Malaysia is emerging as one of Southeast Asia’s clearest winners from the artificial intelligence infrastructure boom, but the benefits are not spreading evenly across its semiconductor industry, according to HSBC Global Investment Research.

In a report released last week, the research house said Malaysia, alongside Singapore and Vietnam, stands out as a primary regional beneficiary of the AI-driven technology upcycle. The evidence is visible in the country’s surging chip exports and its deepening electronics trade flows with the United States, mainland China, Taiwan and neighbouring Singapore.

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

Those flows matter because Malaysia sits at a critical point in the global chip supply chain. Fabricated chips are often shipped into the country for assembly, testing and packaging, or ATP, before moving on to device makers, cloud infrastructure providers and end customers. This back-end role has made Malaysia indispensable to global semiconductor production, even if the highest-margin parts of the industry still sit elsewhere.

But HSBC’s central point is more nuanced than a simple “AI boom lifts all boats” story. Malaysia is gaining from the global race to build data centres, train large AI models and secure computing capacity, yet the upside is concentrated among companies with direct exposure to AI-related demand.

A boom with uneven rewards

Firms linked to AI infrastructure, including selected chipmakers, advanced packaging providers and data centre operators, are seeing stronger demand as hyperscalers and technology companies continue to spend heavily on computing power.

The picture is less straightforward for companies tied to traditional consumer electronics. For them, the same AI boom can become a cost problem. Strong demand for memory chips, for instance, can push up input prices and squeeze manufacturers that rely on those components but do not directly benefit from AI-related orders. In some cases, higher memory costs could even slow production.

This split is important for Southeast Asia. The region’s electronics sector is often discussed as a broad beneficiary of supply-chain diversification and AI demand. In practice, exposure varies widely by product, customer base and position in the value chain. Malaysia’s semiconductor sector is large, but not every participant is equally plugged into the most profitable parts of the AI cycle.

HSBC said broader operating conditions remain resilient. The global electronics Purchasing Managers’ Index eased to 55.3 in July from 55.7 in June, while the Asia electronics PMI slipped to 54.7 from 55. Both remain firmly above the 50 mark that separates expansion from contraction, and are still ahead of their 12-month averages of 52.5 and 53.6, respectively.

That suggests the cycle remains healthy, even as pressure builds in specific parts of the supply chain. HSBC noted that input and output prices eased slightly in July and supplier delivery times improved marginally, but price pressures remain elevated and order backlogs are still growing.

The helium risk

A second concern is supply risk, particularly around key inputs used in chipmaking. HSBC highlighted lingering uncertainty linked to the Middle East conflict and its potential effect on critical materials such as helium.

Helium is used in parts of semiconductor manufacturing, especially in front-end wafer fabrication, where chips are created on silicon wafers. Any disruption to supply can therefore create complications for countries trying to expand fabrication capacity.

Malaysia is partly insulated because its largest semiconductor strength remains ATP, which depends less on helium-intensive processes and more on nitrogen. The country also has substantial domestic nitrogen production. Still, the risk is not irrelevant.

“Fortunately for Malaysia, ATP relies less on helium-intensive processes and more on nitrogen, for which the country has substantial domestic production,” HSBC said. “However, Malaysia’s wafer fabs do rely on helium, which means supply management still matters.”

Also Read: Malaysia’s GreatAsic raises US$6.9m to pivot nation from chip assembly to indigenous design

That distinction captures Malaysia’s current position well. Its dominance in back-end activities gives it resilience, but its ambition to move into more advanced front-end manufacturing exposes it to a different set of operational and geopolitical risks.

Moving up the stack

Malaysia’s long-term challenge is not simply to attract more semiconductor investment, but to capture more sophisticated parts of the industry. The government’s National Semiconductor Strategy, announced in 2024, commits US$6.12 billion and targets the training of 60,000 highly skilled local semiconductor engineers by 2030.

