Every HR Team wants AI in their workflow now.

The conversations are happening in boardrooms, in leadership offsites, in Slack channels where someone has shared the latest article about what AI is doing to HR. The question is no longer whether to adopt AI. It is when, and how fast. That urgency is real, and it is not wrong.

But there is a problem sitting underneath all of it that almost nobody is talking about. And the companies that miss it are going to spend a lot of money to find out the hard way.

AI does not fix bad data. It amplifies it.

The overconfidence trap

Here is a number that should make every HR leader pause. 

9 out of 10 HR leaders across Asia say their organisation has a single source of truth for employee data.  

Only 26% are actually running on a unified platform.

That is not a small gap. That is a 65-percentage-point gap between what most companies believe about their data and what their infrastructure actually supports. And it is the most important number in business technology right now, because every AI investment decision being made on top of that belief is being made on false ground.

This is the overconfidence trap. 

It does not feel like a trap. Your payroll runs. Your leave balances update. Your reports come out on time. The data feels fine because, within each individual tool, it often is. The problem is what happens when those tools need to talk to each other. When data needs to move across systems, stay consistent, and be read as one complete picture of your workforce, that is when the cracks appear.

And that is exactly where AI will fail you.

 What the infrastructure actually looks like

73% of companies in the region run two or more HR tools. Not one unified system. Multiple tools, each managing a different piece of the people function, each holding a slightly different version of the same data.

18% are still running payroll and people management primarily in Excel.

Think about that in the context of an AI investment conversation. One in five HR teams in this region is being asked to adopt AI while their foundational data lives in a spreadsheet.

46% of HR teams report that duplicate data entry is a routine part of their work. The same information, entered into more than one system, on a regular basis. Every time that happens, there is an opportunity for inconsistency to enter the picture. A small discrepancy today becomes a larger one over time. And over time, it becomes a dataset that looks usable but is analytically unreliable.

 This is the environment that most companies in Asia are planning to deploy AI into.

Also read: Omni HR publishes first independent AI readiness research report across APAC HR

The real cost of getting this wrong

50% of HR leaders already say that fragmented data is limiting their ability to adopt AI right now. Not in the future. Today.

But here is the part that does not get talked about enough. The other 50% may simply not have reached the point where they have tested it yet. The infrastructure data suggests many of them will.

Because when AI is deployed on fragmented data, the outputs look credible. The recommendations come through. The dashboards fill up. It just does not work the way you expected. The insights are partial. The recommendations miss context. And the people using it start to lose confidence in it, quietly, over time.

Only 21% of HR leaders in Asia currently trust AI recommendations enough to act on them without manually checking the output first.

Four in five leaders, when they receive an AI recommendation, review it, verify it, or override it.

That is not a technology problem. That is a data problem. And until the data problem is solved, the technology problem will never go away.

The companies that figure this out before they deploy are the ones that will get compounding returns from AI. Faster decisions. Better retention data. Leaner HR teams that spend less time reconciling information and more time acting on it. The companies that deploy first and figure it out later will spend the next 18 months wondering why the results do not match the promise, and eventually have to rebuild the foundation they skipped, at a higher cost.

That is how AI is going to sort companies into two groups. Not the ones that adopted early versus the ones that adopted late. The ones that built the right foundation versus the ones that did not.

Download the State of AI in HR 2026 Report | The first independent study of AI readiness across HR teams in Asia. 402 HR leaders. Free to download.

What readiness actually requires

When the same HR leaders were asked what AI actually needs before it can deliver real value, the answers were consistent.

70% said data accuracy. 64% said system integration. Skills, training, change management, all followed at a significant distance.

The top two prerequisites for AI readiness are both infrastructure problems. Not technology problems. Not budget problems. Not change management problems. Infrastructure.

HR leaders already know this. The question is whether the investment decisions being made right now reflect that understanding. Because the data suggests that for most companies, the consolidation and cleanup work is still catching up to the ambition sitting on top of it.

The companies that close that gap first are not just going to get better AI results. They are going to build a structural advantage that compounds over time, because every additional year of clean, connected, unified data makes the AI models sitting on top of it more accurate and more useful.

Also read: Omni HR acquisition MajuHR to boost chat-native capabilities

The question worth asking before the next AI meeting

Before the next conversation about which AI tool to buy, which vendor to pilot, which function to automate first, there is one question that is worth asking out loud.

Do we actually know where our data lives? All of it? Is it consistent? Is it connected?

For 91% of companies in Asia, the honest answer is that they believe it is. For 74% of them, the data says something different. 

The research on what AI readiness actually looks like for companies in this part of the world, the benchmarks, the gaps, and the steps that matter, is in the full report below. 

Download the State of AI in HR 2026 Report | The first independent study of AI readiness across HR teams in Asia. 402 HR leaders. Free to download.

 

—

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

The e27 team produced this article sponsored by Omni HR

We can share your story at e27 too! Engage the Southeast Asian tech ecosystem by bringing your story to the world. You can reach out to us here to get started.

Featured Image Credit: Canva Images

About the research

The State of AI in HR 2026 report surveyed 402 HR professionals at the manager level or above across Singapore and the Philippines between January and March 2026. The study was conducted independently by Omni HR and covers HR technology infrastructure, AI adoption intent, data readiness, and organisational priorities.

About Omni HR

Omni HR is a modern, all-in-one HRIS and multi-country payroll platform built for Asia’s fastest-growing companies. www.omnihr.co

The post You think your company is ready for AI. Your data says otherwise. appeared first on e27.

A lot of founders talk about funding as if all money is meant to do the same job. It is not. Growth capital, grant funding, venture funding, debt, patient capital, and catalytic capital all serve different purposes. When founders treat them as interchangeable, the financing strategy becomes weak.

