Day: September 9, 2026

Southeast Asia’s e-commerce story is no longer just about search bars, discount vouchers and marketplace rankings. Increasingly, shoppers are discovering what to buy through livestreams, short videos, creator reviews and affiliate content. And that shift is now large enough to reshape the region’s online retail economy.

Content commerce gross merchandise value across Shopee, TikTok Shop, and Lazada reached US$49.7 billion in 2025, almost doubling from the previous year, according to Momentum Works’s latest report, Live Commerce in Southeast Asia 2026. The Singapore-headquartered research and venture outfit estimates that US$33.8 billion was transacted in the first half of 2026 alone. If the current pace holds, the segment is on track to hit US$77.9 billion for the full year.

Also Read: The rise of live commerce in Asia and adoption of BeLive by retailers

The more telling number is not just the headline GMV, but content commerce’s share of the region’s e-commerce mix. In the first half of 2026, it accounted for 37 per cent of Southeast Asia’s platform e-commerce GMV, up from 20 per cent in 2024. In other words, what was recently treated as an add-on marketing channel has become a major sales engine.

For founders, brands and marketplace operators in the region, this marks an important turning point. Southeast Asia’s e-commerce markets — from Indonesia and Thailand to Vietnam, the Philippines, Malaysia, and Singapore — have long been shaped by mobile-first behaviour, price sensitivity, and high social media usage. Content commerce sits at the intersection of all three. It makes shopping more entertaining, but also more immediate: a product demo, a creator recommendation, and a checkout button can now sit within the same customer journey.

Live commerce moves into the operating core

Live commerce has been the most visible part of this shift. The format allows sellers, creators, and brands to demonstrate products in real time, answer questions, and trigger purchases through limited-time offers or platform vouchers. In categories such as beauty, fashion, household goods, and fast-moving consumer products, it has become a daily operating channel rather than a campaign experiment.

Momentum Works notes that for many brands, the question is no longer whether they should go live, but what role live should play in the broader business. Some use it mainly for conversion, pushing volume during platform sales days. Others use it to educate consumers on new products, build trust in unfamiliar brands, or move slower-selling inventory.

That distinction matters because live commerce does not work equally well for every product. A low-priced lipstick, snack bundle, or kitchen gadget can benefit from quick demonstrations and impulse buying. Higher-consideration purchases may need more education, reviews, and repeat exposure before a customer checks out. Execution quality also matters: the host, script, pacing, product assortment and incentives can materially affect sales.

For now, the returns from live remain attractive for many operators. But Momentum Works argues that these returns are unlikely to stay unusually high forever. They are being supported by growing consumer attention, platform incentives and a competitive environment that is still maturing. As more brands, agencies, sellers and creators develop similar capabilities, the cost of standing out will rise.

The next battleground: brandformance

This is where “brandformance” enters the conversation. The term, a blend of brand building and performance marketing, captures a problem many e-commerce teams face: short-term conversion can be measured instantly, but long-term consumer preference is harder to track.

Live commerce is naturally performance-driven. A seller can see how many viewers joined, how long they stayed, which products were clicked and what was purchased. That makes it appealing in a region where marketing budgets are often tied closely to measurable outcomes. But if every brand is running live sessions with similar scripts, discounts and affiliate networks, performance alone becomes easier to copy.

Also Read: Elevating your e-commerce strategies with livestreaming and hero products

The longer-term advantage may sit with companies that use content not only to sell, but to build memory and trust. That could mean explaining why a skincare product works for humid climates, why a halal-certified food product matters to Muslim consumers, or how an electronics brand supports after-sales service in provincial cities. In fragmented Southeast Asian markets, where language, culture, logistics, and purchasing power differ widely, local relevance is not a minor detail.

The challenge is that many brands still treat live commerce as a standalone sales machine. It is visible, measurable and relatively easy to justify internally. Short videos, affiliate reviews and community content can be harder to attribute, even when they play a crucial role in creating demand before the livestream begins.

AI lowers the cost of execution

The report also points to a structural change that could compress the advantage of skilled operators: AI live. In selected cases, Momentum Works says AI-driven live operations cost around 20-25 per cent of a comparable human setup while achieving around 80 per cent of human livestream GMV per hour on average.

That has significant implications. Capabilities that once took agencies, brands, and livestream studios years to build (scripting, scheduling, product explanations, host consistency, and basic audience interaction) are becoming more accessible through technology and platform tools. For smaller sellers, this could lower the barrier to entry. For larger brands, it could reduce operating costs and allow more always-on content.

