For companies selling, hiring or operating across multiple countries, the promise of going global often runs into a very old problem: banking still behaves as if borders are hard walls. Opening local accounts can require entities, paperwork and long waits. Payments move through correspondent banks. Foreign exchange fees are not always clear. Finance teams end up stitching together banks, payment providers, cards, spreadsheets and treasury tools just to keep money moving.

Limited, a San Francisco-based fintech startup founded in 2024, is trying to simplify that stack. The company has grown its seed round to US$18.5 million less than ten months after launch, after existing investor Third Prime preempted the round. Singapore-based 1982 Ventures participated through 1982 Ventures Fund II, joining new backers ParaFi Capital, Pharsalus Capital, Digital Currency Group and Onigiri Capital.

Also Read: Nium acquires Cypher as fiat and stablecoin payments converge

Existing investors North Island Ventures, which led Limited’s original seed round, The House Fund and Collab+Currency also returned. Early backers Arche Capital and SevenX Ventures remain on board.

Limited offers what it describes as a global business account for multinational companies. Its platform provides business accounts across the US, EU, UK, Latin America and Africa, and supports payments to more than 170 countries. It also offers local-currency payouts in more than 60 currencies, corporate cards, spend controls, accounting integrations and stablecoin rails for real-time transfers.

The company is not a ban but a fintech building the layer that helps businesses access accounts, move money and manage spending across jurisdictions.

“Customer demand pulled this round forward,” said Hussein Ahmed, founder and CEO of Limited. “We are a lean team with strong revenue growth, so this capital is about accelerating what is already working: senior hires across go-to-market, operations and compliance, more local corridors, and deeper treasury features for larger, multi-entity companies.”

The cross-border finance gap

Limited’s pitch is straightforward: international companies still face a fragmented financial system. A business operating across Mexico, Dubai and Hong Kong, for example, may need to work with local banks in each market, manage slow wire transfers, absorb unclear foreign exchange spreads and reconcile multiple systems.

That problem is familiar in Southeast Asia. Startups in Singapore, Indonesia, Vietnam, the Philippines and Malaysia often expand regionally earlier than their US or European peers because home markets can be smaller or more fragmented. Even before they become large enterprises, many need to pay overseas suppliers, receive revenue from foreign customers, manage remote teams and move capital between entities.

Also Read: Circle to acquire Tazapay for US$400M as USDC push moves into cross-border payments

The challenge becomes sharper for companies with ambitions beyond the region. A Singapore-headquartered startup selling into the US, hiring in Latin America and sourcing from China may quickly outgrow a domestic business bank account. Traditional banks can serve these needs, but onboarding, compliance checks and account opening across markets can be slow. Newer fintech platforms are trying to win customers by collapsing that complexity into one interface.

This is where Limited wants to compete. Its model combines local accounts, cross-border payments, corporate cards and spend management, while also using stablecoin rails for faster transfers. Stablecoins are digital tokens designed to track the value of fiat currencies such as the US dollar. In business payments, advocates argue that they can reduce settlement times, especially where traditional banking rails are slow or expensive. The trade-off is that companies still need to manage regulatory, compliance and counterparty risks carefully.

Why 1982 Ventures is backing the company

For 1982 Ventures, the investment fits its focus on fintech infrastructure and financial services businesses that can scale across markets. The Singapore-based fund manager has backed Limited through its second fund, placing a Southeast Asian investor on the cap table of a US-headquartered company aiming at a global customer base.

“Hussein is a proven founder who has done this before, and it shows,” said Herston Powers, Founding Managing Partner at 1982 Ventures. “In under ten months, Limited has built business accounts across five regions and payments to 170-plus countries on a very lean team.”

Scott Krivokopich, Founding Managing Partner at 1982 Ventures, added that cross-border money movement remains “stitched together from wires, local banks and FX providers”, and that Limited is trying to put those functions into one account.

The emphasis on founder experience is notable. In fintech, especially in cross-border payments, execution is not only about product design. It also depends on licensing strategy, banking partnerships, compliance processes, transaction monitoring, fraud controls and the ability to support customers across time zones. Scaling too quickly without the right controls can create regulatory and operational risks.

Limited said the fresh capital will go towards senior hires across go-to-market, operations and compliance. It also plans to add more local corridors across Latin America, Asia Pacific and the Middle East and North Africa, and build deeper treasury features for larger companies with multiple entities.

Rivals in a crowded global fintech category

Limited is entering a competitive market with both global and regional rivals. Airwallex, founded in Australia and now a major player in Asia Pacific, offers multi-currency accounts, cards and international payments for businesses. Wise Platform and Wise Business are widely used for cross-border transfers and multi-currency accounts. Revolut Business targets companies with accounts, cards and foreign exchange tools, while Payoneer serves many exporters, marketplaces and digital businesses.

Also Read: SBI joins dtcpay’s US$25M round to bridge Japan, SEA stablecoin corridors

In Southeast Asia, Aspire has built a regional business finance platform for startups and SMEs, while Singapore’s Thunes focuses on cross-border payment infrastructure. Limited’s differentiation will depend on how well it can combine global account access, stablecoin-enabled settlement, compliance and treasury tools in a way that is reliable enough for larger multinational customers.

A broader shift in business banking

The round also points to a broader movement in fintech. The first wave of neobanks focused heavily on consumers and small businesses. The next opportunity may sit in the messy financial operations of companies that are global by default.

Remote work, cross-border commerce, global supply chains and digital services have changed how companies operate. A startup can be incorporated in one country, sell into another, hire developers in a third and raise money from investors in a fourth. But the banking infrastructure supporting that company often remains local, manual and slow.

For Southeast Asian founders, this is not an abstract issue. Regional expansion usually means navigating different currencies, regulators, banking norms and payment preferences. A company moving from Singapore into Indonesia, Thailand or the Philippines may need different local partners and workflows in each market. If it expands further into the US, Europe or the Middle East, complexity multiplies.

That makes cross-border finance infrastructure an attractive investment theme, even in a tougher funding environment. Investors have become more selective, but they continue to back fintech companies that solve clear operational problems and can show revenue traction.

Limited has not disclosed customer numbers, revenue figures or valuation. That leaves open the usual questions around early-stage fintech businesses: how defensible the product is, how expensive compliance will become, and whether it can scale without taking on too much operational risk.

Also Read: The end of manual finance? AI agents are coming for startup payments

For now, the company has secured a sizeable Seed round and a group of investors who believe the pain point is large enough to support a global business. The next test is whether Limited can move beyond early demand and become a trusted financial operating layer for companies that no longer fit neatly inside one country’s banking system.

The post 1982 Ventures joins Limited’s US$18.5M seed round to simplify cross-border business banking appeared first on e27.

Samsung Electro-Mechanics is making its largest single-product investment to date, committing US$4.9 billion to expand production of chip-packaging substrates in South Korea and Vietnam as the artificial intelligence boom reshapes demand across the semiconductor supply chain.

In two filings with the Korea Exchange dated September 28, the Samsung Electronics affiliate said it will spend about US$3.1 billion in South Korea and around US$1.8 billion in Vietnam to increase capacity for package substrates, the high-performance boards that connect advanced chips to the wider electronic systems around them.

Also Read: A*STAR and EDB unveil SG Semiconductor as partnerships target AI-era chis

The larger portion will go into new production lines at Samsung Electro-Mechanics’s Sejong plant in central South Korea, where the company will make flip-chip ball grid array substrates, commonly known as FC-BGA. These substrates are used in high-density semiconductor packages for AI accelerators, server processors and other advanced computing chips.

Construction of the South Korean expansion is expected to run until May 2028, with mass production scheduled to begin in September 2028. Separately, Samsung Electro-Mechanics’ Vietnamese subsidiary will expand its package-substrate plant in Vietnam by the end of April 2028.

The investment is sizeable even for a Samsung group company. Samsung Electro-Mechanics said the commitment is equivalent to about 44 per cent of its consolidated equity at the end of 2025, underlining how central advanced package substrates have become to its long-term strategy.

Why packaging now matters more

For years, the most visible part of the semiconductor race centred on smaller transistor sizes and more advanced chip fabrication. That remains important, but AI has pushed another part of the industry into the spotlight: packaging.

As AI models become larger and more computationally demanding, chipmakers need processors that can move huge volumes of data quickly and efficiently. That depends not only on the chip itself, but also on the substrate and package architecture that connect processors, memory and other components.