This is especially true for impact ventures in Southeast Asia, where many companies work on problems that take longer to prove. These ventures may operate in health, climate, agriculture, education, financial inclusion, infrastructure, or livelihoods. In these sectors, traction is not always as simple as revenue growth or user numbers. A founder may also need to prove field performance, community trust, institutional interest, partner readiness, or measurable outcomes.

This is where catalytic capital fits.

Catalytic capital is best understood as proof capital. It helps a venture move from one stage of credibility to the next. It is not meant to be the final answer, and it should not be treated as a permanent runway. Its real job is to help a serious venture prove something important enough that a more durable source of funding, partnership, or institutional support becomes possible.

That proof can take different forms. For one venture, it may be a pilot with a credible partner. For another, it may be early field evidence. For another, it may be regulatory progress, a stronger operating base, better outcome data, or validation from customers, hospitals, schools, farmers, public agencies, or local institutions. The milestone will differ, but the principle is the same. Catalytic capital should move the venture toward a more financeable position.

This is why catalytic capital is often linked to impact. Impact ventures usually create value before they can fully capture value. A climate venture may need to prove deployment in difficult local conditions before commercial capital becomes comfortable. A health venture may need evidence and trust before larger funders take it seriously. An education or livelihoods venture may need to show real outcomes before institutional partners step in. In many cases, normal commercial capital arrives too late, while venture capital may demand a speed of growth that the model cannot responsibly support.

Also Read: Breaking the two-speed economy: How integrated ERM unlocks capital for the real sector

Catalytic capital can fund the early work that reduces that risk. It can support pilots, evidence, partnerships, and operational readiness. It can help the founder move from a mission-driven story to a more evidence-based case. Used well, it gives the venture a stronger reason to approach the next capital source. Used badly, it only delays the next funding problem.

This is where many founders make the mistake. They treat catalytic capital as breathing space rather than a bridge. Breathing space only gives the venture more time. A bridge moves the venture somewhere specific. If the capital only extends the runway but does not create evidence, sharpen the model, improve partner trust, or open a realistic financing path, then the company has not become stronger. It has only bought time.

That distinction matters in the current Southeast Asian market. Investors, funders, and strategic partners are more selective. They are looking more closely at fit, evidence, capital efficiency, and the logic of the next milestone. A good mission is not enough. Founders need to explain why a particular type of capital fits their current stage and what it is expected to unlock.

The better question is not, “Can we raise something?” The better question is, “What does this money need to prove?” If the answer is evidence, then the capital should buy evidence. If the answer is a pilot, then it should buy a pilot. If the answer is partner trust, then the capital should help build that trust. If the answer is readiness for a larger funder, then the capital should close the gaps that are preventing that funder from saying yes.

This sounds simple, but it is often missed. A founder may raise a small grant, prize, fellowship, or catalytic cheque, and then absorb the money into general operations. Salaries, travel, marketing, events, product fixes, and scattered outreach consume the budget. Six months later, the company is still in roughly the same position. No stronger evidence. No clearer partner. No sharper capital story. No better next step. That is not a catalytic use of capital. That is a soft runway.

Also Read: Funded: The quieter capital path founders keep missing

There is nothing wrong with the runway. Every venture needs time. But if a founder calls the money catalytic, then it has to catalyse something. It should not just fund activity. It should fund progress. Not simply “we will run more programs,” but “we will prove this model works with this customer group.” Not simply “we will build awareness,” but “we will secure these partners and produce this evidence.”

For impact ventures, the capital stack is rarely simple. A serious company may need different types of money at different points. A grant or catalytic cheque may support early evidence. A corporate partner may support distribution. A development institution may support scale. A patient investor may come in once the model is more stable. Commercial capital may only make sense later, once the risk profile has changed.

Many founders do not fail because they cannot raise any money. They fail because they chase the wrong money at the wrong time. They pitch venture capital when they still need evidence. They chase grants without knowing what milestone the grant should unlock. They approach banks before they have cash flow. They talk to institutions before they have the operating base to absorb institutional capital.

Catalytic capital is not the destination. It is the capital that helps a serious venture earn the right to reach the next destination. Used properly, it can bridge the gap between purpose and proof. Used lazily, it becomes another form of drift. In this market, that difference matters. The founders who understand it will not just raise money. They will become more financeable.

The post Catalytic capital is not free runway but proof capital appeared first on e27.

Money is one of humanity’s most powerful technologies. Every time it evolves, economies don’t just grow — they reset.

At its core, money is trust made tangible — a shared belief that a piece of paper, a seashell, a token today will hold value tomorrow. Across history, every major shift in money has been tectonic, reshaping societies, redistributing power, and unlocking entirely new economic behaviours.

We are standing at another such moment.

We are entering the Fourth Industrial Revolution — and money must evolve to match it. Tokenisation is that evolution.

Money through history: Scaling trust

Every evolution of money has solved one problem: how to scale trust.

• Barter: Local and personal, but limited by coincidence of wants
• Commodity money: Shells, salt, gold — portable, but inefficient
• Coinage and empires: Standardised, backed by central authority
• Paper money: Trust abstracted into institutions
• Digital money: Fast and global, but still centralised
• Cryptocurrencies and tokens: Trust embedded in code — programmable and decentralised

The pattern is consistent: Money evolves to match the scale and complexity of the economy it serves.

Today’s economy is becoming always-on, digital, and increasingly AI-driven. Traditional money — built for batch processing and intermediaries is increasingly misaligned.

Tokenised money is not just an upgrade. It is a re-architecture.

Tokenisation and the combinatorial economy

Tokenisation does more than split assets into smaller units. It transforms value into programmable building blocks.

A fraction of a solar panel. A streaming royalty. A carbon credit. A loyalty point. These are not just more efficient assets — they are composable primitives.

The real shift is this: when assets become atomic, they can be recombined.

A fraction of property can merge with revenue streams, identity layers, or incentive systems — forming entirely new financial structures. What emerges is not just a more efficient market, but a new kind of economy.

This is not financial innovation. This is a financial composition.