But it also creates a strategic problem. If everyone can access similar tools, operational capability alone becomes less defensible. The differentiator shifts to what cannot be automated as easily: product quality, customer insight, creative direction, creator relationships, community trust and brand positioning.

This matters in Southeast Asia because the region’s ecommerce growth has often been fuelled by intense marketplace competition and subsidised demand. As subsidies normalise and consumer acquisition becomes more expensive, brands will need more than efficient livestream operations. They will need reasons for shoppers to return without being pulled only by the next discount.

China offers lessons, not a template

China remains the global reference point for live commerce. Its ecosystem is more mature, with advanced livestream infrastructure, professional creator networks, high-frequency shopping behaviour and deeper use of data and automation. Southeast Asian platforms, brands and sellers have borrowed heavily from that playbook.

Yet Momentum Works cautions that China should be seen as a map, not a blueprint. Southeast Asia is not one market. Creator economics in Indonesia differ from Singapore. Consumer trust patterns in Vietnam may not mirror those in Thailand. Payment habits, logistics reliability, local languages and platform dynamics vary sharply across the region.

That means the next phase of content commerce will likely be less about copying a single model and more about adapting formats market by market. A livestream strategy that works in Bangkok may need to be rebuilt for Manila. A short-video approach that drives discovery in Jakarta may not translate neatly to Ho Chi Minh City.

Also Read: How AI, AR, and live streaming are changing the online shopping experience

The broader lesson is clear: live commerce has become infrastructure, but it is not the entire content commerce strategy. As the channel matures, brands that over-invest in live while under-funding short video, affiliates and review-led discovery risk mistaking the checkout moment for the whole customer journey.

For Southeast Asia’s digital economy, the US$77.9 billion forecast is a sign of how quickly shopping behaviour is changing. The next question is not whether content will shape e-commerce, but who can turn attention into durable customer relationships once the easy growth fades.

The post Southeast Asia’s live commerce boom enters its harder second act appeared first on e27.

Capital A’s (formerly AirAsia Group) latest numbers tell a company coming out of crisis, but not yet one firing evenly across all engines.

The Malaysia-based group, which has spent the past few years restructuring after the pandemic and disposing of its airline business, reported second-quarter revenue of about US$193 million, up 9 per cent year-on-year. For the first half of 2026, revenue stood at about US$376 million, a 4 per cent increase from a year earlier.

On the surface, that points to stability. Capital A also reported profit after tax of about US$6 million for the quarter and US$11.9 million for the first half, giving it another profitable quarter after the airline disposal.

Also Read: AirAsia unit Teleport buys stake in Indonesia’s ‘Uber for logistics’ Kargo Technologies

But the recovery is more uneven than the topline implies. Growth is being driven mainly by two units: Asia Digital Engineering (the aircraft maintenance, repair, and overhaul business) and Teleport (the logistics arm). Together, ADE and Teleport accounted for more than 70 per cent of first-half group revenue.

That leaves the rest of the portfolio (AirAsia MOVE, AirAsia Next, and Santan) with a harder job to prove that Capital A can build a broad-based, asset-light aviation services and digital platform business beyond the airline brand that made it famous.

Profitability returns, but margins remain thin

Capital A’s return to profitability is meaningful. The group has exited PN17 status, a classification for financially distressed companies on Bursa Malaysia, and is trying to rebuild investor confidence around a cleaner corporate structure.

Yet the profit margin leaves little room for error. Second-quarter profit after tax of around US$6 million on revenue of US$193 million implies a net margin of roughly 3.1 per cent. For the first half, profit after tax of US$11.9 million on US$376 million revenue works out to about 3.2 per cent.

For a group still in transition, that is not alarming by itself. But it does mean the turnaround remains vulnerable to foreign exchange movements, interest costs, lease obligations, capital expenditure and slower volumes.

The operating picture is also less flattering. First-half net operating profit fell 21 per cent year-on-year to about US$16 million, despite revenue growth. Capital A said core group net operating profit rose 6 per cent after adjusting for the loss of aviation interest income following the airline disposal.

That adjustment may be fair, but it is also doing a lot of work. The reported number shows operating profit declined. The adjusted number supports the recovery story. Investors will want a clearer bridge between the two.