FC-BGA substrates are especially important for high-performance chips because they allow dense electrical connections, better signal performance and improved heat management. In simple terms, they are the foundation that lets powerful chips communicate with the rest of the system without bottlenecks.

Demand is being driven by AI accelerators used in data centres, as well as high-end server central processing units and graphics processing units. The rise of generative AI has led cloud providers and technology companies to spend heavily on computing infrastructure, creating pressure across the semiconductor value chain, from foundries and memory suppliers to equipment makers and packaging specialists.

Samsung Electro-Mechanics is trying to position itself in that chain. The company already makes electronic components including multilayer ceramic capacitors, camera modules and semiconductor package substrates. With this investment, it is making a clearer bet that AI-related packaging will be one of its main growth engines over the next decade.

Vietnam’s role in the semiconductor supply chain

The Vietnam portion of the investment is particularly relevant for Southeast Asia. Samsung Electro-Mechanics has been building up package-substrate production in Vietnam since 2021, adding to Samsung’s broader manufacturing footprint in the country.

Also Read: Thailand targets US$80B semiconductor push as it moves beyond assembly

Vietnam has become one of Southeast Asia’s most important electronics production hubs, helped by its labour force, export-oriented industrial zones and deepening role in global supply chains. Samsung is already one of the country’s largest foreign investors, with major smartphone and electronics operations there. A larger package-substrate plant adds another layer to that relationship, moving Vietnam further into higher-value electronics manufacturing.

For Southeast Asia, the significance goes beyond one factory. Governments across the region are trying to attract more semiconductor and advanced manufacturing investment as companies diversify supply chains beyond China and seek resilience after pandemic-era disruptions. Malaysia has long been strong in chip assembly and testing, Singapore remains a key node for semiconductor equipment and manufacturing, and Vietnam is trying to climb from electronics assembly into more specialised semiconductor-related production.

Samsung Electro-Mechanics’s expansion does not turn Vietnam into an AI chipmaking hub overnight. Substrates are only one part of a complex industry that includes wafer fabrication, advanced packaging, memory, equipment, chemicals and design. But the investment strengthens Vietnam’s claim as a serious electronics manufacturing base at a time when AI hardware demand is redrawing global supply chains.

It may also deepen the supplier ecosystem around Samsung’s Vietnamese operations. Large anchor investments often attract materials providers, logistics firms, automation specialists and component suppliers. For local companies, the challenge will be moving beyond basic support services into higher-specification manufacturing and engineering work.

Korea keeps the most advanced lines at home

While Vietnam gets a major expansion, the larger investment remains in South Korea. That is not surprising. Advanced substrate production requires precision manufacturing, tight process control and close coordination with customers building cutting-edge processors.

Also Read: Malaysia’s chip suppliers face rising pressure to prove cyber resilience

South Korea’s semiconductor ecosystem gives Samsung Electro-Mechanics access to engineering talent, materials suppliers and proximity to Samsung Electronics, one of the world’s biggest chip and electronics companies. Keeping the largest FC-BGA push in Sejong also reflects a broader pattern in the chip industry: companies may internationalise parts of production, but the most sensitive or technically demanding capacity often stays close to home.

The investment also lands amid intense competition among governments to secure advanced semiconductor supply chains. The US, Japan, South Korea, Taiwan and the EU have all pushed policies to support domestic chip capabilities. For Korea, expanding advanced materials and packaging capacity is important because its semiconductor strength has traditionally been associated with memory chips and manufacturing scale. AI has made the wider supply chain more strategically important.

Rivals are also chasing the AI packaging boom

Samsung Electro-Mechanics is not alone in chasing this market. Japan’s Ibiden and Shinko Electric Industries are major suppliers of high-end package substrates and have benefited from demand linked to advanced processors. Taiwan’s Unimicron and Nan Ya PCB are also significant players in IC substrates, while Austria-headquartered AT&S has invested heavily in high-end substrates used in servers and data centres. In South Korea, LG Innotek and Daeduck Electronics also operate in related substrate segments.

The competitive question is whether Samsung Electro-Mechanics can scale capacity while meeting the exacting quality requirements of AI chip customers, where yields, reliability and long qualification cycles matter as much as headline investment size.

A long-term bet, not a quick AI trade

The timeline shows this is not an immediate revenue boost. Construction and expansion will run through 2028, meaning Samsung Electro-Mechanics is betting that demand for AI processors and server infrastructure will remain strong well beyond the current investment cycle.

That is a reasonable but not risk-free assumption. AI infrastructure spending has surged, led by hyperscale cloud providers and large technology companies. Yet semiconductor cycles can turn quickly if customers overbuild capacity, delay data-centre projects or shift architectures. Package substrates also require significant upfront capital, and returns depend on securing long-term orders from major chip customers.

Also Read: Southeast Asia’s chip-hub ambition is colliding with its chip-smuggling problem

Still, Samsung Electro-Mechanics’ move highlights a crucial point about the AI economy: the winners will not only be model developers or chip designers. Much of the value will sit in the less visible infrastructure that makes AI computing possible — substrates, memory, power components, cooling, testing and manufacturing equipment.

For Vietnam and Southeast Asia, the investment is another sign that the region is becoming more embedded in the hardware supply chain behind AI. For Samsung Electro-Mechanics, it is a high-stakes attempt to capture a deeper role in the next phase of semiconductor growth.

The post Vietnam gains from Samsung Electro-Mechanics’s US$4.9B AI substrate expansion plan appeared first on e27.

Artificial intelligence is moving deeper into asset management, but not only through chatbots, research summaries or back-office automation. The bigger question is whether AI can help investment firms do what has become increasingly difficult in public markets: find differentiated returns at scale.

Blue Fire AI, a technology company focused on AI-driven investment management, is making that bet in Japan through a strategic commercial agreement with Mizuho Financial Group and Asset Management One (AM-One).

As part of the collaboration, AM-One will take a minority equity stake in Blue Fire AI, closing the company’s latest funding round with US$9 million in aggregate commitments.

Also Read: Why Japan’s booming AI market is harder to crack than it looks

The deal links Blue Fire AI with one of Japan’s largest financial groups and an asset manager with more than US$500 billion in assets under management. AM-One plans to offer new equity investment products powered by Blue Fire AI’s proprietary neuro-symbolic AI decision engine to institutional and retail clients in Japan.

The companies did not disclose the size of AM-One’s stake or Blue Fire AI’s valuation.

For Blue Fire AI, the partnership provides a route into one of the world’s largest pools of managed capital. For Mizuho and AM-One, it is a move to strengthen active management at a time when the industry is under pressure from passive investing, lower fees and growing scepticism over whether traditional stock-picking can consistently outperform benchmarks.

Why AI matters in active management

Active fund managers have always sold judgement: the ability to analyse companies, understand markets and identify mispriced securities before others do. The problem is that markets have become faster, information is more abundant, and many strategies that once produced excess returns have become crowded.

This has pushed asset managers to look for structural advantages. Scale helps. Proprietary data helps. So does technology that can process more information than human teams can handle on their own.

Blue Fire AI says its system enables portfolio managers to perform bottom-up fundamental analysis at scale, identify overvalued securities and generate repeatable investment insights. Bottom-up analysis refers to studying individual companies — their financials, competitive position, valuation and prospects — rather than simply making top-down calls on sectors or economies.

Also Read: AI governance is moving from promises to proof

The company describes its technology as a neuro-symbolic AI decision engine. In simple terms, neuro-symbolic AI combines the pattern recognition associated with machine learning and neural networks with more structured reasoning systems. In investment management, the appeal is that such systems may be able to analyse large volumes of data while still offering a more explainable framework than purely black-box models.

That explainability matters. Institutional investors, regulators and investment committees are unlikely to be comfortable with strategies that cannot be interrogated. Asset managers using AI need to show not only that a model works, but also why it reaches certain conclusions, how risks are controlled and how decisions fit within fiduciary responsibilities.

Blue Fire AI says it has spent ten years developing its technology and has a seven-year live investment track record. That history is important in an industry where many AI claims remain untested across cycles.

Japan’s asset management opening

The partnership comes at a significant moment for Japan’s financial industry. The country has been trying to make better use of household savings, encourage investment and strengthen Tokyo’s role as a global financial centre. Policy changes such as the expansion of Nippon Individual Savings Accounts have helped push more retail money into markets, while corporate governance reforms have drawn renewed foreign investor interest in Japanese equities.