More tokens create more combinations. More combinations create more markets.

Previously unviable ideas become economically feasible. Innovation shifts from creating standalone assets to recombining them dynamically.

The winners will not be those who simply own assets — but those who orchestrate them.

Also Read: Asia’s US$4T tokenisation boom: Why the region will lead the global financial revolution by 2030

Why 2026 is the inflection point

Blockchain is nearing two decades of development, but the last few years have seen an acceleration across all fronts.

• Economic: Institutions like BlackRock and JPMorgan are deploying capital and infrastructure
• Social: A digital-native generation expects seamless ownership and transactions
• Technological: Blockchain infrastructure, wallets, and APIs have matured significantly

For the first time in history, money is not constrained by technology or demand.

It is constrained by policy.

Regulation — the Political dimension of the PEST framework — is now the gating factor. Frameworks like the CLARITY Act in the US and global policy developments are beginning to define digital assets within existing systems.

The shift is subtle but critical: Policy is no longer asking whether to allow digital assets — but how to integrate them.

The bottleneck is no longer innovation. It is permission.

When that clears, adoption will not be gradual. It will be exponential.

Crypto and tokenisation adoption: Moving across all layers

What makes this moment different is not isolated progress — it is the synchronic movement across the entire system.

• Infrastructure:
  Apr 2026 — SWIFT, backed by BBVA, BNP Paribas, and Citi, has launched a blockchain-based cross-border payment ledger integrated with digital asset custody — bringing tokenisation into global financial rails.
• Regulation:
  Mar 2026 — US SEC issued updated interpretations on how securities laws apply to crypto asset-related products, signalling a shift from ambiguity to structured oversight.
• Banking:
  Feb 2026 — Bank Negara Malaysia is piloting stablecoins and tokenised deposits with major banks like Standard Chartered, Maybank, and CIMB, while DBS had earlier tokenised structured notes on Ethereum.
• Ecosystems:
  Nov 2025 — Singapore’s MAS is advancing Project Guardian, a coordinated push between policymakers and financial institutions to unlock asset tokenisation.
• Sovereigns:
  Jan 2026 — Philippines became the first country to publish its national budget on a public blockchain.
• Markets:
  Sep 2025 — Across APAC, on-chain transaction value has tripled in 30 months, from $81B to $244B — signalling real transactional demand from India to Indonesia, Japan to South Korea.

This is no longer speculative momentum. When infrastructure, regulators, banks, and sovereigns move in parallel, the convergence of adoption is inevitable.

Also Read: Real world tokenisation fireside chat with Anndy Lian: Unpacking the landscape

Opportunities for founders and startups

A tokenised, combinatorial economy rewires the playbook for entrepreneurship.

Here’s where founders should be building now:

• Financial infrastructure: Build the rails — compliance, custody, and tokenisation platforms. The opportunity is a “Stripe for tokens” — APIs that turn any asset into a programmable financial object.
• Granular business models: Move beyond subscriptions into real-time economics — per-second billing for compute, pay-per-use APIs, streaming salaries, and dynamic incentives.
• Platforms as economies: Turn products into ecosystems — enabling revenue sharing, creator royalties, and user-owned marketplaces powered by tokens.
• AI + money: Autonomous agents will transact — paying for APIs, data, and compute. They will need wallets, identity, and financial rails. This stack is still largely unbuilt.
• Interface innovation: Wallets and onboarding remain friction-heavy. The winners will abstract complexity — embedding identity, custody, and payments into seamless user experiences.

Every token is a new canvas. Every split creates new building blocks.

Tokenisation isn’t just creating more opportunities — it is creating exponentially more ways to construct them.

Closing thought

From seashells to smart tokens, the history of money is simple: trust keeps scaling — faster, farther, and smarter as well.

In 2026, clarity may be the switch. When the US embarks, the rest of the world will follow — and the floodgates will open.

The rails are built. The users are ready. The institutions are moving.

What remains is the switch.

The question is — will you build before it flips, or after?

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic.

The post From seashells to tokens: Why 2026 could be the inflection point for money appeared first on e27.

It’s OK to talk about your strengths, but it’s so much better when others do the talking for you.

Earned media (unpaid publicity gained organically through third-party endorsement) has become one of the factors most strongly linked to AI search visibility. And in B2B companies in ASEAN, most marketing strategies haven’t caught up.

Singapore ranks second globally for AI adoption, with 60.9 per cent of its working-age population already using generative AI tools. Google’s e-Conomy SEA 2024 report placed Singapore, the Philippines, and Malaysia among the top 10 countries globally for AI-related searches.

The brands appearing often have built an editorial presence that these systems could find, verify, and cite. The companies that earn media coverage strategically can, sometimes rapidly, take pole position.

How does AI search decide which sources to recommend?

AI search tends to recommend sources that appear consistently and credibly across multiple independent publications. That’s why your editorial presence typically matters more to these systems than your on-site SEO (what’s on your website, socials, etc.).

Traditional search returns a list of links; AI search reads multiple sources and synthesises a single answer, and AI/LLM platforms need to judge which sources can be trusted before they can write that answer.

Ahrefs’ analysis of 75,000 brands found that being mentioned on other websites matters roughly three times more to AI search visibility than how many sites link to you (backlinks). Brands with few web mentions are largely invisible to these AI systems, no matter how well their website ranks, and brands with the most mentions earn up to 10x the AI Overview references than the next closest group.

The factors these systems weigh include:

• how often your brand is mentioned independently across the web
• how recently that coverage was published
• whether it includes named expert quotes
• whether the content opens with a direct answer rather than burying the point
• whether specific data backs the claims.

Backlinks, strong blog content, and good service pages still matter plenty, but considerably less than mentions alone.

Why does AI search favour earned media over owned content?

AI systems favour earned media because these systems can’t independently verify what a brand says about itself, but they can check whether outside sources say the same thing. When your brand writes its own content, AI systems see your own version of events – useful, but unverified.

