currently benefits from a stronger growth narrative.

This does not mean Vietnamese startups always raise more than Thai companies. Southeast Asian funding data remain incomplete, many transactions are undisclosed, and a few large deals can distort national totals.

But a broader distinction is visible: investors are often more willing to finance Vietnamese startups against expected growth. Thai founders are more frequently required to demonstrate regional scale before receiving comparable backing.

Investors price future growth

Venture capital is a wager on what a company could become several years from now.

That makes national economic expectations important, even when investors are evaluating an individual startup.

Vietnam’s economy expanded by 8 per cent in 2025, while the World Bank expects growth of 6.8 per cent in 2026. Thailand, by comparison, is expected to grow by about 1.6 per cent in 2026.

Economic growth does not determine whether a particular software, healthcare or logistics startup will succeed. But it affects the assumptions investors place around that company.

Vietnam offers a population of more than 100 million, rising household incomes, manufacturing expansion and growing demand for digital services. An investor can reasonably expect some companies to expand alongside the economy.

Thailand is wealthier and has stronger infrastructure in many areas. It is also home to sophisticated banks, retailers, telecommunications groups and industrial companies.

These are valuable assets for startups seeking customers and partnerships. But they can also make the venture case more difficult.

A Thai startup may need to displace established companies in a relatively mature market. A Vietnamese company may be able to grow by serving demand that is still being created.

As a result, the Vietnamese startup can sometimes receive more credit for future scale, even when the Thai company has stronger revenue today.

Also Read: Inside SEA’s AI gold rush: The 20 investors writing the biggest cheques

Market size changes the fundraising conversation

Vietnam’s population is significantly larger than Thailand’s. This gives consumer-facing companies a broader domestic market from which to build.

A Vietnamese startup can often present domestic expansion as a venture-scale opportunity. A Thai startup in the same category may be asked almost immediately about Indonesia, Vietnam, Malaysia or the Philippines.

Thailand’s market can produce substantial companies. But venture funds are not simply looking for good businesses. They need a small number of investments to generate unusually large returns across a portfolio in which many companies will fail.

This pushes investors towards businesses that can reach large markets.

For Thai founders, the result is an execution discount. Investors may believe that the domestic business is sound while assigning limited value to regional growth that has not yet been demonstrated.

This is why the first customer outside Thailand can matter so much. It shows that the company’s opportunity is not restricted by the size or maturity of its home market.

Capital follows other capital

The composition of the investor ecosystem also influences funding rounds.

Vietnam attracted nearly 150 active venture investors in 2024, according to the Vietnam Innovation and Private Capital Report. Funds from Singapore and Japan were among the most active international participants.

Funding remains difficult. Vietnamese technology startups experienced a sharp decline in investment after the global venture boom, and national private-capital figures often include large buyouts that are unrelated to early-stage startups.

The important point is not that Vietnam has unlimited capital. It is that a growing number of regional investors already include the country in their investment strategies.

Also Read: Inside Singapore’s startup boom: The 21 firms investors can’t stop funding

Venture capital depends on networks.

A seed investor wants to know who might lead the next round. A Series A investor considers whether growth funds will be available later. Every investor eventually asks who might acquire the company or purchase its shares.

When many funds already follow a market, investors know the potential co-investors, corporate buyers and later-stage funders. This makes rounds easier to assemble.

Thailand does not lack capital. It has independent funds, family offices, government programmes and a substantial corporate venture sector.

Large Thai companies can provide startups with distribution, customers, regulatory knowledge and technical expertise. Yet corporate venture capital is not always a substitute for independent institutional funding.

Corporate investors may prioritise strategic alignment over financial returns. They may avoid companies that compete with another group subsidiary or require several layers of internal approval before investing.

They may also be willing to join a round without leading it.

A lead investor sets the terms, conducts extensive due diligence and gives other investors confidence to participate. Without one, a startup may receive interest from several organisations but still fail to close a substantial round.

The shortage of investors able and willing to lead larger early-stage rounds remains one of Thailand’s most important financing constraints.

The exit question begins early

Founders often discuss exits as a distant issue. Investors consider them before making the first investment.

A venture fund earns its return when it can sell its shares through an acquisition, a secondary transaction or a public listing.

Thailand has a large stock exchange and some of Southeast Asia’s most powerful corporate groups. Yet the country has not developed a predictable exit path for venture-backed technology companies.

This can create a cycle.

Limited exits attract smaller funds. Smaller funds write smaller cheques. Startups then have less capital to expand regionally, making large exits even less likely.

Vietnam’s exit market is not mature either. Its improving public-market narrative does not yet provide a reliable listing route for technology startups.

However, Vietnam’s role in regional manufacturing, trade and supply chains gives strategic investors several reasons to acquire local technology, logistics and enterprise businesses.