At the same time, Japan’s asset managers face the same pressures seen globally. Passive funds and exchange-traded funds have reduced fees across the industry. Large global firms have used scale to compete aggressively. Retail and institutional clients are asking harder questions about performance, cost and differentiation.

AM-One, established in 2016 and backed by major Japanese financial institutions, sits at the centre of this shift. With approximately JPY80 trillion (more than US$500 billion) in assets under management across institutional and retail businesses as of December 31, 2025, it has the distribution reach to bring AI-enhanced investment products to a broad client base.

Noriyuki Sugihara, President and CEO of AM-One, said the firm plans to use Blue Fire AI’s capabilities to enhance its investment solutions and make them available through AM-One’s product platform.

“This partnership reflects a shared conviction that the next era of active management will be built by firms willing to combine deep institutional expertise with genuinely differentiated technology,” said Luke Waddington, CEO of Blue Fire AI.

Why Southeast Asia should watch

Although the deal is centred on Japan, it carries lessons for Southeast Asia’s financial ecosystem. Singapore, in particular, has built itself into a regional wealth and asset management hub, with global managers, family offices, private banks and fintech companies using the city-state as a base for Asia.

Also Read: Japan is moving into Southeast Asia faster than the West, and most brands haven’t noticed yet

Across Southeast Asia, asset managers are also facing fee pressure, rising client expectations and the need to offer more sophisticated products. Markets such as Singapore, Malaysia, Thailand and Indonesia have growing pools of retail investors, pension money and institutional capital, but local managers often compete against global firms with deeper research budgets and technology platforms.

AI could narrow some of that gap if applied carefully. A regional manager covering hundreds of listed companies across Southeast Asia may not have the same analyst headcount as a global asset manager. Tools that scale fundamental research, flag valuation anomalies and organise company-level data could become useful, especially in less-covered markets where information is fragmented.

But the Japan example also shows that distribution and trust remain critical. Blue Fire AI is not entering the market alone; it is partnering with Mizuho and AM-One, institutions with established client relationships and regulatory credibility. Southeast Asian AI-fintech startups aiming to sell into asset management may need similar partnerships with banks, brokerages, insurers or licensed fund managers rather than trying to bypass the existing system entirely.

Rivals in AI investing

Blue Fire AI operates in a growing field of investment technology companies applying AI and data science to portfolio management. Global players such as BlackRock have long used technology platforms, including Aladdin, to support risk and portfolio analytics. Firms such as Two Sigma and AQR have built quantitative investment businesses around data, models and systematic decision-making, though they are not direct product equivalents.

In the AI investment tools market, companies including Boosted.ai, Auquan and Toggle AI provide machine learning-driven research and analytics for investment professionals. In Asia, South Korea’s Qraft Technologies has developed AI-powered investment strategies and exchange-traded funds. Blue Fire AI’s challenge will be to prove that its neuro-symbolic approach can translate into durable performance inside large institutional product platforms.

The next test: performance and governance

The promise of AI in active management is compelling, but the bar is high. Investment products are ultimately judged by performance, risk management, transparency and client outcomes. A model that works in one market regime may struggle in another. Data quality can vary. AI systems can overfit, meaning they appear powerful in historical testing but fail in live markets.

Also Read: StashAway acquires MakeGoodwill to add digital wills to its wealth platform

There are also governance questions. Asset managers must decide how much authority to give AI systems, how human portfolio managers should use model outputs, and how to explain decisions to clients and regulators. For retail investors, the language around AI can easily become marketing unless firms are clear about what the technology does and does not do.

That may be why the partnership between Blue Fire AI and AM-One is framed around combining institutional expertise with differentiated technology, rather than replacing human managers outright. The more realistic future of AI in asset management is not fully autonomous investing, but augmented investment teams that can examine more companies, test more ideas and respond faster to changing market conditions.

Further details of the collaboration are expected later. For now, the agreement gives Blue Fire AI a powerful Japanese partner, gives AM-One a stake in an emerging investment technology platform, and signals that the next fight in active management may be as much about data and AI infrastructure as it is about traditional investment judgement.

The post Blue Fire AI closes US$9M round with AM-One stake under Mizuho partnership appeared first on e27.

Anthropic has told prospective investors that it is growing at a pace few companies have matched, losing money at a scale few could survive, and building a product that could pose an existential threat to humanity. The prospectus behind what could become the largest IPO on record shows revenue rising twelvefold to nearly US$4.6 billion in 2025, while an operating loss topped US$8 billion as operating expenses approached US$13 billion.

The trajectory has steepened since: second-quarter 2026 revenue reached US$11.5 billion, and the company is on course for a second consecutive quarter of adjusted operating profit. Backers believe it could list above US$2 trillion, more than twice its US$965 billion May valuation.

The filing details plans to spend US$518 billion on cloud, compute and infrastructure, and flags that two customers generated nearly a quarter of last year’s revenue. Close to a third of the document covers risk factors, including model behaviour such as resisting shutdown, manipulating information and conduct resembling blackmail.

For Southeast Asian founders and investors, the listing will set a public benchmark for frontier-AI valuations and compute economics, and a reminder that the companies supplying the region’s AI stack now describe safety as a material business risk.

REGIONAL

GoTo shares drop 14% after Indonesia scraps price floor: Indonesia’s decision to remove the stock price floor triggered an immediate sell-off in GoTo, one of the country’s most closely watched listed tech companies, raising fresh concerns about retail investor protections.

Governance gaps slowing cloud and regtech adoption: KPMG: A KPMG Singapore report finds that unclear internal accountability and weak data governance frameworks are the primary barriers preventing financial institutions across the region from scaling cloud and regtech solutions.

IFC takes US$20M stake in Axiata-backed Boost to scale digital lending: The World Bank Group’s private-sector arm brings development-finance discipline to a Malaysian fintech tied to Boost Bank, Axiata’s venture with RHB, as regulators scrutinise how lenders underwrite thin-file SMEs.

VinFast folds R&D spin-off back in, lifting charter capital to US$8.2B: Manufacturing arm VFTP will absorb Tuong Lai‘s assets and debts, a year after the 2025 carve-out moved factories and liabilities off the Nasdaq-listed carmaker’s books. VinFast took 42% of Vietnam’s August car sales.

Temasek buys 9% of Italy’s FSI in push for European mid-market deals: The stake sits in the US$5.7B manager’s management company, alongside commitments to future funds, as Temasek aims to deploy US$15.9B-19.3B across EMEA by 2029 after investing US$14.8B in two years.

Igloo cuts net loss 60% as revenue climbs 46% with costs held flat: The Singapore embedded-insurance platform posted SGD80.9M (US$63M) in FY2025 revenue and a US$6.7M loss, and targets adjusted EBITDA breakeven by end-2026, though it has not disclosed margins, cash or claims ratios.

1982 Ventures joins Limited’s US$18.5M seed for cross-border banking: Third Prime pre-empted the round for the San Francisco fintech, which offers business accounts across five regions, payouts in 60-plus currencies and stablecoin rails; it will add corridors in Asia Pacific and the Middle East.

Japan’s Kawaijuku backs Do Ventures Fund II for Vietnam education push: The Tokyo education group’s undisclosed commitment gives it a window on Vietnamese startups for future partnerships and direct deals; the Ho Chi Minh City VC’s first fund targeted US$50M with NAVER, Sea and Vertex backing.

Antom reshuffles SEA leadership to knit 2C2P and DOKU into one stack: Ant International’s merchant-payments unit named DOKU co-founder Nabilah Alsagoff regional product head, ex-DOKU CEO Chris Yeo Philippines head and Himelda Renuat DOKU CEO, pursuing a unified product roadmap across its regional brands.

INTERVIEWS AND FEATURES

Korea’s Autonomous A2Z bets on fixed-route shuttles, not robotaxis: The full-stack Level 4 developer, first Korean firm to clear Singapore LTA’s Milestone 1, ran a Grab staff shuttle pilot and signed a US$6.8M UAE supply deal; its CSO explains why localisation is the hard part.

INTERNATIONAL

Peak XV lifts Surge seed cheques to US$5M as the Series A bar rises: The firm put over US$50M into its 18-startup Surge 12 cohort; 13 target global markets though more than half are India-based, and Rajan Anandan says deeptech founders are raising bigger seed rounds.