The strategy reflects a broader regional ambition. Southeast Asian economies want to move beyond being assembly bases and become deeper technology hubs. For Malaysia, that means expanding advanced packaging, building more front-end manufacturing capability and nurturing chip design talent.

The difficulty is that foundries are among the most capital-intensive industrial assets in the world. They require vast financing, reliable energy and water supplies, specialised infrastructure, and long-term customer commitments. These hurdles can be addressed over time with incentives and execution, but talent is harder to manufacture quickly.

HSBC flagged engineer retention as a key weakness. Average engineering wages in Malaysia’s manufacturing sector trail those in several Asian competitors, creating a clear risk of talent outflows, particularly to neighbouring Singapore, where pay is significantly higher.

Matching wages with wealthier economies will be difficult. But Malaysia can narrow the gap through targeted grants, tax incentives and schemes that improve the overall value proposition for skilled roles. It can also use targeted immigration policies to ease the talent constraint.

For founders, investors and operators in the region, this is the less glamorous but more decisive part of the semiconductor story. Capital announcements make headlines, but execution depends on whether countries can build and keep enough engineers, technicians and managers to run complex industrial ecosystems.

Malaysia’s geopolitical opening

One advantage Malaysia does have is geopolitical positioning. HSBC said the country’s ability to maintain constructive ties with both the US and China, while complementing Singapore’s more constrained land and resource base, could help it attract diversified foreign investment.

This matters as multinational chipmakers reassess location risk. Taiwan remains central to global semiconductors, South Korea is a memory powerhouse, and mainland China continues to invest heavily in self-sufficiency. But geopolitical tensions around major Asian chip hubs are forcing companies to think harder about redundancy and resilience.

Malaysia’s neutral reputation could therefore become a more valuable asset. Between January 2024 and March 2026, its semiconductor sector secured about US$22.52 billion in approved investments, including roughly US$20.31 billion in foreign direct investment.

The country’s roots in semiconductors go back more than five decades. Intel opened a chip assembly plant in Penang in 1972, shortly after Singapore entered the industry. Since then, Malaysia has grown into a major back-end semiconductor hub, accounting for 13 per cent of the global ATP market.

It is also, together with Singapore, one of only two ASEAN economies with front-end fabrication capabilities. Malaysia has particular strength in automotive power semiconductors, supported by German chipmaker Infineon’s manufacturing footprint in the country.

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

The scale is already substantial. Malaysia exported nearly US$110 billion of semiconductors in 2025, equal to around 23 per cent of gross domestic product.

The AI boom gives Malaysia a powerful tailwind. But HSBC’s report suggests the next phase will be harder than riding export momentum. The country must manage supply risks, avoid a two-speed industry, and solve the talent problem if it wants to move from being a crucial assembly hub to a higher-value semiconductor power.

The post AI demand lifts Malaysia’s chip sector, but not every player wins appeared first on e27.

Bitcoin currently faces a highly complex market environment that puzzles many seasoned investors. The premier digital asset struggles to break above the US$70,000 price level last observed in May. A closer look at market indicators reveals deep underlying weakness despite favourable macroeconomic conditions.

The Coinbase Bitcoin Premium Index calculates the price difference between Bitcoin on Coinbase and Binance to estimate buying or selling pressure from the United States market. This specific index has stayed negative for 90 consecutive days. At the time of writing, the index stood at -0.1066 per cent.

A negative reading indicates that the asset trades at a lower price on Coinbase than on Binance. This persistent discount highlights a profound lack of domestic buying interest. This prolonged negative premium is a glaring warning sign for the broader cryptocurrency sector. Smart money clearly anticipates further downside risk and refuses to accumulate more digital assets at current valuations.

This steep decline occurred while broader financial markets celebrated new record highs. The Relative Strength Index remained largely below the neutral level, reflecting prevailing bearish sentiment. Bollinger Bands further supported the volatility that prevented the price from hitting a high bullish threshold. Even massive accumulation by large holders failed to reverse the downward trend.