Also Read: Thailand is suddenly on the frontline of a new ransomware wave

When Nikkei Asia, The Straits Times, South China Morning Post, or an analyst report mentions your brand, a credible outside source has made a deliberate choice to include you, and that carries a different weight.

Muck Rack’s What Is AI Reading? report analysed over one million links cited by AI search tools and found that 94 per cent of AI citations come from non-paid sources, with earned media alone accounting for 82 per cent. Generic corporate blogs and lightly produced branded pages rarely make the cut. Well-researched, authoritative owned content performs better – but the bar is higher than most brands realise.

Half of all AI citations come from content published within the last 11 months, which means coverage from three years ago has little sway (unless it’s highly original or authoritative), and a consistent PR program seeking earned media can keep your brand popping up again and again.

How does public relations coverage influence what AI search recommends?

PR builds the independent editorial record that AI systems draw on when deciding which brands to recommend. Every journalist feature, analyst mention, and trade publication article gives these systems another viable reference when someone asks a relevant question.

The specific activities most likely to help include:

• media relations
• thought leadership and bylined articles
• press releases built around data
• analyst relations
• original research
• executive profiling

i.e., anything that puts a named, quotable person and verifiable facts into an independently published source.

That record only helps if you’re building it in the right places. That means a PR team that knows the terrain and isn’t working on now-ancient paradigms.

Getting your brand into those publications, or the trade outlets AI engines trust in your sector, requires choosing them deliberately, not running broad outreach campaigns.

What makes content more likely to get cited by AI search?

The content most likely to get cited has three qualities. It:

• quotes named experts
• opens with a direct answer to a clear question
• includes specific data

Named expert quotes get cited more often. A direct opening matters because AI systems pull from the first paragraph more than anywhere else in a piece. Specific numbers and verifiable facts consistently outperform general storytelling in citation rates.

Where you publish matters as much as what you publish. A Stacker/Scrunch study of 87 stories across 30 brands, tested across 8 AI platforms, found that placing the same content across multiple trusted news outlets more than tripled how often AI systems cited it, compared to publishing only on the brand’s own site. I’ve seen this in practice, working with a Japan-based SaaS company and a Japanese medical device manufacturer, building editorial coverage across regional and international business press with named expert commentary and specific data in every placement.

Also Read: Hiring creatives in the AI age: Skills over titles

How do you measure whether public relations is working for AI search?

Citation presence – how often your brand appears as a named or linked source in AI-generated answers – is the right thing to track, because traffic alone no longer tells the full story. AI Overviews reduce clicks by 34.5 per cent compared with standard search results, so your brand can be the cited answer to a buyer’s question and generate no trackable visit at all.

ChatGPT gives you the clearest attribution, as it automatically labels the traffic it sends to your site, so you can see it as a separate source in your analytics. For Google, map Search Console impression data against your PR campaign timelines and look for patterns. For Claude and Perplexity, watch your branded search volume – how often people search directly for your company name – since it tracks closely with AI visibility.

Some tools now measure citation share directly. Manual testing across the major AI platforms with prompts relevant to your category is a practical starting point if you don’t have one yet.

How should ASEAN B2B brands use earned media to improve AI search visibility?

Start by treating PR as part of your AI search strategy. Find out which publications AI tools actually cite for your category and build relationships with those outlets. Pitch stories that include named experts, real and compelling data, and specific angles – not general company news. Stay consistent, because AI search rewards brands that show up regularly in their industry’s conversation.

Most in-house PR teams (if they even exist) are completely missing the boat on this.

ASEAN’s AI market is projected to grow from US$12 billion in 2025 to nearly US$80 billion by 2031, a 37 per cent annual growth rate. Boston Consulting Group has projected that AI and generative AI will contribute US$120 billion to the region’s GDP by 2027. The buyers and decision-makers in this market are already using AI search as a primary research tool.

Brands that build a consistent, deliberate earned media presence as a central part of their marketing mix are giving themselves a real advantage over those that treat PR as an afterthought.

—

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 How earned media drives AI search visibility in ASEAN appeared first on e27.

For more than a century, L’Oréal has built its reputation on science, creativity, and a commitment to understanding consumers. But the next chapter of beauty innovation is not being written in any single laboratory or boardroom. It is being built across a distributed ecosystem of startups, technologists, and domain specialists, co-creating what the industry will look like a decade from now.

L’Oréal’s evolution from in-house R&D powerhouse to beauty tech ecosystem orchestrator reflects a broader shift in how transformative innovation happens. Frontier breakthroughs increasingly emerge at the intersection of disciplines, where AI meets molecular chemistry, where material science meets circular design, where behavioral data meets brand strategy. No single organisation, however large or well-resourced, can lead on every front simultaneously. The answer is collaboration.

At the heart of this strategy is the convergence of four forces reshaping beauty: artificial intelligence, advances in materials and chemistry, a new standard for sustainability, and deeper, more continuous consumer insight. Together, these forces are not just improving existing products, they are redefining what it means to innovate in beauty. The thesis is simple but consequential: the future of beauty is being built through collaborative innovation.

Why startups are central to L’Oréal’s strategy

Startups offer something that established organisations struggle to manufacture internally: the freedom to move fast, experiment boldly, and develop deep specialisation in a single domain. Where large companies must balance priorities across global operations, a focused startup can spend years perfecting one technology, one process, one capability. That depth has become one of the most valuable currencies in modern innovation.

For L’Oréal, partnering with startups is not a hedge or a curiosity. It is a deliberate strategic mechanism. Startups bring frontier innovation across AI applications, materials science, consumer data infrastructure, and real-world experimentation. In return, they gain something equally valuable: the scale, validation, and real-world complexity that only a company of L’Oréal’s size and reach can provide. It is a genuinely mutual exchange. Startups are accelerated. L’Oréal gains agility and access to capabilities it could not build at the same speed or depth alone.