Meta launches enterprise AI unit, poaches MongoDB CEO CJ Desai to lead: Meta Enterprise Platform will sell Muse, Meta Business Agent, Muse API and Muse Code to companies, a bid to monetise heavy AI spending; MongoDB shares fell over 17% on the abrupt exit.

Blue Fire AI closes US$9M round as AM-One takes stake in Mizuho tie-up: AM-One, which manages over US$500B, will offer equity products built on the startup’s neuro-symbolic decision engine in Japan, a partnership model Southeast Asian AI-fintechs selling into asset management may need to copy.

TikTok settles Alabama addiction case for at least US$100M: The payout could reach US$300M under certain conditions, and TikTok will impose a two-hour daily limit for minors plus overnight and filter curbs, weeks after a US$400M child-privacy settlement with the DOJ.

CYBERSECURITY

OpenAI’s misalignment log reveals a DNS sandbox escape and AI ‘worms’: The new site lists nine incidents, including a model that smuggled a GitHub token and self-replicating prompt injections; Axios reports major labs have logged up to 10,000 cases of models exceeding evaluator instructions.

Truecaller opens web Scam Checker, with Southeast Asia on its roadmap: The free tool needs no sign-in and checks numbers, links and messages against community reports and risk data, launching in India first as telcos, Apple and Google chip away at its caller-ID business.

FBI tells staff ShinyHunters breach exposed their personal data: The group exploited an Oracle PeopleSoft flaw behind the FBIJobs.gov portal, reportedly taking medical and psychiatric records too; a Lawfare analyst calls it a counterintelligence disaster exposing personnel to foreign profiling.

SEMICONDUCTOR

VSMC opens US$7.8B Singapore fab in record time for specialty chips: The Vanguard-NXP venture’s Tampines plant will make 44,000 wafers a month on 40-130nm nodes by 2029 for automotive and industrial uses, creating about 1,600 jobs, three-quarters of them professional or technical.

Samsung Electro-Mechanics steers US$1.8B of AI chip bet to Vietnam: The US$4.9B plan, its largest single-product investment and 44% of its equity, keeps advanced FC-BGA lines in Sejong, Korea, while expanding the Vietnam substrate plant by April 2028 as AI accelerators drive packaging demand.

Thailand approves US$80B chip strategy with an 86,600-worker target: The three-phase plan bets on photonics, power chips and sensors, yet the headline figure rose over US$6B since January without a new fab; BOI applications worth US$27.2B are pledges, not capital spent.

Singapore brands its chip sector SG Semiconductor, backed by US$626M: A*STAR and EDB launched the national identity with partnerships spanning Applied Materials, KLA, STATS ChipPAC and GlobalFoundries in packaging, photonics and yield; the test is whether R&D turns into manufacturable technology.

AI

AMD buys Fei-Fei Li’s World Labs for US$8.2B to bet on physical AI: The spatial-intelligence lab builds models that generate and simulate 3D worlds for robotics; Li becomes AMD’s chief scientist reporting to Lisa Su, giving the chipmaker in-house insight into next-generation AI workloads.

OpenAI shelves Astra 6.1 after model shows higher levels of deception: Safety chief Saachi Jain told the WSJ it tested poorly on alignment; critics argue the industry’s safety push may also entrench frontier labs at the expense of smaller rivals.

Nvidia pitches hardware guardrails to keep rogue AI agents in the box: Its Open Agent Safety Platform pairs OpenShell software with Sentry, a monitor on BlueField-4 chips that quarantines escaping agents; Anthropic and Microsoft signed on, OpenAI did not, and Nvidia opposes a development slowdown.

Ropedia opens its physical-AI data kit to university researchers: The NTU spin-off’s head-mounted HOMIE Gen2 records video, motion and depth as people handle objects for robotics and world-model research; its academic network includes Princeton and Carnegie Mellon.

THOUGHT LEADERSHIP

AI safety is shifting from tech policy to national security: Shawn Balakrishnan reads UK calls to ban superintelligence, and Amodei, Altman and Musk backing a slower frontier, as signs of looming capability-based rules; Singapore’s AI assurance work could help shape verification standards.

SIA runs 160 AI apps; the harder question is what they may do: John Tan urges airlines to govern the point where AI recommendations become transactions, via authority registers, human checkpoints and tamper-evident logs, while keeping safety-critical systems under statutory oversight.

Meta’s Muse agent threatens brands built on reach and habit: SOMIN’s Aleks Farseev argues AI agents pick the brand that plainly answers a stated need, leaving a sponsored slot beneath; WhatsApp-heavy Southeast Asia, used to delegating via super-apps, may feel it first.

Singapore funds scale-ups with substance, not company registrations: Gerald Yap says foreign founders should pick their Singapore strategy before the grant, noting EDB’s RIS(C) and Refundable Investment Credit reward R&D and hiring, while EnterpriseSG schemes need 30% local ownership.

Your startup’s rival for VC money may be the fund’s own portfolio: Jun Yan notes only four SEA-focused VC funds closed in 2025, down from 33 in 2023, so managers weigh each new cheque against follow-on reserves for companies they already own.

SEA’s next unicorn may be a one-person company, not the next Grab: Astrid Dang argues AI agents let lean teams in Vietnam, Indonesia and the Philippines build products once reserved for well-funded Silicon Valley firms, as investors swap scale metrics for revenue per employee.

Past 1.5°C, Southeast Asian firms must fund adaptation, not pledges: Adam Goulston cites SM Investments’ climate-driven data-centre exit and ACEN’s early coal retirement to argue verifiable capital spending, not net-zero banners, separates real ESG from theatre; resilient infrastructure returns about US$4 per dollar.

Bitget hack and 5% yields put Bitcoin’s US$82,000 support to the test: Anndy Lian links the selloff to a US$387.5M exchange theft possibly tied to Lazarus and a US$120M long liquidation wave; a daily close below US$82,000 could open a slide to US$77,000.

Bitcoin dominance at 58.5% keeps a full altseason on hold: With ETF inflows of up to US$3B pulling capital into Bitcoin first, Anndy Lian says the Altcoin Season Index at 57-70 signals selective rotation; he wants dominance below 55% and the index above 75.

Why credential-free outsiders may win in signal intelligence: Faheem Aizat Kamsan profiles a Singaporean paramedic building a timestamped, proof-logged signal platform for retail users, arguing institutions’ real moat was timing and that it is eroding.

The post Ecosystem Roundup: Anthropic’s IPO filing: 12x revenue, US$8B loss and an extinction risk appeared first on e27.

A founder pitched me earlier this year on his semiconductor company. The deck was clean. The market was real. The technology was genuinely differentiated. The Singapore entity was properly incorporated, the cap table looked orderly, and the revenue was growing.

Then I asked him three questions.

Where does your intellectual property legally sit? Who owns the entity that owns it? And if I wire money into your Singapore company tomorrow, what exactly am I buying?

He could not answer any of them cleanly. The IP sat in China. The Singapore entity he was raising on owned almost nothing of substance. The structure he was pitching and the business he was running were two different things, connected mostly by hope.

The raise was over before it started. Not because the company was bad. Because he had prepared the wrong story.

Founders prepare the product story. Investors underwrite the structure story

Most founders raising across borders spend their preparation time on the things they can see: the product, the traction, the market size, the team slide. These matter. But they are not where a cross-border raise actually succeeds or fails.

A domestic investor and a cross-border investor are not doing the same job. A domestic investor backing a Singapore company operating in Singapore can largely take the entity at face value. The company is incorporated where it operates, the revenue is earned where it is booked, the assets sit where the company sits. The investor underwrites the business.

A cross-border investor cannot do that. When the founder is in one jurisdiction, the IP in another, the revenue booked in a third, and the holding company in a fourth, the investor is no longer underwriting the business. They are underwriting the structure. And if the structure does not hold up to scrutiny, the quality of the underlying business becomes irrelevant, because the investor cannot safely own a piece of it.

This is the single most common reason promising Southeast Asian companies fail to close cross-border rounds. Not weak fundamentals. Unprepared structure.

The environment has made this sharper. The eFishery accounting fraud reset diligence standards across the region. Beijing’s unwinding of a two-billion-dollar acquisition of a Chinese-founded, Singapore-headquartered AI company put every cross-border structure under brighter light. Investors who two years ago might have taken a Singapore wrapper at face value now open it and look inside. Founders who have not looked inside it themselves get caught.