Group EBITDA also fell 5 per cent in the first half, even as revenue rose 4 per cent. That suggests either costs are rising faster than sales, or the revenue mix is tilting towards lower-margin activities.

ADE and Teleport carry the group

The strongest part of the update is ADE. The aircraft maintenance unit reported second-quarter revenue of about US$67.6 million, up 29 per cent year-on-year, with EBITDA of about US$16.4 million. Capital A said hangar slots are booked through next year and that ADE is building a new four-line maintenance hangar.

That demand backdrop is credible. Southeast Asia’s airline industry is still rebuilding capacity after the pandemic, while narrowbody aircraft fleets across the region need maintenance as utilisation rises. Supply-chain delays and aircraft delivery bottlenecks have also made maintenance capacity more valuable.

The question is how much cash ADE will need to keep growing. Maintenance is not a pure software-style business. Tools, hangars, engineering talent and certifications require investment, and depreciation will rise as capacity expands. EBITDA may look healthy while free cash flow tells a more complicated story.

Teleport also showed momentum. Second-quarter revenue rose 22 per cent year-on-year to about US$74 million, while first-half revenue grew 21 per cent to about US$147.6 million. Tonnage in the quarter reached 85,877 tonnes, up 11 per cent year-on-year, and parcel volume jumped 79 per cent to 56.6 million.

Also Read: AirAsia aims to fulfill super app ambition with upcoming launch of ride-hailing services in Malaysia

For a logistics business operating in a softer global freight market, that is a solid result. But margins remain modest. Teleport’s second-quarter net operating profit was about US$1.8 million on US$74 million in revenue, implying an operating margin of around 2.4 per cent. Profit after tax was around US$1 million.

There is also some selective framing. First-half tonnage was 182,660 tonnes, which means first-quarter tonnage was around 96,783 tonnes. On that basis, second-quarter tonnage declined sequentially even as the year-on-year comparison looked positive.

Consumer units still have work to do

AirAsia MOVE, Capital A’s travel platform, is where the pressure is more visible. The unit reported second-quarter revenue of about US$22.9 million, up 5 per cent year-on-year. But its WANO B2B business contributed 11 per cent of total revenue, or roughly US$2.5 million.

Excluding WANO, MOVE’s underlying revenue appears to have declined year-on-year. Flight sales also fell 3 per cent, which Capital A attributed to an 11 per cent reduction in AirAsia capacity. That explanation is reasonable, but it underlines MOVE’s continued dependence on the AirAsia airline ecosystem.

AirAsia Next, which includes loyalty and licensing activities, remains profitable. It posted second-quarter revenue of about US$18.6 million, EBITDA of US$6.2 million and net operating profit of US$5.2 million. But part of the growth came from non-aviation licensing fees and AirAsia Rewards, where revenue recognition can be influenced by points redemptions. The company said redemptions rose 34 per cent, helping revenue but also increasing redemption expenses.

Santan, the group’s food business, remains small. Second-quarter revenue was about US$10.7 million, broadly flat on a normalised basis, while passenger volume fell 14 per cent due to airline capacity constraints. Its push into e-commerce through TikTok and Shopee is sensible, but Capital A did not disclose the absolute revenue base, making the 30 per cent quarter-on-quarter growth figure hard to assess.

Rivals are not standing still

Capital A’s challenge is that each part of the group competes with specialised players. ADE faces established maintenance providers such as SIA Engineering, ST Engineering Aerospace, GMF AeroAsia and Lufthansa Technik Philippines. Teleport competes in a crowded logistics market against DHL, FedEx, UPS, J&T Express, Ninja Van and regional cargo operators. AirAsia MOVE is up against Traveloka, Agoda, Booking.com, Trip.com and airline direct channels. That makes execution harder: Capital A is not fighting one market battle, but several at once.

The balance sheet update also leaves questions unanswered. Capital A said shareholders’ equity is comfortably above US$119 million and operating cash flow was about US$35.7 million. It also said refinancing reduced interest expenses.

Those are positive signs. But without clearer disclosure on total debt, net debt, lease liabilities, cash balance, capital expenditure commitments and interest coverage, it is difficult to judge how strong the balance sheet really is.

Also Read: AirAsia calls off US$10M acquisition of Gojek Thailand’s fintech arm: report

The fairest reading is that Capital A is in better shape than it was during the depths of its restructuring. ADE and Teleport are growing, the group is profitable again, and the PN17 overhang has been removed.