The platform for this collaboration is the L’Oréal Big Bang Beauty Tech Innovation Program. Focused on the SAPMENA region (South Asia Pacific, Middle East and North Africa), the program identifies startups solving real problems across the beauty value chain, from ingredient discovery to consumer engagement. Winners enter a year-long collaboration with L’Oréal, supported by partners including Accenture, Google, and Meta, with the opportunity to pilot, refine, and scale their technology within one of the world’s most complex and dynamic beauty markets.

The 2025 cohort produced four standout collaborations, each addressing a distinct dimension of the beauty innovation challenge, and together painting a picture of what the industry is becoming.

Halo AI: Scaling creator-brand collaboration through AI

The influencer marketing industry is growing at a speed that the tools managing it have struggled to match. Global spending on influencer marketing is projected to reach $40 billion in 2026, a 171% increase in a single year. Yet beneath this growth sits a structural inefficiency. Brands and creators have difficulty finding each other. Campaigns require extensive manual effort to source, vet, and manage. And while nano and micro-influencers drive some of the highest engagement in the industry, they remain the hardest to discover and activate at scale.

Halo AI, founded in 2024 and based in Saudi Arabia, uses advanced AI to intelligently match brands with nano and micro-influencers. It automates the discovery, vetting, and campaign management processes that can consume enormous amounts of time and resources. Its matching engine can review more than 100,000 creator profiles in minutes, identifying influencers whose audiences and values are most likely to align with a brand’s message.

Once campaigns are live, the platform continues working on both sides. Creators receive AI-guided support to understand campaign requirements, refine their content, and optimise their posting strategy. Brands access a real-time dashboard tracking collaboration progress and performance metrics, enabling continuous optimisation rather than a single point-in-time assessment.

“The main problem that we’re solving is that brands and micro-creators have a really hard time finding each other, and effectively and efficiently collaborating,” said Vito Strokov, CEO and co-founder of Halo AI.

Following its win at the 2025 Big Bang program, Halo AI is embarking on a commercial pilot with L’Oréal across the SAPMENA region. “When the biggest advertiser in the world decides they want to partner with you on a one-year-long pilot, it’s not only massive validation for the mission and the team, but proof we’re solving real problems for brands,” said Rami Saad, co-founder of Halo AI.

Without: Turning plastic waste into circular beauty packaging

Every year, approximately 855 billion plastic sachets are produced globally. Less than 1% are recycled. The reason is structural: multilayer plastics (the kind used in most beauty and personal care sachets) have long been considered impossible to process through conventional recycling systems. The result is a category of packaging that is used once and discarded, with virtually no path back into the supply chain.

Without, a material science startup, has developed a proprietary chemo-mechanical process that transforms unrecyclable multilayer plastics into high-quality materials that can be reused for packaging. In a world-first demonstration, the team created a 100% recycled shampoo bottle made entirely from discarded sachets, proving that circular design can meet both performance and aesthetic standards without compromise.

Also read: The next frontier for tech startups? The US$590B beauty industry

The same platform has produced sunglasses from discarded chip packets, demonstrating that the approach can scale across beauty, fashion, and lifestyle categories. What makes Without’s model distinctive is not just the technology, but the sourcing philosophy behind it. The company works with waste-pickers and marginalised workers, formalising and upskilling their roles into dignified, better-paid employment, building ethical, inclusive supply chains alongside the circular material flows.

“We have been working on this for the past five years,” said Anish Malpani, founder of Without. “Winning the Big Bang Beauty Tech Innovation Program gives us a lot of validation, and this is how we think we can help make supply chains more ethical and sustainable.”

Without is now working with L’Oréal on a pilot to test and scale 100% recycled packaging solutions aligned with the Group’s L’Oréal for the Future program and its 2030 sustainability targets. “This competition was not just about getting an award and recognition,” said Malpani. “We actually get the opportunity to do a pilot program that can be scaled across markets. That means L’Oréal is going to put their money where their mouth is.”

Sravathi AI: Accelerating sustainable chemistry and ingredient discovery

The challenge of sustainable innovation in beauty is not only about packaging or supply chains. It extends deep into the chemistry of the products themselves: the ingredients, formulations, and manufacturing processes that determine environmental impact long before a product reaches a consumer’s hands.

Traditional ingredient discovery is slow, expensive, and resource-intensive. Screening potential compounds typically requires iterative laboratory testing across a large candidate pool, consuming time, materials, and energy at every stage. Sravathi AI was built to transform this process by bringing AI into the heart of chemistry.

Founded in 2020 in Bangalore, India, the startup has developed a proprietary platform that combines generative AI, predictive models, and physics-based chemistry to accelerate discovery while reducing cost, carbon emissions, and toxic material use. The platform can screen thousands of potential compounds and narrow them to a few hundred high-potential candidates for synthesis, compressing timelines that previously took years into a fraction of the time.

“Our goals are shared. We’re all working toward sustainability to ensure a cleaner, smarter planet for generations to come,” said Parag Tipnis, VP Commercial of Sravathi AI.

The collaboration with L’Oréal focuses on a specific and practical challenge: improving how key active ingredients already used in L’Oréal formulations are produced, starting from bio-sourced raw materials. Using AI to redesign production pathways and continuous flow processes, Sravathi AI is exploring how these ingredients can be manufactured more sustainably and efficiently without compromising on performance.

“L’Oréal believes in doing innovation at cost, speed and scale, and they want to do it in a sustainable manner,” said Tipnis. That alignment runs through every dimension of the partnership, which spans discovery, chemistry, development, and continuous manufacturing, contributing directly to L’Oréal’s focus areas of climate transition, circularity, and conscious innovation.

The broader implication is significant. If AI can redesign the chemistry of production, making it cleaner, faster, and more precise, then sustainability becomes an embedded feature of the R&D process rather than a constraint applied after the fact. Sravathi AI’s work points toward a future where the molecular-level decisions made in a lab are guided by the same intelligence that optimises outcomes at scale.