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

The questions to be able to answer before you pitch

If you are raising from an investor outside your home jurisdiction, you should be able to answer each of these without hesitation, with documents to back them.

• Where does your IP legally sit, and who owns it?

If your patents, code, or core technology are held by an entity other than the one you are raising on, the investor is buying a company that does not own its own product. This is fixable, but only before the raise, not during diligence.

• Can an investor independently verify your overseas revenue?

Revenue that flows through entities or jurisdictions an investor cannot diligence is revenue an investor will discount to zero. If a meaningful share of your traction sits in a market where contracts, banking, and customers cannot be verified, prepare to prove it or prepare to lose credit for it.

• Who really owns what across your cap table and holding structure?

Layered holding companies, nominee arrangements, and undocumented founder agreements are not red flags because they are illegal. They are red flags because they signal the founder either does not understand their own structure or is hoping the investor will not ask. Both end the conversation.

• Are your intercompany flows arm’s length?

If money moves between your entities in ways that inflate revenue, shift costs, or would not survive a transfer-pricing review, an investor’s lawyers will find it. Find it first.

• What happens to your structure if regulators act?

If a regulator in any jurisdiction you touch changed its stance tomorrow, what happens to your ownership, your IP, and your ability to operate? If you have never asked this question, you are not ready to raise across borders.

None of these are product questions. All of them are structure questions. And the founders who close cross-border rounds are the ones who have answered them before the investor asks.

What this looks like from the other side of the table

For the investors reading this, the same checklist is the discrimination that separates a real cross-border thesis from a hopeful one.

Underwriting a cross-border deal is not about liking the product. It is about being able to answer one question for your own LPs: what, precisely, am I buying, and can I defend my ownership of it if scrutiny comes? A founder who can walk you through their IP ownership, their verifiable revenue, their clean structure, and their regulatory exposure is not just better prepared. They are demonstrating the exact discipline that predicts whether the company can be owned, scaled, and eventually exited across borders.

Also Read: How a cross-border tech team built a fintech MVP in 3 months

The founders who cannot are not necessarily running bad businesses. They are running businesses that have not yet been built to be owned by someone in another jurisdiction. That is a different problem from product-market fit, and capital does not solve it.

This is the lens we apply to every company we look at across the markets we work in. The strongest signal in a cross-border pitch is rarely the product. It is whether the founder has done the structural work to be investable by someone who is not sitting in the same country.

Prepare the story that actually gets underwritten

The semiconductor founder I turned away was not a weak operator. He had built something real. But he had prepared to be evaluated as a product, when he was going to be evaluated as a structure. By the time he understood the difference, the conversation was over.

If you are planning to raise across borders, prepare both stories. The product story gets you the meeting. The structure story gets you the money.

—

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 cross-border due diligence questions most founders cannot answer appeared first on e27.

For years, Southeast Asia’s insurtech promise has rested on a simple idea: insurance should be bought where people already spend, borrow, shop, travel or top up their phones. The harder part has been turning that distribution advantage into a business that can scale without burning ever larger amounts of capital.

Singapore-headquartered Igloo is now trying to show that the model can move closer to profitability.

The company’s audited accounts for the year ended 31 December 2025 show revenue rising 45.9 per cent year on year to SGD80.9 million (~US$63 million). Net loss narrowed 60.4 per cent to US$6.7 million, from US$17 million in FY2024.

Also Read: Igloo expands Thailand footprint with Eazy Digital acquisition amid insurance reform push

The top-line growth is notable, but the more important signal sits underneath it. Igloo said revenue rose by US$19.8 million while total operating expenses remained broadly flat.

In other words, the company claims it added scale without adding cost at the same pace. That is the operating leverage many venture-backed technology companies have been under pressure to prove since the funding market cooled.

Igloo is targeting adjusted EBITDA breakeven by the end of 2026, with revenue growth continuing and no material increase in operating expenditure.

“Revenue grew 46 per cent year on year while OPEX stayed much the same,” said Raunak Mehta, co-founder and CEO of Igloo. “The way we’ve designed our operating system for insurance means that the cost of serving the next partner and the next million policies keeps falling. We are targeting adjusted EBITDA breakeven at the close of 2026.”

From insurance distributor to infrastructure layer

Igloo describes itself as an “operating system for insurance” in Southeast Asia. In practical terms, it provides the technology that allows insurers, digital platforms and financial institutions to build, distribute and manage insurance products online.

That puts the company in the embedded insurance market, where coverage is offered inside another customer journey. A shopper may buy device protection at checkout, a driver may access accident cover through a mobility platform, or a gig worker may receive microinsurance through a fintech or telecoms app. The product is insurance, but the point of sale is often not an insurer.

This model is particularly relevant in Southeast Asia, where insurance penetration remains low across many markets and traditional agency-led distribution can be expensive. The region’s large digital platforms, mobile-first consumers and fragmented regulatory landscape create both the opportunity and the complexity for companies such as Igloo.

Igloo operates across Indonesia, the Philippines, Thailand, Vietnam and Malaysia, with technology centres in China and India. It says its platform processes more than 100 million policies a month and has facilitated more than 2.2 billion policies cumulatively. Its partners include Shopee, Lazada, Tokopedia, GCash and Telkomsel, alongside more than 100 commercial and insurer partners.

Also Read: PolicyStreet’s US$21M raise signals a shift from insurtech hype to infrastructure reality

The company also runs Igloo Tech Solutions, which licenses its modular technology stack to insurers and enterprises. The aim is to shorten insurance product launch cycles from months to days by digitising product configuration, underwriting, claims adjudication and financial reconciliation.

Why flat costs matter

Igloo attributes its FY2025 performance to operating leverage in its embedded insurance business. The company said partnership volumes scaled without a proportionate increase in fixed costs, helped by what it calls AI-native infrastructure.

The phrase can sound vague, but the business logic is straightforward. If product setup, partner operations and claims processing can be automated, Igloo can serve more platforms and more policyholders without hiring large teams for every new product or market.

That matters in insurtech because distribution scale alone does not guarantee profitability. Companies still need to manage integration costs, customer support, claims workflows, compliance and reconciliation with insurers and partners. If each new partnership requires a heavy manual build, growth becomes expensive. If those functions are repeatable through software, margins can improve over time.

Igloo’s reported net loss includes US$1.6 million in non-cash share-based compensation, down from US$3 million in FY2024. It also includes US$860,000 in foreign exchange translation losses. These items do not erase the loss, but they suggest the underlying cash profile may be improving faster than the statutory bottom line shows.

Still, the company has not disclosed gross margins, cash balance, adjusted EBITDA figures, claims ratios or quarterly performance. Those numbers would give a clearer view of how close the business is to sustainable profitability, and whether growth is spread evenly across markets or concentrated in a handful of major partners.

A tougher market for insurtech

Igloo’s improved numbers come at a time when Southeast Asian startups are being judged less on expansion narratives and more on capital efficiency. During the peak of the funding cycle, insurtech companies could raise large rounds on the promise of digitising a vast underinsured population. Today, investors are asking whether those models can survive lower liquidity, higher scrutiny and slower follow-on funding.

Igloo has raised more than US$100 million from investors, including Eurazeo, Openspace Ventures, Cathay Innovation and BlueOrchard. That backing gives it room to build across markets, but it also raises expectations. A path to adjusted EBITDA breakeven by end-2026 is therefore not just a financial milestone; it is a credibility test for the embedded insurance category in the region.

The competitive field is also active. Singapore-founded bolttech is one of the most prominent global insurtech platforms with a strong Asia presence, while Australia-born Cover Genius works with digital companies worldwide on embedded protection. In Indonesia, PasarPolis has long focused on microinsurance and digital distribution, while Qoala operates across Southeast Asia with an agent-assisted and digital insurance model. Igloo’s differentiation lies in its infrastructure-led pitch and deep platform partnerships, but rivals are chasing the same broad shift: making insurance available through everyday digital channels rather than traditional sales routes.

The next test

For Igloo, the next 12 months will be about proving that FY2025 was not a one-off improvement. Revenue growth of 45.9 per cent is strong, but the company’s more consequential claim is that it can keep expanding without a material rise in operating expenditure.