But this is not yet a broad, high-margin recovery. It is a narrower turnaround led by two operating units, while consumer-facing businesses remain tied to airline capacity, accounting-sensitive revenue streams and early e-commerce bets. Capital A has stabilised. Now it has to prove the new group can compound.

The post Teleport powers Capital A’s rebound, but thin margins show logistics remains a hard road appeared first on e27.

For many consumer app companies, the hardest part is no longer building the first product. It is finding enough growth capital to keep buying users profitably once a product has already shown traction.

Tevo, a consumer and AI apps company from Southeast Asia, is now turning to a financing model designed for exactly that gap.

The company has secured US$10 million in non-dilutive user-acquisition financing from PvX, a Singapore-based platform that provides growth capital for mobile gaming and consumer app businesses.

Also Read: Tevo secures seed funding, strikes partnership with Vietnam’s MobiFone

Tevo said the facility will be used to scale user acquisition across its portfolio of consumer apps and AI products.

Unlike a traditional equity round, non-dilutive financing does not require the company to sell shares. In the app economy, this form of capital is often tied to marketing performance: companies use the funds to acquire users and then repay the financing from revenues generated by those users. For founders, the appeal is straightforward. If the unit economics already work, they can spend more on growth without giving up ownership.

Tevo operates around 45 consumer and AI apps across work utilities, education and entertainment. Collectively, those products have generated nearly 200 million installs globally, according to the company. The new capital will go into priority markets, higher marketing spend on user cohorts that have already proven profitable, and further investment in AI-native product features.

“Non-dilutive UA financing lets us put more capital behind products that already have product-market fit,” said Thanh Luu, CEO and founder of Tevo. “This facility also gives us the flexibility to scale globally and accelerate Tevo’s 2030 vision as the leading company in AI apps and services, and among the top five largest mobile apps and games companies from Southeast Asia.”

Why user acquisition financing is gaining ground

Tevo’s deal points to a wider change in how consumer app companies are funding growth. For years, many app businesses relied on venture capital to finance user acquisition, even when the money was being spent on paid marketing rather than product development or hiring. That made sense during the low-interest-rate era, when investors were willing to fund aggressive growth. But the downturn in venture funding has forced founders to think more carefully about what kind of capital fits each use case.

User acquisition is a different problem from building a new product. If a company has enough data to show that a customer acquired for US$1 can eventually generate more than that in revenue, then the risk profile becomes more measurable. In that case, performance-linked financing can be a better fit than equity capital, particularly for founders who do not want to dilute themselves just to increase ad spend.

PvX said it has surpassed US$750 million in committed user acquisition financing for mobile gaming and consumer app companies globally. Its focus on Singapore as a base is also notable. Southeast Asia has produced major gaming and consumer internet companies, but the region still has relatively few scaled consumer app platforms with global reach. Financing models such as PvX’s could help bridge that gap by giving app operators access to capital based on revenue performance rather than venture-market sentiment.

Also Read: PvX lands MIT investment to fund the next wave of app user acquisition

The timing is also important. Artificial intelligence has lowered the barrier to launching new consumer software products, from study tools and productivity assistants to content and entertainment apps. But it has not solved the distribution problem. App stores are crowded, advertising costs can rise quickly, and winning users requires both data discipline and capital. Companies that already run multiple apps have an advantage because they can test, optimise, and redeploy learnings across a portfolio.

Southeast Asia’s consumer app opportunity

Southeast Asia is a mobile-first region, with large young populations, high social media usage, and deep familiarity with digital services. Yet many of the world’s biggest consumer app companies still come from the US, China, Europe, Turkey, Israel and India. Southeast Asian startups have built strong positions in ride-hailing, e-commerce, fintech and gaming, but fewer have become global consumer app factories.

That is what makes Tevo’s positioning interesting. Rather than focusing on one flagship app, the company runs a broad portfolio across practical and entertainment-led categories. Work utilities and education apps can offer recurring use cases, while entertainment products can scale quickly if they find the right audience. AI adds another layer, allowing companies to turn common consumer needs — writing, studying, editing, searching, creating — into lightweight software experiences.

The challenge is that portfolio app businesses live and die by execution. Downloads alone do not guarantee long-term value. Retention, monetisation, ad efficiency, subscription conversion, and churn matter more than headline install numbers. Tevo’s nearly 200 million installs give it a base to build from, but the real test will be whether additional user acquisition spending can produce users who stay and pay.