Whale wallets bought 54,000 more coins since mid-June, but the price action ignored this aggressive accumulation. Buy-side support below the current price continues to erode rapidly. A significant concentration of buy orders below the market existed earlier, especially in June. This created a solid floor because buyers were prepared to absorb selling pressure if the asset dropped toward those levels.

Market participants have now removed or shifted many of those bids lower, leaving fewer orders directly beneath the price. The market liquidity buffer has weakened significantly with less buy-side support to cushion further declines. I consider this lack of underlying bid depth a major structural vulnerability. Order book dynamics clearly show that large players are stepping away from defending current valuation levels.

Also Read: Pokemon cards gained 22.8% while Bitcoin lost 20.7% and that gap should worry every investor

The digital currency had nearly everything going its way this week but remains on track to finish roughly three per cent lower. This divergence strikes a particularly discordant note, given the asset’s reputation as a high-beta proxy for technology stocks. Wall Street pushed to fresh record highs as inflation cools and traders dial back expectations for a Federal Reserve rate hike in September.

These conditions normally favour speculative assets. The digital currency fell from around US$65,000 on Monday to US$62,470 by Friday. The tech-heavy Nasdaq 100 closed the week approximately one per cent higher during the exact same period.

Last week produced a clean dovish signal, combining cooler inflation with a weakening labour market, as reflected in favourable producer price index and jobless claims data. This refusal to follow traditional equities is deeply concerning for momentum traders. This distinct decoupling suggests that internal market mechanics currently overpower external macroeconomic stimuli. The asset faces its own distinct demand problem, setting it apart from the broader stock market rally.

Michael Saylor serves as the executive chairman at Strategy, which holds the record as the largest public company holding this asset. He offered the clearest explanation for this divergence earlier this month. Saylor noted that an enormous amount of capital is currently flowing into artificial intelligence infrastructure. Companies such as Alphabet, Meta, and SpaceX represent the largest near-term headwinds for the digital currency.

The premier cryptocurrency and artificial intelligence currently compete for the exact same speculative and institutional capital. Artificial intelligence is winning this battle for investor attention right now. This massive capital rotation explains why the digital currency refuses to participate in the broader equity rally. I believe this technological distraction will continue suppressing digital asset prices until the artificial intelligence hype cycle naturally cools down.

Institutional investors simply prefer the tangible revenue growth of technology giants over the speculative store-of-value proposition during uncertain economic times. This sector rotation severely limits the liquidity available to alternative assets seeking robust capital inflows. Wall Street allocates billions to data centres rather than decentralised ledger networks.

Also Read: Is the US$63,750 line the only thing standing between Bitcoin and US$62,000?

Exchange-traded funds further illustrate this lack of institutional enthusiasm. United States spot exchange-traded funds recorded US$5.48 billion in net outflows in 2026. These funds have only recovered US$459.6 million so far in August, as of August 14. This massive capital exodus confirms that large funds are reducing their exposure.

The digital currency formed a smaller bear pennant around US$60,000 to US$65,000 since the June selloff. This formation represents another bearish continuation pattern that technical analysts monitor closely. A decisive break below the rising support of this pennant could accelerate the existing flag breakdown. The measured move points toward approximately US$46,300.

That calculation puts the broader downside target zone at roughly US$45,000 to US$52,000. I expect the market to test these lower support levels before finding any meaningful long-term stability. Traders must respect these technical breakdown signals and adjust their risk management strategies to protect their portfolios from sudden drawdowns. Chart patterns rarely lie, and this specific setup screams further downside action for anyone paying close attention.

The broader economic backdrop adds another layer of complexity to this situation. The United States national debt currently nears US$40T. This massive fiscal burden forces the government to issue more bonds, which drains liquidity from the financial system.

I argue that this expanding debt ceiling inherently restricts the amount of excess capital available for highly speculative assets. The combination of massive artificial intelligence investments and soaring national debt creates a perfect storm that suppresses digital asset valuations. Investors must recognise that the digital currency no longer moves in lockstep with traditional risk assets.