Also Read: Health, wealth, and legacy planning converge as new wave of SEA insurtechs emerges

If it reaches adjusted EBITDA breakeven by the end of 2026, Igloo would stand out in a sector where many players have struggled to balance growth, regulation and unit economics. If it misses, investors will likely look more closely at the cost of partner acquisition, market-level profitability and dependence on large distribution channels.

For now, the audited FY2025 accounts show a company moving in the right direction: bigger revenue, smaller losses and a clearer profitability target. In Southeast Asia’s still-developing insurtech market, that may be the most important policy Igloo is trying to underwrite.

The post Igloo narrows FY2025 loss as embedded insurance bet edges closer to breakeven appeared first on e27.

Singapore is putting a clearer name and sharper frame around one of its most important industrial bets.

On Friday, the city-state announced SG Semiconductor, a national identity for its semiconductor sector, jointly developed by A*STAR and the Singapore Economic Development Board (EDB). The move is not a new agency or a standalone company. Rather, it is an attempt to make Singapore’s chip capabilities easier to understand, navigate and sell to global companies, researchers and talent at a time when semiconductors have become central to economic strategy.

The first focus will be public sector research and development (R&D). That is a telling choice. Singapore is already a major manufacturing base, but the next phase of the global chip race is increasingly about whether countries can connect lab work, pilot production, advanced manufacturing and commercial scale in one ecosystem.

Also Read: Southeast Asia’s chip-hub ambition is colliding with its chip-smuggling problem

SG Semiconductor is meant to package that proposition under a single national banner.

The initiative brings together capabilities across R&D, advanced manufacturing, infrastructure, industry partnerships and talent. It covers seven technology areas: advanced packaging, silicon photonics, power electronics, radio frequency gallium nitride, piezoelectric micro-electro-mechanical systems, flat optics and integrated circuit design.

Put simply, these are not consumer-facing technologies. They sit deep inside the devices, networks, vehicles, data centres and industrial systems that now power the digital economy. They matter because artificial intelligence, high-performance computing, electrification and connected machines are placing heavier demands on chips: they must process more data, consume less energy, communicate faster and fit into more complex systems.

A national brand for a strategic industry

Singapore’s semiconductor story is not new. The country has spent nearly six decades building a base that spans chip design, wafer fabrication, assembly and testing, semiconductor equipment, materials development and R&D.

Today, it accounts for one in ten chips produced globally and one-fifth of global semiconductor manufacturing equipment output, according to EDB. Those figures explain why chips remain one of Singapore’s strongest anchors in advanced manufacturing, even as neighbouring economies across Southeast Asia court electronics and chip-related investment.

Malaysia, Vietnam, Thailand and the Philippines are all strengthening parts of the semiconductor value chain, especially in assembly, testing, electronics manufacturing and supply-chain diversification. Singapore’s pitch is different: it wants to sit closer to the frontier of R&D, engineering, process innovation and high-value manufacturing.

That positioning has become more important as the global semiconductor industry reorganises around supply-chain resilience and technological sovereignty. The US, China, Japan, South Korea and Europe are pouring capital into chip capacity and research. For a small country like Singapore, the challenge is not to outspend them, but to remain a trusted, specialised node where companies can develop and scale complex technologies for global markets.

“Singapore’s strength in semiconductor innovation has been built through decades of sustained investment and close partnership across public research, universities and industry,” said Beh Kian Teik, CEO of A*STAR. “SG Semiconductor brings this collective endeavour under a national identity.”

Also Read: Nexstrom lands US$12M to bring 2D semiconductors to 12-inch wafers

That national identity is backed by money. Singapore has committed SGD800 million (about US$626 million) from 2026 to 2030 through the Research, Innovation and Enterprise (RIE) Flagship in Semiconductors. The funding is intended to deepen capabilities and tighten the link between research and industry.

From lab work to manufacturing scale

The launch was accompanied by a series of partnerships announced at Innovate Together 2026, offering a glimpse of how SG Semiconductor is expected to work in practice.

In advanced packaging, Applied Materials and A*STAR will move into Phase 4 of their long-running collaboration, expanding the joint laboratory’s infrastructure, equipment and headcount. Advanced packaging has become critical because chip performance is no longer improved only by making transistors smaller. Increasingly, companies are combining multiple chips, memory components and optical links in sophisticated packages to deliver more computing power and energy efficiency.

A*STAR and KLA will also establish a new process control collaboration framework to explore ways to improve manufacturing reliability and yield. In chipmaking, yield — the share of usable chips produced from a wafer — can determine whether a technology is commercially viable.

A*STAR and STATS ChipPAC will work together on co-packaged optics, a technology that brings optical communication components closer to computing chips. The goal is to create a pathway towards high-volume production. This is especially relevant for AI data centres, where moving data quickly and efficiently between chips and servers is becoming one of the biggest bottlenecks.

In silicon photonics, A*STAR and GlobalFoundries will deepen their R&D collaboration to develop next-generation technologies on 300 mm wafers in Singapore. Silicon photonics uses light, rather than only electrical signals, to transmit data. It is increasingly important for high-performance computing and data communications, where speed and energy efficiency are both under pressure.

There is also a sensing angle. Tacta Systems and ASTAR will develop intelligent sensing technologies for robotics using ASTAR’s Lab-in-Fab platform, which combines research and manufacturing capabilities for prototyping and validation. Tacta has also opened its Singapore operations, adding to the country’s technology development base.

Another collaboration brings together the National Center for Advanced Integrated Photonics, hosted at Nanyang Technological University (NTU), and Battery Age Minerals. They will explore germanium-based devices for data communications, linking NTU’s research expertise with Battery Age Minerals’ access to raw germanium. The partnership points to a broader ambition: connecting upstream materials with higher-value semiconductor applications.

Why it matters for Southeast Asia

For Southeast Asia, Singapore’s move comes at an important moment. The region is benefiting from multinational companies diversifying manufacturing beyond China, but much of the opportunity has so far centred on production, assembly and supply-chain redundancy. Singapore is trying to show that Southeast Asia can also play a deeper role in semiconductor innovation.

That matters for startups, too. While chip companies are capital-intensive and harder to build than software firms, demand is rising for enabling technologies around AI infrastructure, robotics, sensors, power electronics, mobility and industrial automation. A stronger research-manufacturing bridge in Singapore could create more openings for deep-tech startups, corporate spinouts and university-led commercialisation.

EDB Managing Director Jermaine Loy said companies are looking for locations where innovation, manufacturing and talent come together as semiconductor technologies become more complex. SG Semiconductor, he said, is intended to provide a “gateway” to Singapore’s ecosystem.

Also Read: Taiwan bets on Gen Z founders to move beyond its chip-supplier image

The word gateway is doing a lot of work. Singapore cannot be all things to the global chip industry. But it can make itself easier to access for companies that need specialised R&D, trusted manufacturing partners, process expertise and regional connectivity.

The test for SG Semiconductor will be whether it becomes more than a branding exercise. Its success will depend on how quickly research projects turn into manufacturable technologies, how effectively talent is developed and retained, and whether Singapore can keep attracting global partners despite intensifying competition.

For now, the message is clear: Singapore wants its semiconductor sector to be seen not just as a manufacturing base, but as a place where the next generation of chip technologies can be built, tested and scaled.

The post A*STAR and EDB unveil SG Semiconductor as partnerships target AI-era chis appeared first on e27.

Nine months ago, Thailand unveiled its first national semiconductor roadmap, a 25-year plan to graduate from backend electronics work to high-value chipmaking. This month, Bangkok approved its first national semiconductor and advanced electronics strategy, again. The headline number has grown, though: roughly US$80 billion in cumulative investment and more than 230,000 new jobs by 2050.

The repetition is less odd than it sounds. January’s announcement set out the vision; the version now cleared by the National Semiconductor and Advanced Electronics Policy Board, chaired by Deputy Prime Minister and Finance Minister Ekniti Nitithanprapas, puts formal targets and a workforce programme behind it, according to the Thailand Board of Investment (BOI). But the drift is worth noting. In January, investment figures of around US$73.5 billion were doing the rounds. That number has since grown by more than US$6 billion without a single new fab breaking ground.

Also Read: 15 Southeast Asian semiconductor startups moving beyond assembly

Such is the nature of 25-year industrial plans: the targets are aspirational, execution is never guaranteed, and the people who announce them are rarely in office when the deadline arrives.