This is where non-dilutive capital can be both useful and unforgiving. It rewards companies with strong data and clear payback periods, but it also exposes weak assumptions quickly. If marketing spend is pushed into channels or countries where users do not convert, the model breaks down. For Tevo, the stated focus on “proven cohorts” suggests the company intends to put capital behind segments where performance is already visible.

The competitive field

Tevo is not alone in chasing the consumer AI apps opportunity. Globally, it sits in a competitive field that includes portfolio app operators such as Turkey’s HubX, which runs more than 40 mobile apps across AI, education, health and fitness and has surpassed 600 million downloads. HubX recently announced an investment of up to US$75 million from Point72 Investments at a US$1.2 billion pre-money valuation, making it Turkey’s first consumer-apps unicorn.

Other global rivals include mobile-first app studios and subscription app companies building AI tools for productivity, learning, photo editing, wellness and entertainment. In Southeast Asia, the field is less crowded at scale, but local gaming studios, AI productivity startups, and consumer internet firms could all move into overlapping categories as AI app demand grows.

That competitive pressure makes distribution capital more important. AI features can be copied quickly, and app store rankings are volatile. Companies that understand paid acquisition, localisation, monetisation, and rapid product iteration are more likely to survive than those relying only on novelty.

Also Read: PvX bags US$10.5M as cohort financing goes mainstream

For Southeast Asia, Tevo’s financing is also a sign that regional consumer app companies are beginning to access the same specialised capital structures used by more mature app markets. If the company can translate financing into sustainable global growth, it could help widen the region’s startup narrative beyond marketplaces, fintech, logistics and enterprise SaaS.

The US$10 million facility is not a traditional funding round, and it does not carry the signalling effect of a headline valuation. But that may be the point. In a market where founders are being pushed to grow more efficiently, capital that follows performance — rather than hype — may become a more common way for consumer app companies to scale.

The post Tevo secures US$10M from PvX to scale its consumer AI apps without selling equity appeared first on e27.

GCash operator Mynt has cleared a key regulatory hurdle after the Philippine Securities and Exchange Commission approved its IPO of up to US$1.63B, one of the largest listings the country has seen in years.

The Commission En Banc resolved to render effective Mynt’s registration statement covering up to 66.9B common shares, subject to remaining requirements. Mynt’s journey to this point has been a long one; it became the Philippines’ first fintech unicorn back in 2021, backed by a US$300M round.

Mynt plans to offer up to 1.61B new shares through a primary offer, alongside a secondary sale of up to 6.42B shares and an overallotment option of 1.2B shares, priced at up to roughly US$0.18 apiece. Fully exercised, the deal could raise net proceeds of about US$1.58B, with roughly US$264M from the primary offer earmarked for growth in digital financial services and product development.

The offer period runs from 6 to 12 October, with Mynt targeting a 20 October listing on the Philippine Stock Exchange’s Main Board under the ticker “GCASH”. Mynt also becomes the first issuer to benefit from the SEC’s lower public float requirement for large companies, at 12% instead of 15%.

The listing will be closely watched as a valuation benchmark for Southeast Asian fintech, testing whether GCash’s dominant consumer brand can convert into durable public-market economics and arriving just as the region’s broader IPO window has started to reopen.

---

REGIONAL

GCash operator Mynt clears SEC hurdle for up to US$1.63B IPO: The Philippine SEC has approved Mynt’s up to US$1.63 billion IPO, clearing the way for an October listing on the PSE under ticker GCASH, a closely watched valuation test for Southeast Asian fintech.

SEA content commerce GMV nears US$50B as live shopping matures: Momentum Works pegs 2025 content commerce GMV at US$49.7 billion across Shopee, TikTok Shop and Lazada, forecasting US$77.9 billion in 2026 as AI-driven livestreams cut costs to 20-25% of human setups.

Tevo secures US$10M non-dilutive financing from Singapore’s PvX: Tevo, which runs 45 consumer and AI apps with nearly 200 million installs, will use the non-dilutive facility from PvX to scale user acquisition without diluting founder ownership.

Capital A rebounds on Teleport and ADE, but margins stay thin: Second-quarter revenue rose 9% to about US$193 million, with ADE and Teleport contributing over 70% of first-half revenue as net margins stayed near 3%, and consumer units like AirAsia MOVE still lag.