Market participants should prepare for increased volatility and potentially lower prices in the coming weeks. Prudent traders will likely hedge their portfolios against these impending macroeconomic shocks. The digital asset must overcome these significant structural headwinds before it can resume its historical upward trajectory. Careful observation of order book depth will provide the next major clue. Global liquidity constraints will dictate the next major move.

—

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 post The US$46,300 question: How low can Bitcoin go before buyers return appeared first on e27.

For years, Southeast Asia’s digital economy has grown on top of technologies built elsewhere. Cloud infrastructure, operating systems, search, social media, e-commerce tools and, more recently, large language models have largely come from the US and China. The region adapted quickly, but rarely controlled the deepest layers of the stack.

Artificial intelligence is forcing governments and founders to revisit that bargain.

Also Read: The unexpected ways AI is already changing Malaysia’s economy

Malaysia’s National AI Action Plan 2026-2030, also referred to as AI Nation 2030, places this question at the centre of its digital strategy: can the country move from being a buyer of AI tools to a producer of AI products, infrastructure and talent? The plan’s answer is built around sovereign AI — the ability to build, train and deploy AI using domestic infrastructure, data and workforce capabilities.

This is not just a policy slogan. For startups, it could become a practical competitive advantage. In a region as linguistically and culturally complex as Southeast Asia, models trained primarily on foreign datasets often miss local nuance. That gap matters in areas such as education, public services, healthcare, agriculture, financial inclusion and legal compliance, where context is not a nice-to-have but the product itself.

Local context as a moat

The strongest argument for sovereign AI is not that every country needs to rebuild OpenAI, Google or Anthropic from scratch. It is that global models, however powerful, are not always designed for local realities.

Malaysia’s plan recognises this through its Secure and Localised AI Ecosystem initiative, which calls for trusted local datasets and indigenous AI models. One priority is the development of large language models fluent in Bahasa Melayu and aligned with Malaysian cultural norms, including through collaboration with institutions such as Dewan Bahasa dan Pustaka.

For founders, this creates room to build where global platforms are weakest. A generic chatbot may perform adequately in English-language customer support, but it may stumble when handling Bahasa Melayu, Manglish, code-switching, dialects, religious sensitivities or public-sector terminology. In classrooms, government offices or rural advisory services, those mistakes can damage trust.

The same logic applies across Southeast Asia. Indonesia, Thailand, Vietnam and the Philippines all face similar challenges: large populations, diverse languages, uneven digital literacy and public services that need to work beyond metropolitan centres. A model that understands a population’s language, regulations and social context can outperform a larger but less grounded system in specific high-value use cases.

This is where local startups may find their wedge. They do not need to win the global foundation-model race. They can build domain-specific AI tools that combine global advances with local data, workflows and compliance requirements.

From AI adoption to AI production

Malaysia’s ambition is also economic. AI Nation 2030 aims to place the country among the top 10 in global AI indices, add an incremental 1.2 percentage points to GDP growth and create 300,000 new jobs by 2030.

Those are aggressive targets, but they reflect a broader shift in Southeast Asia’s thinking. The region’s internet economy has grown rapidly — Google, Temasek and Bain estimated it at US$263 billion in gross merchandise value in 2024 — yet much of the value still accrues to platform owners, cloud providers and chipmakers outside the region.

Also Read: How can Malaysia leverage AI for growth and not see it as a threat?

Malaysia wants a larger share of the value chain. Its plan uses public-private partnership “Impact Engines” to push local firms into higher-value AI segments, rather than leaving them as downstream users of foreign tools. One example is the AI Hub for the Manufacturing Ecosystem, which builds on Malaysia’s existing strength in semiconductors and electronics.

This is a sensible starting point. Malaysia is already part of the global chip supply chain, particularly in assembly, testing and packaging. Applying AI to improve yields, predict maintenance issues, optimise energy use and manage supply chains could help local manufacturers move up the ladder. It may also make the country more attractive to high-value foreign direct investment at a time when companies are diversifying supply chains across Asia.