Thailand has similar strengths in tourism, healthcare, food, energy, automotive manufacturing and services. The challenge is to connect these sectors to regional buyers rather than treating acquisition by a domestic conglomerate as the only possible outcome.

What Thai founders can control

Founders cannot change Thailand’s demographics, economic growth or fund structure. They can change how dependent their company appears to be on the domestic market.

Regional expansion must be presented as an operating plan, not a collection of flags in a pitch deck.

A Thai software company might follow an existing corporate client into Malaysia. A hospitality platform could expand through Thai hotel groups operating abroad. A healthcare startup could target countries with similar private hospital systems.

Internationally comparable metrics are also essential. Recurring revenue, retention, gross margin, customer acquisition costs and contribution margin help investors compare the company with businesses in other markets.

Thai founders should also approach regional investors before they urgently need capital. A fund that has followed a company for a year can evaluate its progress more confidently than one receiving a pitch shortly before the runway expires.

Most importantly, founders need to identify which investors can actually lead a round. Interest from corporate funds and smaller investors is useful, but it may not be enough to establish the valuation and bring the full syndicate together.

An expectations premium versus an execution discount

The difference between the two ecosystems is not that Vietnamese founders consistently build better companies.

Vietnam benefits from an expectations premium. Investors see a large market, faster economic growth and a growing network of international funds. They are sometimes willing to finance the scale a company may eventually achieve.

Thailand faces an execution discount. Startups are more often expected to show regional revenue, efficient economics and clear evidence that they can grow beyond the domestic market.

Both perceptions are incomplete. Vietnam remains exposed to trade disruption, regulatory risk and limited exits. Thailand has sophisticated infrastructure, strong corporations and real competitive advantages.

But investor narratives affect how capital is allocated.

Vietnamese startups can sometimes raise against the future investors expect their market to create. Thai founders are more often required to begin building that future before investors will pay for it.

—

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 Why investors often back Vietnamese startups more aggressively than Thai peers appeared first on e27.

Southeast Asia is in the midst of a workforce transformation paradox that has quietly become the region’s most pressing business challenge.

Governments have invested billions in upskilling initiatives. Singapore alone has trained over 555,000 workers through SkillsFuture programmes. Across the region—from Indonesia’s digital transformation drive to Vietnam’s emerging tech ecosystem—organisations are spending heavily on employee development. SMEs are sending their teams to AI courses, data science bootcamps, and digital literacy programmes. On paper, the workforce has never been better prepared.

Yet when these trained employees return to their jobs, something breaks.

The CEO who approved the training gets a report that the new “AI-capable” team member isn’t delivering AI-ready outputs. The employee who completed certification feels anxious despite their credential. The hiring manager who reviewed a resume with “AI Skills Certified” discovers during the onboarding period that the candidate struggles with real-world application. Nobody is lying. Everyone invested in good faith. But the signal—the credential—isn’t predicting actual capability.

This gap between certification and demonstrable capability has become the hidden cost of Southeast Asia’s digital transformation. And for SMEs, it’s catastrophic.

The training paradox: Credentials without capability

Here’s what the data reveals: training completion is not the same as job readiness. The distinction matters more than we’ve admitted.

Researchers across multiple industries have documented this phenomenon. Cloud Range’s 2025 research on technical workforce readiness is unambiguous: “Knowledge is what you learn. Readiness is what you can perform—and those are not the same. In a live incident, the difference between knowing what to do and being able to execute in real time under uncertainty is dramatic.”

This isn’t a criticism of training programmes. It’s a description of a fundamental gap between learning and performance.

Consider Google’s experience, documented by Cornerstone OnDemand. For years, the company screened job candidates using traditional credentials: transcripts, GPAs, test scores. After hiring thousands of people, Google researchers concluded these credentials were essentially “worthless” for predicting actual job performance. Only 43 per cent of workers in STEM roles even possess STEM degrees—yet those roles are filled nonetheless, suggesting that credentials and actual capability are loosely correlated at best.

In Southeast Asia, this gap has been replicated at scale. The SHRM Global Worker Project (2025) found that globally, 37 per cent of workers hold jobs that don’t align with their skills, while 53 per cent report their roles don’t match their education and training. But the regional data is more alarming: Singapore’s Ministry of Manpower and National Trades Union Congress (NTUC) study (2025) found that hiring challenges are increasingly driven by “skills specificity rather than qualification mismatches”—meaning employers aren’t struggling to find people with credentials; they’re struggling to find people with demonstrated expertise in the specific capability needed.

Translation: The credential exists. The capability doesn’t.

Also Read: “The AI did it” is not a defence; it is a confession

The hidden cost: What credential-capability mismatch actually costs

When certification becomes divorced from capability, three cascading problems emerge for organisations, particularly for resource-constrained SMEs.