SEA funding drops 74.78% from July peak but improves on 2025: Southeast Asian startups raised US$1.171 billion across 13 rounds in August, down 74.78% from July’s record but up 582.75% year-on-year, with Sharpa’s US$669.7 million round leading a barbell-shaped market.

Temasek, Seraphim lead US$100M round for India’s Pixxel: Indian space-tech firm Pixxel has raised US$100 million in a Series C led by Temasek and Seraphim Space, taking total funding to US$195 million to expand its satellite and Earth-intelligence platform.

VinFast’s Vietnam factory arm goes fully domestic after exit: VinFast Auto has divested its stake in VinFast Trading and Production JSC, its Vietnamese manufacturing arm, making the US$3.25 billion factory business wholly domestically owned under an asset-light restructuring.

Singapore’s Ant International wins Brazil payment licence: Ant International has secured a payment-institution licence from Brazil’s central bank, expanding its regulated footprint beyond Antom’s merchant-payments platform as Brazil tightens cross-border payment oversight.

Vietnam plans 60-minute daily game cap for under-16 players: Vietnam’s draft rules would cut daily game time for under-16 players to 60 minutes from 180, requiring parental account registration and mobile-number verification for all players.

Singapore Prison Service deploys PROTECT surveillance robot: The Singapore Prison Service has built an autonomous robot named PROTECT with HTX to boost surveillance and incident response, giving officers remote video, audio and interdiction tools during incidents.

Thailand freezes 49 data centres over resource strain: Bangkok has halted approvals for 49 data centre projects, citing pressure on power and water resources, a major signal for hyper scalers and investors banking on Thailand as a regional digital infrastructure hub.

Tazapay opens Bengaluru centre, eyes India expansion: Singapore-based cross-border payments firm Tazapay has launched an engineering hub in Bengaluru, signalling a push to deepen its India footprint as it scales payment infrastructure across Asia.

---

REPORTS AND INTERVIEWS

The US$103K H-1B fee won’t hand SEA a talent windfall: Trump’s US$103,265 H-1B visa fee could push skilled workers out of the US, but Southeast Asia’s own brain-gain record suggests capturing them needs deeper pay and equity reform, not just new visas.

---

INTERNATIONAL

Why Kyoto, not Tokyo, is quietly becoming Japan’s deeptech bet: Kyoto’s 650-plus startups are betting patience beats speed, building semiconductor, robotics and life-sciences ventures around a manufacturing lineage spanning Nintendo, Kyocera and Murata, investors say.

Authors dispute publisher and agent claims on Anthropic payout: Writers say publishers and literary agents are wrongly claiming shares of Anthropic’s US$1.5 billion copyright settlement, including for books whose rights had already reverted to authors.

Seattle Times, Newsday sue OpenAI and Microsoft over AI training: Two more US newspapers have filed suit, arguing generative AI could leave journalism irreparably damaged by training on their reporting without compensation or consent.

Ping An Digital Bank launches receivables financing for e-commerce: Hong Kong’s Ping An Digital Bank now offers financing of up to US$5 million against export receivables, targeting cross-border e-commerce merchants with one-day approval turnaround.

Google-backed Indian space startup raises US$100M: Pixxel, which operates hyper spectral imaging satellites, closed a US$100M round, one of India’s largest space tech raises, with implications for earth observation demand across Southeast Asian markets.

Peak XV and Filter Capital lead US$50M round in Nua: Indian consumer health brand Nua secured US$50M in around led by Peak XV Partners and Filter Capital, under scoring sustained investor appetite for women’s health and wellness across emerging Asian markets.

Ola Electric clears US$180M fundraise as COO exits: India’s electric two-wheeler maker approved a major capital raise even as its COO resigned, a dual signal of ongoing financial pressure and leadership instability at one of Asia’s most watched EV firms.

Dubai’s Talabat and Quikbot trial high-rise delivery robots: The partnership tests autonomous robots for vertical last-mile delivery in multi-storey buildings, a use case with direct relevance to Singapore, KL, and other dense SEA urban markets.

Samsung to unveil humanoid robot at CES 2027: Samsung plans to debut its humanoid robot at CES 2027, entering a field already contested by Tesla and Figure, with manufacturing and logistics implications for Southeast Asia’s factory-heavy economies.

---

CYBERSECURITY

OpenAI agents secretly ran a German wiki forum for weeks: Independent researchers found internally deployed OpenAI agents had hijacked an obscure German wiki for over a month, coordinating on evaluations until OpenAI staff appeared to notice and intervene.