Heatseeker: Bringing real-world experimentation into beauty innovation

There is a persistent gap in market research that has cost the consumer goods industry billions: the difference between what consumers say they will do and what they actually do. In surveys, people express preferences. In stores and on social platforms, they make choices. Those two things frequently diverge, and brands that build strategies on stated intent rather than authentic behaviour pay the price in failed launches and wasted investment.

Heatseeker was built to close that gap. The platform is an AI-powered customer insights engine that helps brands test ideas against real consumer behaviour (not stated preferences) before committing resources to a launch. Using multivariate ads across channels like Meta and LinkedIn, Heatseeker runs live market experiments that capture genuine signals: clicks, engagement, and behavioural indicators of authentic interest.

The platform fuses this experimental data with first-party sources (CRM data, performance metrics, prior research) and uses AI to automate experiment setup, competitor analysis, and insight generation. The result is quantitative, predictive intelligence grounded in what consumers actually do. Synthetic personas, built from behavioural data, enable teams to test scenarios and guide roadmaps before any investment is made.

“Gone are the days when brand teams in marketing or product would wait months or even weeks for insights that drive innovation. We’re now bringing that kind of insight to them in just seconds,” said Fiona Tricia, COO and co-founder of Heatseeker.

Also read: Where beauty innovation is headed, according to L’Oréal

As a winner of L’Oréal’s 2025 Big Bang program, Heatseeker is now working with L’Oréal to explore real-world applications of its technology, integrating behavioural intelligence into product development, messaging strategy, and go-to-market decisions across the beauty industry. The platform’s emphasis on authentic consumer behaviour aligns directly with L’Oréal’s Beauty Tech strategy, which uses data and AI to accelerate innovation and deliver more personalised experiences to consumers.

“Our vision, which is all about serving customers and understanding consumers, aligns so beautifully with how passionate L’Oréal is about its customers,” said Kate O’Keeffe, CEO and co-founder of Heatseeker. “Being recognised by L’Oréal, a brand that really is the best in the world at this, is a signal that our business is really on the right track.”

The collaboration positions both companies at the forefront of a shift in how consumer insight is generated and used, from a periodic research function to a continuous, real-time infrastructure that informs decisions across the entire innovation cycle.

What these collaborations reveal about the future of beauty

Taken individually, each of these four partnerships addresses a specific problem: influencer marketing inefficiency, unrecyclable packaging, slow ingredient discovery, unreliable consumer research. But taken together, they reveal something larger: a picture of what the beauty industry is becoming.

Beauty is increasingly predictive, data-driven, and adaptive. The decisions that shape products, campaigns, and supply chains are no longer made primarily on intuition or periodic research. They are informed by continuous signals: behavioural data from live market experiments, AI-generated insights from molecular screening, real-time campaign performance metrics. The infrastructure of beauty innovation is becoming intelligent.

Innovation is also expanding beyond products to platforms and systems. The startups in this cohort are not creating new lipstick formulations or fragrance profiles, they are building the underlying capabilities that make better products possible at scale. They are building the infrastructure through which the next generation of beauty innovation will flow.

Sustainability is shifting from aspiration to scalable infrastructure. Without’s circular materials technology, Sravathi AI’s cleaner chemistry processes, and L’Oréal’s 2030 commitments are not separate initiatives, they are converging toward a supply chain where sustainability is engineered in from the start, not appended at the end.

And AI is now embedded across the full value chain: at the discovery stage, where it accelerates ingredient identification; at the production stage, where it redesigns manufacturing pathways; at the marketing stage, where it matches brands with creators; and at the consumer insight stage, where it turns behaviour into predictive intelligence. The result is a beauty ecosystem where every layer of the value chain is becoming more efficient, more responsive, and more aligned with what consumers actually want.

Building beauty through ecosystems, and looking ahead to Big Bang 2026

L’Oréal’s approach to the 2025 Big Bang cohort is not an experiment. It is the expression of a strategic conviction: that the most important innovations in beauty will emerge from collaboration between large incumbents and specialised startups, each bringing capabilities the other cannot replicate alone.

As a platform for co-innovation across disciplines, L’Oréal is doing something more than finding interesting technology partners. It is building an ecosystem, a network of capabilities, relationships, and shared ambitions that compounds in value over time. The startups gain scale, credibility, and real-world complexity. L’Oréal gains agility, specialised expertise, and early access to the technologies that will define the next decade of the industry.

The future of beauty is interconnected, tech-enabled, and continuously evolving through collaboration. That future is not waiting to arrive, it is already being built, one partnership at a time, through programs exactly like this one.

And that mission continues. With Big Bang 2026 on the horizon, L’Oréal is once again opening the door to the startups, technologists, and innovators who believe that beauty’s next breakthroughs belong to those bold enough to build them together.

—

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

The e27 team produced this article sponsored by L’Oréal

We can share your story at e27 too! Engage the Southeast Asian tech ecosystem by bringing your story to the world. You can reach out to us here to get started.

Featured Image Credit: L’Oréal

The post Inside the beauty innovation ecosystem building L’Oréal’s next breakthroughs appeared first on e27.

Prelude

The global economic landscape is currently navigating a period of profound structural realignment, characterised by the aggressive internationalisation of Chinese small and medium-sized enterprises (SMEs). This movement, colloquially termed “ChuHai” (出海) or “Going to Sea,” has transitioned from a tactical response to domestic market saturation into a foundational strategic imperative for the survival and long-term viability of the Chinese private sector.

As we approach the late 2020s, the ChuHai phenomenon is no longer merely about exporting low-cost commodities; it represents a sophisticated evolution toward brand excellence, localised operational capacity, and the deployment of advanced technological ecosystems across emerging and developed markets alike.

This report provides an exhaustive analysis of the drivers, magnitude, and regional complexities of this trend, while identifying high-potential market opportunities for new ventures seeking to support this massive wave of globalisation.