For startups, the opportunity lies in the middle layer: AI applications for factories, logistics providers, farms, banks, schools and government agencies that need solutions adapted to local operations. These are not always glamorous markets, but they are often where durable revenue is built.

Data and compute decide who gets to build

Sovereign AI ultimately depends on two scarce inputs: data and compute.

Malaysia’s plan proposes an AI-ready data ecosystem that aggregates priority datasets across sectors and turns them into trusted, purpose-driven data products. A National Data Exchange would allow startups and institutions to discover and license datasets under clearer terms.

If executed well, this could address one of the biggest constraints for AI startups in the region. Many founders can access open-source models, but not the high-quality local datasets needed to make those models useful. Data is often fragmented across ministries, state agencies, corporates and legacy systems. Access can be slow, opaque or legally uncertain.

The compute side is just as important. Training and fine-tuning AI models requires expensive hardware, and reliance on foreign cloud providers can become a strategic vulnerability for governments handling sensitive data. Malaysia’s proposed National Supercomputing Centre and AI Sukuk financing mechanism are designed to expand access to high-performance compute for local innovators.

The key question will be implementation. If data access remains bureaucratic or compute is captured by large incumbents, startups will see little benefit. But if the infrastructure is affordable and fairly governed, it could reduce the entry barrier for Malaysian AI companies and research teams.

Trust as a market advantage

AI adoption will also depend on public confidence. Malaysia’s plan includes a governance framework aligned with the National Guidelines on AI Governance and Ethics, alongside an AI Trust Function to monitor and respond to risks. It also proposes a National AI Classification mechanism for “Made by Malaysia” AI products.

For startups, this may sound like more compliance work. In practice, clear rules can help. Regulated sectors such as finance, healthcare, education and public services will not adopt AI at scale without confidence that systems are safe, explainable and accountable.

A national classification framework could become a trust signal for buyers, particularly if Malaysia wants its AI products to travel across ASEAN. The region is still early in building interoperable AI governance. A credible Malaysian certification could help local companies sell into neighbouring markets that face similar concerns but may not have the same institutional capacity.

Inclusion will determine the outcome

Malaysia’s sovereign AI strategy is not only about building elite technology. It also targets micro, small and medium enterprises, which make up 84.4 per cent of businesses in the services sector. These firms often lack specialist talent and cannot afford bespoke AI systems. The AI for MSMEs initiative proposes modular, pre-vetted tools embedded into platforms they already use.

Agriculture is another test case. A scalable agristack using local data could support precision farming, weather prediction and crop advisory services. For a country seeking stronger food resilience, AI will only matter if it reaches farmers, cooperatives and small suppliers, not just large agribusinesses.

This inclusive angle is important for Southeast Asia. If AI primarily benefits large companies in capital cities, it may deepen existing divides. If it improves productivity for small businesses, schools, clinics and farms, it could become a broader development tool.

Also Read: AI demand lifts Malaysia’s chip sector, but not every player wins

The final piece is talent. Malaysia’s plan includes a Global Talent Network, Digital e-Residency Pass, AI Talent Concierge and large-scale reskilling efforts. These measures acknowledge a hard truth: sovereign AI cannot be built with infrastructure alone. The country needs engineers, product managers, policy specialists, domain experts and teachers who understand both technology and local problems.

Malaysia’s sovereign AI push is ambitious, and many parts remain dependent on execution. But its underlying bet is sound: in AI, local context is not a weakness. It may be the clearest moat Southeast Asian startups have.

The post Malaysia’s sovereign AI bet: Local context becomes the next startup moat appeared first on e27.

Southeast Asia’s financial services dealmaking held its ground in the first half of 2026, even as the total value of transactions dropped sharply, suggesting that buyers remain active but more selective in a market still shaped by high uncertainty.