Starting where Thailand already has credibility

To its credit, the strategy does not pretend Thailand can become the next Taiwan. It runs in three phases. Until 2030, the focus is on strengthening the country’s existing assembly and testing base while moving into advanced packaging — the increasingly valuable craft of combining multiple chips into one compact module, and a frontline of the AI hardware race. The same phase is meant to lay the groundwork for front-end wafer production, the capital-hungry business of fabricating chips on silicon that demands cleanrooms, uninterrupted power and water, and a deep bench of engineers.

By 2040, Bangkok hopes to have attracted chip design and wafer fabrication. By 2050, it wants a complete domestic supply chain.

The three technology bets are the plan’s most sensible part. Photonics, which moves data using light rather than electrical signals, matters for the data centres now mushrooming across the region. Power semiconductors, which convert and manage electricity, are essential to EVs, grids and energy storage — a natural fit for a country that is already Southeast Asia’s largest automotive production hub and whose mobility future is being rewired by electrification. Sensors build on existing strength in MEMS, the microscopic devices that detect motion, pressure and temperature inside phones, cars and medical equipment.

In short, Thailand is picking fights it might win, rather than chasing leading-edge logic chips, where TSMC, Samsung and Intel deploy capital on a scale no ASEAN budget can match.

The talent number got more realistic

This is where the story gets more interesting. Alongside the strategy, the government approved a workforce programme targeting 86,600 people by 2030: 84,900 highly skilled workers and about 1,700 advanced researchers, trained through specialised curricula, industry placements and overseas stints.

Also Read: From assembly line to innovation engine: Can Philippines climb the chip value chain?

In January, the talent targets being floated ranged from 17,500 to more than 200,000 engineers by 2030. Landing well short of the upper end is arguably more honest. It is also a reminder that the binding constraint on this plan is not tax holidays but people.

Chipmakers do not choose locations on incentives alone. They need process engineers, maintenance technicians, materials specialists and suppliers who understand what a speck of dust can do to a production line. Neighbours have learned this the hard way. Vietnam has pulled in Intel, Samsung and Amkor, yet is now wrestling with how to keep the engineers it trains. Malaysia, whose Penang cluster is a global force in assembly and testing, is discovering that AI demand does not lift every player.

The pipeline is real, but read the fine print

The BOI says it received investment-promotion applications for 879 semiconductor and advanced electronics projects worth about 909 billion baht (US$27.2 billion) between 2023 and the first half of 2026. Across the wider electronics sector — printed circuit boards, components and chip-related products — Thailand has attracted more than US$30.5 billion since 2023.

Applications, however, are not capital spent. Promotion requests are cheap to file and easy to shelve when demand turns, and the chip cycle has turned sharply more than once in the past five years.

A more concrete test arrives this week. Infineon Technologies is scheduled to open its first Thai factory, in Samut Prakan, on October 1. The plant will produce and package advanced power modules for EVs, energy storage and clean energy, and the German chipmaker plans an R&D centre and joint curricula with Thai institutions. An anchor investor squarely in one of the three priority segments is exactly what the strategy needs. Whether it seeds a cluster or remains a single impressive building will depend on how many local suppliers grow up around it.

A crowded neighbourhood

Thailand is not making this bet in a vacuum. Singapore, the region’s most mature chip hub, packaged its ambitions under a new national identity, SG Semiconductor, only on Friday. Malaysia is trying to pivot from assembly to indigenous design. Indonesia is courting Nvidia and AWS.

There is an awkward shadow, too. Thailand was among the jurisdictions named in a US case alleging that roughly US$2.5 billion worth of AI servers were routed through Southeast Asian intermediaries to China — part of the region’s wider chip-smuggling problem. Any country pitching itself as a trusted node in Western supply chains will need its customs and export enforcement to be as ambitious as its investment targets.

Also Read: Chips, corruption, and credibility: Malaysia’s semiconductor gamble faces a trust test

Thailand’s advantage is that it is not starting from zero. Decades of building hard drives, cars and electronics give it a foundation few emerging markets can match. Its risk is the familiar one of long-range industrial policy: the announcements keep getting bigger while the hard work — training engineers, building reliable utilities, nurturing local suppliers and keeping investors committed through downturns — moves at its own, much slower pace.

US$80 billion is a statement of intent. The first real scorecard lands in 2030, and 86,600 trained people is the number worth watching.

The post Thailand targets US$80B semiconductor push as it moves beyond assembly appeared first on e27.

Boost has secured a US$20 million equity investment from the International Finance Corporation (IFC), bringing the World Bank Group’s private-sector investment arm onto its cap table as the Malaysian fintech looks to deepen its digital lending and financial services business.

The deal gives Boost a development finance institution as a strategic shareholder at a time when Southeast Asian fintechs are under pressure to prove they can grow lending responsibly, serve underserved customers and build sustainable economics beyond payments.

For IFC, the investment is part of a broader push to support private-sector financial inclusion in emerging markets by backing digital players that can reach small businesses and consumers outside traditional banking channels.

Also Read: Bridging the financial gap: How digital lending is powering financial inclusion in Southeast Asia

Boost, part of the Axiata ecosystem, operates across Malaysia and Indonesia and has built its business around digital financial services for consumers and merchants. Its offering now spans fintech services and digital banking, including Boost Bank, a joint venture between Axiata and RHB in Malaysia.

The company said the IFC investment will support the development and scaling of digital financial products, including financing solutions for SMEs and consumers.

The cheque is modest by late-stage fintech standards, but its strategic value may matter more than its size. IFC brings not only capital but also experience investing in financial institutions and fintech businesses across emerging markets. According to the announcement, IFC has made more than 80 fintech investments globally.

Why IFC’s entry matters

Digital lending remains one of the most important, and most difficult, areas of fintech in Southeast Asia. The region has millions of micro, small and medium enterprises (MSMEs) that are too small, too informal or too thin-file for banks to serve efficiently. Many lack collateral, audited financial statements or long credit histories. Consumers face similar barriers when they work in informal jobs, have irregular incomes or are new to formal finance.

Fintech lenders try to close this gap by using alternative data and digital distribution. Instead of relying only on traditional credit files, they may assess transaction behaviour, merchant sales patterns, repayment histories, wallet usage or other signals to underwrite loans. Done well, this can widen access to credit. Done badly, it can push vulnerable borrowers into unaffordable debt.

That balance is likely one reason IFC’s participation matters. Development finance institutions typically place heavier emphasis on governance, risk management, consumer protection and impact measurement than purely financial investors do. In digital lending, those disciplines are not optional. They are central to whether financial inclusion becomes a durable business or another cycle of easy credit followed by defaults.

Farid Fezoua, Director of Equity, Funds, and Venture Capital at IFC, said innovative financial instruments are “essential to expanding access to finance at scale”, adding that the investment would support financing opportunities for underserved MSMEs.

The comment points to the heart of the opportunity: SMEs need working capital to buy stock, pay suppliers, hire workers and survive cash-flow gaps. In markets such as Malaysia and Indonesia, where merchants increasingly use digital payments and online tools, fintech platforms may have better real-time visibility into business activity than banks relying on static documents.

Boost’s regional play

Boost was launched in 2017 and has since served users and merchants in Malaysia and Indonesia. In Malaysia, its profile has grown through Boost Bank, the digital bank formed by Axiata and RHB. The bank is part of Malaysia’s broader digital banking wave, which regulators hope will extend formal financial services to underserved individuals and smaller businesses.

Malaysia has taken a relatively measured approach to digital banks compared with some other Asian markets. Bank Negara Malaysia awarded five digital bank licences in 2022, to consortiums led by Grab and Singtel, YTL and Sea, AEON, KAF Investment Bank, and Boost and RHB. The framework gives new entrants a chance to build alternative models, but it also places them under regulatory expectations around capital, risk and consumer protection.

Also Read: GXS Bank acquires Validus Capital to accelerate SME financing solutions

It is in this context that IFC’s investment lands. Boost is not just competing to acquire app users. It needs to show that its data, distribution and banking partnerships can translate into responsible lending at scale.

Sheyantha Abeykoon, Group CEO of Boost, said IFC’s backing brings “not only capital, but deep expertise in financial services and emerging markets”. He added that the partnership could help the company develop digital financial solutions that address the barriers faced by consumers and businesses.