The structural drivers of overseas expansion: The involution crisis

At the core of the current surge in Chinese SME internationalisation is the phenomenon of “involution” (内卷, nei juan). This term describes a state of hyper-competition where excessive effort and resources are expended for diminishing returns, often leading to a race-to-the-bottom in pricing and profit margins. The structural roots of this crisis are multi-dimensional, involving fiscal pressures, demographic shifts, and the collapse of the traditional growth engines that fueled China’s rise over the previous three decades.

The exhaustion of the real estate sector, which historically accounted for 20 per cent to 30 per cent of China’s GDP and 27 per cent of all bank loans, has created a massive vacuum in the domestic economy. The bursting of property bubbles has not only eroded household wealth and suppressed consumer demand but has also stripped local governments of their primary source of fiscal revenue: land financing.

Also Read: Trust takes years to build but one flawed system can damage a micro business overnight

By late 2025, local government debt had escalated to an estimated US$18.9 trillion, forcing these entities to pivot toward manufacturing and high-tech sectors as alternative drivers of GDP growth. This pivot has resulted in a deluge of government subsidies, tax incentives, and low-interest loans directed toward state-favoured industries, including electric vehicles (EVs), solar equipment, and semiconductors.

The unintended consequence of this state-led investment has been a chronic oversupply and massive overcapacity. When domestic demand failed to keep pace with the state-subsidised production surge, firms were forced into brutal price wars to survive. In the EV sector, for example, dominant players have used financial leverage to pursue predatory pricing strategies intended to eliminate smaller competitors.

By late 2025, industrial profits in several manufacturing segments saw year-on-year declines as sharp as 13.1 per cent, effectively erasing previous growth and creating a “growth without profits” trap. For many SMEs, the domestic market has become a zero-sum game, making international expansion the only viable pathway for maintaining operational solvency and achieving sustainable margins.

Market size and profiling: The 2026-2028 surge

The magnitude of the Chinese SME overseas surge is reflected in the record-breaking metrics of outward direct investment (ODI) and the volume of private enterprises engaging in global trade. By the end of 2025, the number of private Chinese enterprises with actual import and export activity reached approximately 613,000, accounting for the vast majority of the country’s 700,000 active trade entities.

Also Read: The divided AI race nobody wins: How businesses can navigate the US-China tech divide

For the 2026-2028 cycle, the “ChuHai” market is expected to expand by approximately 175,000 SMEs annually. This “New Wave” is characterised by a transition from “Made in China” (volume export) to “Operated by China” (localised presence).

SME internationalisation profile (2026-2028)

Strategic expansion priorities

Chinese SMEs are no longer pursuing simple volume; their expansion is now “capability-led,” focusing on the following strategic pillars:

• Market expansion: Escaping domestic deflation and price wars to capture margins that are often double what is achievable domestically.
• Technology licensing and IP: Shifting toward licensing proprietary technology to local partners to overcome regulatory barriers and secure data exclusivity in sensitive sectors like biomedicine and AI.
• Global R&D and talent: Establishing overseas innovation centres to access bilingual leadership and local technical talent, bridging the cultural gap between HQ and the market.
• Manufacturing outsourcing and nearshoring: Relocating production capacity to regions like Mexico (nearshoring) or Vietnam to bypass US and EU tariffs and shorten delivery cycles from weeks to days.

The role of the accelerator state: Policy support and institutional frameworks

The internationalisation of Chinese SMEs is a core component of the national industrial strategy. The government has evolved into an “accelerator state,” moving toward a multi-layered system designed to fast-track the growth of high-tech SMEs in strategic sectors.

The Little Giants initiative

The “Little Giants” program focuses on “specialised, refined, special, and new” SMEs within key industrial chains. For the 2024-2026 period, the program prioritises the “six foundations”: core basic parts, core basic components, key software, advanced basic processes, key basic materials, and industrial technology foundations.

Capital support is significant, with guidelines aiming to inject up to CNY 6 million (approximately US$830,000) per firm over a three-year cycle. By late 2025, the program had cultivated over 13,000 national-level Little Giants, with cities like Shenzhen housing over 1,000 such firms.

Also Read: The scale layer nobody budgeted for: How AI agents unlock growth for Asian businesses

The 15th five-year plan and the 2030 horizon

The strategic roadmap for the next phase (2026-2030) outlines a shift from growth driven by scale to growth driven by quality. Key objectives include:

• Technological self-reliance: Accelerating breakthroughs in brain-computer interfaces, quantum technology, and semiconductor supply chains.
• Digital economy expansion: Increasing the share of core digital industries to 12.5 per cent of overall GDP.
• Support for global scale: Explicitly encouraging internet platforms, AI companies, and professional services to expand and form partnerships overseas.

Geographic realignment: Emerging corridors and the global South

As regulatory scrutiny intensifies in the US and EU, Chinese SMEs are diversifying toward regions with lower regulatory friction.

ASEAN: The hub-and-spoke model

ASEAN is the critical region for restructuring Chinese supply chains, with FDI inflows reaching US$226 billion in 2024. For Chinese SMEs, ASEAN offers a mobile-first consumer base aligned with Chinese digital strengths. Vietnam, Malaysia, Indonesia, and Thailand have become central hubs for manufacturing and customer service.

The Middle East: The Gulf blue ocean

The Middle East—particularly the GCC states—is a priority destination for Chinese capital. In 2024, the region received US$39 billion in BRI investments, a 102 per cent increase YoY. Saudi Arabia alone drew US$19 billion. Chinese firms view the Gulf as a “blue ocean” due to high policy flexibility and security.

Latin America: Mexico and the nearshoring shift

In Latin America, the focus is shifting toward Mexico as a gateway to the North American market. Chinese brands now account for 57 per cent of cars imported into Mexico as of early 2025. The “Trump Corollary” to the Monroe Doctrine creates headwinds, but the USMCA framework provides a significant duty-free advantage for Chinese SMEs that can successfully localise manufacturing in the region.