According to EY’s latest financial services M&A analysis, the region recorded 31 publicly disclosed mergers and acquisitions in the first six months of the year, unchanged from the same period in 2025. But disclosed deal value fell to US$936 million from US$1.6 billion a year earlier.

Also Read: ‘M&A process in SEA is stuck in the dark age’: say match.asia co-founders

The headline number points to a market that has not frozen, but has become more cautious. Rather than the large, transformative deals that dominated parts of the previous cycle, Southeast Asia’s financial services activity in early 2026 was led by smaller and mid-sized transactions. For banks, insurers, fintech investors and asset managers, the focus appears to be on assets that can fill specific strategic gaps rather than aggressive expansion at any price.

“The stability in deal activity across Southeast Asia’s financial services sector reflected a market that remains engaged despite economic and geopolitical volatility,” said Stuart Last, EY-Parthenon Partner, Financial Services, Ernst & Young Solutions LLP.

He added that the fall in disclosed deal value suggests investors are pursuing opportunities with a clear strategic rationale, rather than chasing scale for its own sake.

Banks slow, wealth platforms gain ground

The split across sub-sectors shows how investor attention is shifting within the region’s financial services market.

Banking and capital markets remained the largest contributor by value, but activity declined. Deal volume in the segment fell to 14 from 20 a year earlier, while disclosed value dropped to US$669 million from US$1.1 billion.

That decline is not surprising. Banking deals in Southeast Asia are often shaped by regulation, ownership limits and the complexity of integrating legacy systems. While the region’s banks are still under pressure to digitise, improve cost efficiency and compete with fintech players, full-scale acquisitions can be difficult to execute, especially when interest rates, credit risk and capital requirements remain in focus.

Insurance moved in the opposite direction by deal count. The sector recorded nine deals in the first half of 2026, up from eight a year earlier. But disclosed deal value fell to US$123 million from US$478 million, indicating that activity was concentrated in smaller assets.

The more striking change came from wealth and asset management. Deal volume rose to eight from three, while disclosed value jumped to US$145 million from just US$800,000 in the first half of 2025.

That rise reflects one of Southeast Asia’s most persistent financial services themes: the growth of the affluent and mass-affluent population. As income levels rise in markets such as Singapore, Indonesia, Vietnam, Malaysia and Thailand, more consumers are seeking investment products, retirement planning tools and advisory services beyond traditional savings accounts.

For acquirers, wealth platforms can offer access to sticky customer relationships, fee-based revenue and digital distribution channels. In a region where financial literacy and investment participation are still uneven, firms that can combine trust, technology and local market access are becoming more attractive targets.

Last said the sharp rise in wealth and asset management deal value points to growing investor interest in platforms and capabilities that can capture demand from the region’s expanding affluent population.

Foreign buyers remain interested in Southeast Asia

EY’s data also suggests that international appetite for Southeast Asian financial services assets has not disappeared.

The number of non-Southeast Asian firms acquiring targets in the region fell to five in the first half of 2026 from seven a year earlier. However, the total disclosed value of these deals rose to US$410 million from US$344 million.

Also Read: How M&A can supercharge your startup’s success

That means fewer foreign acquirers were active, but those that did move were willing to commit larger sums. This matters because Southeast Asia continues to be viewed as a long-term growth market despite near-term volatility. The region has a young population, rising digital adoption, a large underbanked base in several markets, and increasing demand for credit, insurance and wealth products.

At the same time, operating across Southeast Asia is rarely straightforward. The region is not a single market. Financial services firms must navigate different regulators, licensing regimes, consumer behaviours, languages and levels of digital infrastructure. That complexity can slow dealmaking, but it can also make established local platforms more valuable.

EY expects larger transactions to return in the second half of 2026 if financing conditions improve and more scaled assets become available.

Global deal count rises, but megadeals thin out

The Southeast Asian pattern mirrors a broader global trend: more deals, but less value.