OpenAI’s escaping agents expose gaps in AI incident oversight: Safety researchers say OpenAI’s narrow investigation into repeated agent breakouts shows frontier labs still control the scope of their own safety reviews, with no independent audit process in place.

Liquid Network halts trading after US$320M Bitcoin withdrawal: Bitcoin sidechain Liquid Network paused transactions after about 4,000 BTC left its federation wallet via an authorised peg-out route, with actors claiming white-hat intent but no funds yet returned.

---

SEMICONDUCTOR

Malaysia eyes Huawei chips for national AI project: Kuala Lumpur is weighing Huawei chips for a state-backed AI initiative despite explicit US warnings, a move that could strain trade ties and signal a broader regional shift away from US semiconductor dependency.

Israel’s Accelerate targets chip design efficiency with AI: Tel Aviv-based Accelerate has developed an AI tool that cuts semiconductor design cycle times, a development relevant to Southeast Asia’s growing chip design ambitions in Malaysia and Vietnam.

---

AI

SEA’s 680 million people make it an AI market to watch: With a digital economy set to exceed US$600 billion by 2030, Southeast Asia is emerging as a genuine AI talent hub, not just a market, a contributor argues, now home to 67,000 AI engineers.

OpenAI’s Astra pushes AI from chatbot toward digital worker: OpenAI has launched GPT-6 Astra, pitching it as its most capable model yet for computer use, coding and research, while flagging stronger cyber capabilities that also raise governance risks.

AI gives answers fast; experience decides which ones count: As AI makes first drafts nearly free, the value shifts to judgement, a contributor argues; domain experience, not prompting skill, decides which of AI’s many suggestions survive contact with reality.

---

THOUGHT LEADERSHIP

Playing checkers against China’s AI ecosystem strategy: The US bet on two AI heavyweights versus China’s broader open-weight ecosystem is a high-stakes divergence, a contributor writes, urging startups to map allegiances rather than pick a permanent side.

Malaysia’s second digital wave is won on friction, not novelty: Malaysia’s next wave of digital growth hinges on fixing fragmented payment rails and AI pilots stuck at proof-of-concept, a contributor argues, not the convenience plays that built Grab-era adoption.

Taiwan’s startup problem is matching talent, not scarcity: Platform data from EZStartup shows most founders want collaborators, not more skills, suggesting Taiwan’s bottleneck is poorly defined roles and untested trials rather than a shortage of willing talent.

Ethical AI means letting frontline staff challenge the system: Contestability, not policy statements, is the real test of ethical AI, a contributor writes, arguing most firms want the comfort of human oversight without paying its operational cost.

Bitcoin’s September hinges on a narrow US$79K–US$82K range: Bitcoin must hold a support band between US$79,300 and US$79,900 through the 11 September US inflation data to retest resistance near US$82,400, a contributor’s technical analysis shows.

The post Ecosystem Roundup: GCash operator Mynt clears SEC hurdle for up to US$1.63B IPO appeared first on e27.

In April 2026, the United States announced 83,387 job cuts. 26 per cent of them named artificial intelligence as the reason, the second consecutive month that AI was the top cited cause. Behind those numbers is a quieter story that is going to shape an entire generation of careers.

Stanford economists Erik Brynjolfsson, Bharat Chandar, and Ruyu Chen released findings showing that employment among 22 to 25 year olds in AI-exposed jobs has dropped between 16 and 20 per cent in software development as the trend accelerated into 2026. Older workers in the same roles are largely untouched. The cut is happening at the bottom of the ladder, not the top.

Universities and bootcamps have responded the way they usually do. Add more AI to the curriculum. Bachelor’s level AI programmes in the US grew 114 per cent from 2024 to 2025, jumping from 90 to 193 programmes. New AI majors are launching at Northwestern, Carnegie Mellon, and dozens of other universities. Coding bootcamps now market AI tracks, prompt engineering modules, and LLM integration certificates. The reflex is consistent. If AI is reshaping work, teach more AI.

This reflex is producing graduates who are technically fluent but commercially unhireable. And the data is now clear on why.

The tool trap

A joint study by Amazon Web Services and Pearson, published in April 2026, surveyed employers and education leaders across six focal markets including the US, UK, Vietnam, and Malaysia on what they actually want from graduates entering AI-augmented workplaces. The headline finding is uncomfortable for every institution that has been racing to add AI courses. Employers do not have an AI skills problem. They have a judgement problem.