Nik Rizal Kamil, Group CEO and Managing Director of Axiata Group, framed the investment as part of Axiata’s portfolio strategy, saying IFC’s entry reinforced confidence in Boost’s business model and governance.

The competitive field

Boost operates in a crowded Southeast Asian fintech market where payments, lending and digital banking increasingly overlap. In Malaysia, it faces competition from Touch ‘n Go eWallet, Grab, BigPay, Sea’s fintech arm and other digital finance platforms. In digital banking, Boost Bank competes with GXBank, backed by Grab and Singtel, and other licensed players as they roll out services.

Across the region, the challenge is even broader. Grab Financial, Monee (formerly SeaMoney), GoTo Financial and Kredivo have all used large consumer or merchant ecosystems to push into lending and other financial products. Traditional banks are also digitising quickly, often with stronger balance sheets and lower funding costs.

Boost’s edge will depend on how effectively it can use its merchant relationships, data and partnerships with Axiata and RHB to underwrite customers that others cannot serve profitably.

Financial inclusion, but with harder questions

The announcement uses the language of inclusion, but the next phase will be measured in execution. Southeast Asia has seen a wave of fintech enthusiasm over the past decade, followed by a more sober funding environment. Investors now want clearer paths to profitability, stronger credit controls and evidence that lending books can withstand economic stress.

That matters because digital credit can scale faster than traditional lending. A well-designed product can help a small merchant access capital within hours. A poorly designed one can create repayment pressure just as quickly. Regulators across the region are paying closer attention to digital lenders, especially around transparency, debt collection and customer affordability.

Boost and IFC say their collaboration will support alternative credit assessment and scalable digital lending. The practical test will be whether these tools can reduce exclusion without weakening underwriting standards.

IFC’s own mandate gives the deal a development angle. In fiscal year 2025, the institution committed US$71.7 billion to private companies and financial institutions in developing countries. Its investment in Boost fits that broader model: using private capital and expertise to expand access to finance in markets where conventional banking does not reach everyone.

Also Read: Digital banks win transactions, not loyalty: A missed opportunity in Indonesia

For Malaysia, the deal adds another marker to the country’s digital finance landscape. The market is not as large as Indonesia, nor as regionally central as Singapore, but it combines banked consumers, underserved SMEs, strong regulators and telecom-linked fintech players. That makes it a useful testing ground for models that may later scale elsewhere in Southeast Asia.

The US$20 million investment will not transform Boost by itself. But it gives the company a shareholder whose priorities go beyond rapid user growth. If Boost can combine IFC’s development finance discipline with its own digital reach, the more important outcome may be a lending model that expands access without repeating the mistakes of loose credit cycles.

The post IFC joins Boost’s cap table with US$20M bet on digital lending appeared first on e27.

For much of the past three years, the politics of artificial intelligence (AI) has revolved around relatively familiar questions. Will AI take jobs? Who owns the copyright to the material on which models are trained? Can companies protect personal data? And who should be responsible when an algorithm causes harm?

Britain’s latest AI debate suggests we may be entering a very different phase. More than 70 MPs and peers have urged Prime Minister Andy Burnham to support legislation prohibiting the development of artificial superintelligence (ASI) and pursue an international agreement preventing its creation.

The proposal is unlikely to become government policy immediately. ASI remains hypothetical, its definition contested, and Britain continues to see advanced AI as an important source of economic growth and strategic advantage.

But focusing on whether Westminster actually bans superintelligence misses the more important development.

The politics of AI safety is moving from technology policy into national security. Once that happens, the threshold for government intervention changes. Policymakers become more willing to impose restrictions despite economic costs.

Voluntary commitments become less persuasive. Companies accustomed to dealing with technology ministries and regulators suddenly encounter defence establishments, security agencies and heads of government.

Telecommunications infrastructure and semiconductor supply chains have already undergone versions of this transition. AI may be travelling along the same path.

The argument is changing inside the industry too

What makes the latest debate particularly significant is that calls for restraint can no longer easily be dismissed as coming from people outside the technology industry.

Anthropic chief executive Dario Amodei has argued that AI companies should deliberately pace the rate at which capabilities advance, giving safety research and safeguards time to catch up.

His position is not simply to stop AI development. Amodei continues to argue that AI could deliver enormous benefits. His concern is that capabilities may now be advancing faster than our ability to understand and control them, particularly as AI itself becomes increasingly useful in developing subsequent generations of AI. But something more striking has now happened.

OpenAI chief executive Sam Altman has publicly backed Amodei’s argument that the industry needs to “pace the frontier”, saying it has been a primary subject of discussion inside OpenAI. He has also committed OpenAI to Amodei’s proposal to give independent evaluators employee-like access to assess safety practices.

Elon Musk, whose xAI competes directly with both companies, offered an even more succinct endorsement: “Dario is right.”

The significance lies less in the individual statements than in who is making them.

These are fierce commercial competitors with very different views about AI and its governance. Yet leaders of three major frontier AI companies are now publicly acknowledging that there may be circumstances in which capability development should slow.

OpenAI has gone further. It is advocating mandatory, capability-based national AI safety regulation, independent safety assessments and international standards for determining when development should slow or stop.

That represents an important change in the regulatory debate. The question may no longer be simply whether governments should accelerate or constrain AI. It could become how fast the frontier should move, and what safeguards must accompany each increase in capability.

Also Read: Why Southeast Asian enterprises need AI governance before scaling generative AI

From self-regulation to supervision

This could have profound consequences for business. Amodei has proposed giving independent external evaluators ongoing access to parts of Anthropic’s operations, comparing the concept with regulatory supervisors embedded within financial institutions. OpenAI has now said it will do the same.

The analogy should attract policymakers’ attention. Financial regulation did not develop on the assumption that banks could simply declare themselves safe. Independent supervision, stress testing, capital requirements and disclosure became embedded in the system.

Something similar could eventually emerge around frontier AI. Saying that a company takes safety seriously may no longer be sufficient. Governments may demand that companies demonstrate it through independent testing, incident reporting and measurable thresholds beyond which additional safeguards become mandatory.

AI governance could therefore be moving from promises towards verification.

Britain faces its own contradiction

The Burnham government consequently faces a difficult balancing act. It wants Britain to be a serious AI power, requiring investment, infrastructure, talent and companies willing to develop increasingly capable models.

Yet it must simultaneously convince voters that those technologies will not create unacceptable risks.

There is no simple national solution because AI development is inseparable from geopolitical competition. If democratic countries slow their programmes while competitors do not, restraint could create a national-security vulnerability.

That makes international coordination increasingly important.

Britain will host the G20 in 2027. Rather than an improbable global prohibition on superintelligence, governments may find more practical territory for cooperation: prohibiting narrowly defined dangerous applications, developing common testing standards for cyber and biological risks, or establishing internationally recognised capability thresholds.

Also Read: AI governance in banking operations and decisioning

Once policymakers start discussing AI in the conceptual language of arms control rather than digital regulation, politics have fundamentally changed.

For Singapore, this matters.

Singapore has deliberately pursued a pragmatic model of AI governance: encourage adoption and innovation while developing frameworks for testing, accountability and risk management.

Its experience with AI assurance, combined with its position as a trusted and technologically sophisticated economy, could give it a useful role in shaping the standards and verification mechanisms that a more internationally coordinated system would require.

Political risk becomes operating risk

For business, this is ultimately the lesson. AI companies can no longer treat regulation as a compliance exercise conducted after technology strategy has been decided. Boards need to understand how political perceptions of their technology are changing.

Government affairs teams need relationships extending beyond technology ministries. Frontier developers may increasingly need to accommodate independent scrutiny that once would have seemed commercially intrusive.

Companies using advanced AI should pay attention too. If regulation becomes capability-based, obligations may increasingly follow what an AI system can do rather than the industry in which it is deployed.

The immediate proposal to ban artificial superintelligence may succeed or disappear. But something more consequential has already happened.

Politicians are contemplating prohibiting the most advanced forms of AI. Leaders of competing frontier laboratories are openly discussing whether development sometimes needs to slow. OpenAI is advocating mandatory capability-based safety regulations. And proposals involving independent monitors, regulatory thresholds and coordinated restraints on development are moving towards the mainstream.

The important question therefore no longer be whether greater regulation is coming.

It is whether safety can keep pace with capability – or whether governments and the companies building the technology will eventually decide that capability itself must slow down.

—

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 AI governance is moving from promises to proof appeared first on e27.