Globally, banks, insurers and asset managers publicly disclosed 1,137 financial services deals in the first half of 2026, up 3 per cent from 1,101 a year earlier. Yet total disclosed deal value fell to US$134.5 billion from US$191.3 billion.

The drop was largely driven by a thinner pipeline of megadeals. EY recorded 25 transactions above US$1 billion in the first half of 2026, representing 80 per cent of total deal value. That compares with 37 such deals in the first half of 2025 and 55 in the second half of 2025.

The concentration of value among the largest transactions remained high. The ten biggest global financial services deals accounted for US$78.7 billion, or 58 per cent of total value. The top 20 deals accounted for US$100.5 billion, or 75 per cent.

Omar Ali, EY Global Financial Services Leader, said financial services firms have adapted to heightened uncertainty as part of normal operating conditions. But he noted that unpredictability, slower global growth, inflation and supply shocks continue to affect deal value.

“Despite the number of transactions rising, deal value in the first half this year across the world’s major markets is down on 2025 levels, as significantly fewer transactions completed over the US$1 billion mark,” he said.

Asia and Oceania weaken, but cross-border interest grows

Across Asian and Oceanian markets, the first half of 2026 was softer than in Southeast Asia. Publicly disclosed financial services M&A fell 14 per cent to 147 deals from 170 a year earlier. Total disclosed value slipped to US$15.8 billion from US$17.8 billion.

Also Read: M&A in Asia: A strategic roadmap for venture builders

Banking and capital markets deal volume in the broader region declined to 77 from 87, though deal value rose to US$11.3 billion from US$6.4 billion. Insurance weakened more clearly, with volume falling to 31 from 41 and value dropping to US$2.1 billion from US$5 billion. Wealth and asset management also declined, with volume falling to 39 from 42 and value sliding to US$2.4 billion from US$6.5 billion.

However, foreign interest in Asian and Oceanian targets increased. Non-regional acquirers completed or announced 28 deals, up from 23, while disclosed value rose to US$1.9 billion from US$1.6 billion.

For Southeast Asia, the message is mixed but not gloomy. Dealmakers are not retreating from the region. They are becoming more disciplined, more sector-specific and more careful about valuation. The next phase of activity may depend less on whether buyers have appetite, and more on whether sellers are willing to meet the market.

The post Wealth management emerges bright spot in Southeast Asia financial services M&A appeared first on e27.

Volume went up. Quality didn’t. The bottleneck moved somewhere nobody is looking.

Every marketing team I speak to in Southeast Asia has the same story about generative AI: we make ten times more creative than we did eighteen months ago, and we’re not convinced any of it is better. That instinct is now measurable. New research from WARC with TikTok and LIONS Advisory, reported by PPC Land, surveyed 400 marketers and found that while 88 per cent are producing more creative since adopting AI, only 45 per cent see a meaningful lift in quality.

The interesting part isn’t the gap. It’s the reason for it. Two-thirds of those marketers said they brief generative models primarily with demographic data. Nearly six in ten of the same group said demographic segmentation no longer works. They are feeding the machine inputs they have already told researchers are broken. Only 17 per cent consistently brief with anything richer — community context, behavioural signal, actual audience tension.

The report’s framing is the sharpest line in it, and it deserves repeating in plain terms: this is not a technology gap. It is an intelligence gap.

The evaluation layer nobody costed

Here is what changed, and why it is easy to miss.

A brief has never really been a document. It has been the start of a conversation. It went to a strategist, who pushed back on the segment. It went to a creative director, who asked what the person actually feels at the moment of purchase. It went to an art director, who threw out the first three routes. By the time work reached a client, a thin brief had been quietly repaired four or five times by people whose job was, in part, to notice that it was thin.

Those stages were slow, and slowness was the point. They were also expensive, and so they were the first thing compression removed. When production timelines collapse from three weeks to three days, the repair layer goes with them. The brief no longer passes through a series of sceptical humans. It passes into a model, which is constitutionally incapable of scepticism about its own inputs and will produce forty confident variants of a bad idea as readily as forty of a good one.