• First, hiring decisions fail silently. An SME manager reviews a resume showing “AI Fundamentals Certified.” The hiring process validates the credential. The candidate onboards. Within weeks, the manager realises the person can apply frameworks in training conditions but freezes when facing real systems. The hire was made on a false signal—and SMEs, lacking large HR infrastructure, often don’t have backup plans or retraining budgets.
• Second, organisational anxiety increases. When 24.3 per cent of Singapore employers report experiencing skills gaps in their workforce, and 49.9 per cent report this causes increased workload for other staff, you’re describing a system where “trained” people can’t actually perform, forcing colleagues to compensate. The trained employee feels inadequate despite their certificate. Their manager feels misled by the training system. The organisation’s confidence in development programmes erodes.
• Third, competitive advantage evaporates. SMEs are racing to adopt AI to compete with larger rivals. But if their hiring signal—the credential—doesn’t predict whether someone can actually build AI systems, deploy models, or integrate AI into operations, they’re hiring randomly and hoping. In a competitive market, hope is a business risk.

This is where the problem reveals itself as a systems issue, not an individual or training-quality issue.

The signal integrity problem: Why credentials fail in APAC

Southeast Asia’s workforce development system has optimised for measurable completion metrics rather than capability verification:

What gets measured:

• Training volume (555,000 trained annually in Singapore alone)
• Certification completion rates
• Course quality and practical components

What doesn’t get measured:

• Can the certified person actually perform on the job?
• Do credentials predict job success, retention and performance?
• Is the certification signal reliable?

The result is a market-wide problem. When 16 per cent of specialised professional, manager, executive, and technician (PMET) roles in Singapore remain unfilled for six or more months, employers specifically cite difficulty finding people with demonstrated technical expertise—not people with credentials.

The credential system hasn’t failed because the training is poor. It’s failed because certifications and actual capability are being treated as equivalent when they’re not.

Also Read: The most sophisticated AI strategy is a puzzle hunt in Toa Payoh

The AI-powered enterprise solution: Bridging signal integrity

This is where AI-powered enterprise solutions become the game-changer for SMEs in Southeast Asia.

Traditional hiring systems can filter for credentials. They struggle to verify capability. AI-powered assessment platforms can do what neither training programmes nor conventional recruitment can: assess demonstrated capability—not just knowledge of frameworks—at scale and with consistency.

These solutions work by distinguishing between three different assessment layers:

• First, deterministic signals: Keyword and semantic analysis identify formal qualifications and technical vocabulary. Someone who says they “trained in Python” appears here. But this doesn’t prove they can debug production code under pressure.
• Second, semantic understanding: Advanced models evaluate whether someone can explain concepts in their own words, suggesting deeper comprehension than memorisation. This is closer to capability but still incomplete.
• Third, capability assessment: This is the layer most SMEs lack access to. AI-powered capability assessment goes deeper: Can this person actually do the work? Can they apply knowledge to novel problems? Can they integrate with existing systems? Will they perform in real conditions?

For SMEs, this third layer is transformative. A small team can now make hiring decisions with the same rigour a large enterprise could afford through expensive assessment centres. An SME can distinguish between “certified” and “actually capable” before hiring. They can identify which trained employees are genuinely ready for deployment in AI initiatives.

The competitive imperative for SMEs

SMEs in Southeast Asia face a unique time constraint. Larger competitors are adopting AI faster. Regulatory environments (EU AI Act, Japan’s ¥10 trillion Trustworthy AI 2030 mandate) are tightening requirements. The window to build AI-ready capability is closing.

But SMEs can’t afford to hire and fail repeatedly. They don’t have the budget to train an entire team, discover half aren’t capable, and retrain. They need to know, before hiring or promoting, whether their team members actually have the capability that their credentials claim.

Also Read: How AI is dismantling the risk pool in insurance

AI-powered enterprise assessment solutions solve this by:

• Reducing mis-hire costs: Verify capability before hiring, not after onboarding failure
• Optimising training ROI: Identify which trained employees are genuinely ready for deployment
• Accelerating AI adoption: Deploy capability with confidence rather than guessing
• Building organisational trust: When capabilities are verified, teams move faster and with less anxiety

The game-changer moment

We’re at an inflection point. Southeast Asia has solved the training problem—the region demonstrates this daily with millions of course completions. What remains unsolved is the verification problem: reliably determining who actually has capability versus who has certification.

SMEs that address this first—that adopt AI-powered enterprise solutions to verify demonstrated capability rather than relying on credentials—will outcompete peers who continue hiring blindly. They’ll deploy trained talent more effectively. They’ll build confidence in their teams. They’ll accelerate their competitive position.

The credential-capability gap that seemed like a training problem is actually an assessment and verification problem. And for the first time, AI-powered enterprise solutions make that verification affordable and scalable for organisations of any size.

That’s the game-changer Southeast Asian SMEs have been waiting for.

—

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 real workforce challenge: Bridging the credential-capability gap appeared first on e27.