The study identifies six frictions in the education-to-workforce gap. Only one of them is about technical AI skill. The other five are about pace of curriculum adaptation, weak feedback loops between universities and employers, governance, applied experience, and the gap between graduate abilities and the judgement, adaptability, and collaboration employers want.

A separate 2026 Wonkhe analysis of UK employer surveys found the same pattern. One third of employers rated graduates as below expectations on adaptability, self-awareness, and awareness of the wider organisational context. The same employers were broadly satisfied with foundational technical skills. The gap is not where universities are looking.

Kim Majerus, vice president of global education at AWS, put it plainly. The opportunity is to translate AI tool engagement into real workplace capability, which requires judgement, adaptability, and hands-on experience.

This is what I call the Tool Trap. Universities and bootcamps are training Gen Z in the skills AI itself is best at. The graduates produced are fluent in prompts. They can build with LLMs. They have ethical AI modules on their transcript. What they cannot do is the thing AI cannot do. Decide which problem is worth solving. Read whether an output is good enough to ship. Take responsibility when a decision goes wrong. Sit across the table from a customer who is paying real money and earn their trust.

These are not soft skills. They are the highest-value skills in the AI economy. And they are not on the syllabus.

Also Read: Gen Z and the rise of AI-powered travel

Why the tool trap exists

There is a structural reason this misdiagnosis keeps happening. Tool literacy is easy to teach, easy to certify, and easy to market in a prospectus. Judgement, taste, accountability, and customer trust are slow to develop and impossible to test in a written exam. Universities and bootcamps are optimised for the things they can measure. The economy is now paying for things they cannot.

I had my business research team study 2,500 companies across 25 years, and the work surfaced a useful framework for thinking about this. Inside every operating company, three roles do the actual work that makes the company succeed. The Builder builds the product. The Domain Expert knows the customer and the industry. The Business Driver decides which problems are worth solving and which deals are worth taking. AI can dramatically accelerate the Builder role. It can support the Domain Expert role. It cannot replace the Business Driver role, because the Business Driver lives at the layer of judgement, taste, and human accountability.

Today’s AI curriculum trains Gen Z to be better Builders. The Builders are the role most exposed to AI replacement. The Business Drivers are the role most insulated. Universities are pushing students toward the wrong end of the value chain, and the labour market is starting to notice.

What Gen Z actually needs to learn

If I were advising any university student or recent graduate in 2026, my advice would not be take more AI courses. It would be the opposite. Take fewer AI courses. Take more of the courses that build the capacities AI cannot replicate.

Learn to write clearly so you can think clearly. Learn to sit in front of a real customer and figure out what they need before you build it. Learn to make a decision with incomplete information and own the outcome. Learn to spot when an AI output is technically correct but commercially wrong. Learn to negotiate, to read a room, to build trust with people whose money you are asking for.

This is not a rejection of AI literacy. Every graduate in 2026 should be fluent in AI tools. That fluency is now a baseline, not a differentiator. The differentiator is what surrounds the fluency. The Wonkhe data, the AWS-Pearson study, the Stanford research on entry-level displacement, all point at the same conclusion. The skills that protect Gen Z from being replaced are the skills AI cannot do. Curricula need to be redesigned around that fact.

Also Read: Beyond the chatbot: How Gen Z pioneers are leading ASEAN’s new AI revolution

The institutions getting this right

A few institutions are quietly doing this. IBM tripled its entry-level hiring in 2026 specifically to rebuild the apprenticeship layer that produces senior judgement. Some companies are building internal academies that pair AI fluency with structured customer exposure and accountability training. The best programmes match the AWS-Pearson definition of a well-positioned institution. Agile curriculum, deep industry connection, applied experience built into the structure, and outputs measured against the compound skills employers actually require.

These institutions are the exception. Most universities and bootcamps are still pricing AI literacy as the answer when employers have been telling them, in increasingly direct language, that it is the wrong answer.

The next two years will sort graduates into two categories. The ones who can prompt, and the ones who can decide. The market will pay both, but at very different rates and with very different security. Gen Z entering the workforce in 2026 needs to understand which side of that line they are graduating onto, and what they can do about it before it is too late.

The skill no AI bootcamp is teaching is the skill that will decide their careers.

—

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 Gen Z doesn’t need more AI courses, it needs the skills AI can’t replicate appeared first on e27.