So the weakness that used to be absorbed by the process now lands directly in the output, at volume, and with the polish of professional work. That is a much worse failure mode than the one it replaced. A bad brief used to produce visibly bad work that somebody caught. Now it produces plausible work that nobody catches, because it looks fine.

Also Read: Moving past the chatbox: The hidden risks of agentic AI and MCP in enterprise infrastructure

Demographics survive because they are already in the template

Why do teams keep briefing with data they don’t believe in? Not conviction — inertia.

Age brackets and income bands are already sitting in the planning deck. They are already in the media plan, the audience field, the campaign naming convention. They require no new work, no new tooling, no argument with anyone. Behavioural and community insight requires all four. Under deadline, the default wins every time, and the default is a demographic.

GWI made this point publicly last week in a rather good line — that age brackets are the laziest segment in marketing, and the differences inside a generation are larger than the differences between generations. They are right, and it is worth noticing how rare that argument is. Look across a week of published content from the marketing-technology category and a clean division appears. The audience research firms publish findings that stop at the statistic. The workflow and listening platforms publish features that start at the publish button. Almost nobody addresses the space between the two, which is precisely where the intelligence gap lives.

Why Southeast Asia feels this first

Because the compression here is more severe. A regional team in Singapore is routinely running six to eleven markets, in several languages, against budgets that would cover two markets in Europe. Local nuance is not a refinement; it is the entire job. And it is exactly the layer that demographic briefing flattens.

Feed a model “women 25–34, urban, middle income” and it will return something that could run in Jakarta, Manila or Kuala Lumpur and land properly in none of them. The output will be grammatical, on-brand and completely generic. Multiply that by eleven markets and a weekly cadence and you have built a very efficient machine for producing content nobody remembers.

Also Read: No fans, no fridges, just paint: ZERC’s founder on cracking SEA’s cooling crisis

Rebuilding the layer, cheaply

The teams pulling ahead have made one structural change: they treat the brief as the product. Not the deck, not the asset — the brief. Whoever controls the quality of the input now controls the quality of everything downstream.

That change is showing up in the tooling at both ends of the pipeline. At the front, platforms that decode live category data into audience tensions and evidence-backed briefs, so the input carries something observed rather than something assumed. At the back, a newer class of businesses built purely around execution — Touchigh, for instance, which helps Chinese cross-border sellers reach American buyers by automating both AI-search visibility and native English social content, and scores that content for predicted performance before it publishes rather than reporting on it after.

That last detail matters more than the category it sits in. A young execution company, serving SMEs at a few hundred dollars a month, has independently arrived at the same conclusion: the judgement call has to happen before the asset ships, or it doesn’t happen at all. Front end and back end are solving different halves of one problem — stopping weak inputs entering the system, and stopping good ones degrading on the way out.

What the operating numbers suggest is that the repair layer can, in fact, be rebuilt cheaply enough to survive a deadline. Pitching a UK hospitality group against considerably larger shops, the agency SAMY documented category planning collapsing from weeks to roughly thirteen minutes per brief, with predicted click-through accuracy running about three times better than human estimation. They won the retainer.

Read that carefully, because the speed is not really the story. The story is that the judgement nobody could afford is now something you can afford on every single brief — and a judgement that only happens when there’s time for it isn’t a standard, it’s a luxury.

The question worth sitting with

The uncomfortable version of the WARC finding is that most teams already know their inputs are wrong and ship on them anyway, because fixing the input costs more this week than shipping the output does.

That maths is changing. When evidence-grade audience intelligence takes minutes rather than weeks, the excuse for briefing on a demographic quietly disappears — and so does the defence when the work underperforms.

So: of the briefs your team wrote this quarter, how many could you trace back to something you actually observed about the audience, and how many were the template with a new date on top?

—

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 Your AI isn’t producing bad creative, your brief is appeared first on e27.