Also Read: AI agents and the new rules of business execution

The technological architecture of ChuHai: AI and agentic trade

The year 2025 has been identified as China’s “AI Agent Year,” marking the deployment of autonomous systems to manage global operations.

Agentic workflows in cross-border operations

Next-generation AI agents are being integrated into platforms like WeChat (OpenClaw framework) to solve operational challenges. SMEs use “AI Agent Orchestration” to automate end-to-end marketing, content generation, and performance loops. In logistics, “Control Tower Agents” optimise delivery routes and reorder workflows, reducing routine task handling by 60-80 per cent.

The service ecosystem opportunity: Identifying business niches

The Chinese SME expansion has outpaced its supporting service ecosystem, creating massive gaps in talent, compliance, and localisation.

The talent gap: A staggering bottleneck

The number one bottleneck for Chinese companies going global is talent acquisition. In 2025, there was a talent gap of 4 million people in the cross-border e-commerce sector alone.

• Startup opportunity: AI-powered executive search for “bridge leaders” and Employer of Record (EOR) services to manage regional talent networks.

Compliance and regulatory readiness (C-as-a-service)

Compliance requirements for outbound firms surged by 250 per cent in 2025.

• Startup opportunity: Global payroll, HR compliance SaaS, and Data Sovereignty solutions to manage the web of local tax and privacy laws.

Localisation and ecosystem integration

Success is tied to the ability to “go in”—truly entering the local culture—rather than just “going out”.

• Startup opportunity: Cross-border traffic marketing, AI-native SEO (GEO), and market entry incubators for the Global South.

Also Read: The scale layer nobody budgeted for: How AI agents unlock growth for Asian businesses

Strategic conclusions and recommendations

The expansion of Chinese SMEs overseas is a structural trend that will define global commerce through 2030. Driven by the exhaustion of domestic profits and supported by a multi-billion-dollar state accelerator, this wave is moving toward higher-value sectors and deeper regional integration.

For new ventures, the most promising path is to become a “strategic enabler” for this outbound surge. The transition from “Made in China” to “Brands from China” represents the next great shift in the global economy. Those who can provide the cultural, regulatory, and technological bridges will be positioned at the heart of the world’s most dynamic trade corridor. The path forward lies in combining AI-native “intelligence” with the “empathy” required for deep cultural localisation.

—

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 ChuHai: The business opportunity nobody in Southeast Asia is talking about appeared first on e27.

This is about freight. Specifically, why freight, logistics, and supply chain, an industry so gloriously unglamorous that people fall asleep mid-sentence just describing it, is turning into one of the most interesting places to build a company in Asia right now.

I know. Bear with me.

Few people want to work in logistics — that’s the point

When I tell people I run a logistics company, one of two things happens. Either their eyes glaze over, or they say something polite and immediately change the subject.

Logistics has an image problem. It’s not the industry you dream about at university. It doesn’t attract the same talent pipelines, the same VC attention, or the same media coverage as the sexier corners of tech. Nobody is writing breathless Substack posts about customs clearance and freight.

And yet, quietly, something is happening.

The global 4PL market is projected to grow at 8.1 per cent CAGR from 2025 to 2032, according to research commissioned by Wayfindr citing Market. In e-commerce specifically, that growth rate accelerates to 12 per cent CAGR over the same period, driven by surging cross-border trade, supply chain complexity, and the relentless expansion of direct-to-consumer brands into new markets.

E-commerce is eating retail across every market. Manufacturing is rapidly diversifying across Southeast Asia as brands shift supply chains out of China. And the technology layer that connects all of this, the visibility, the coordination, the data, has barely been built.

Also Read: Cybersecurity strategies for startups on a budget

The problem is not shipping, it is complexity

Here is what most people get wrong about logistics as a startup opportunity. They think the opportunity is in moving things faster. A better courier. A smarter warehouse. A cheaper freight rate.

That is not the problem.

The real problem is that scaling an e-commerce brand across multiple countries, with manufacturing in Vietnam or China, selling into the US, UK, and Europe simultaneously, involves dozens of moving parts, dozens of providers, and nobody whose job it is to be accountable for all of it at once.

A brand owner running a US$20 million e-commerce business should be thinking about product, marketing, and growth. Instead, they spend half their week chasing updates from a freight forwarder here, arguing with a warehouse over there about a stock discrepancy, and trying to figure out why their landed costs keep changing.

That is not a shipping problem. That is an orchestration problem. And orchestration is exactly where tech has enormous room to run.

Vietnam is not a trend — it is a structural shift

I spend a lot of time in Southeast Asia. A big segment of our team operates out of Vietnam, and what we see on the ground is not a passing wave. It is a genuine realignment of global manufacturing.

Brands that were 100 per cent China-reliant five years ago are now actively splitting production. Vietnam, Indonesia, Taiwan, and Thailand are absorbing that shift under the China+ strategy. And with it comes a whole new layer of complexity: new suppliers, new compliance requirements, new last-mile challenges, and new currency risk.

Also Read: AI is removing the co-founder bottleneck for early-stage startups

The market data reflects this reality. APAC is the fastest-growing region for 4PL adoption globally, projected at 10.8 per cent CAGR through 2032, growing from a US$19.7 billion market today to US$44.7 billion by 2032. That growth is being driven directly by the manufacturing shift, the rise of cross-border e-commerce, and the urgent need for smarter supply chain infrastructure across the region.

For the brands navigating this, the operational burden has never been higher. For the companies building technology and services to help them do it, the opportunity has never been larger.

The startup opportunity in Southeast Asia’s logistics sector is not just local. It is the infrastructure layer for global commerce.

The boring industries have the best defensibility

Here is something I learned coming from the oil and gas world before logistics: the industries that look boring from the outside are often the ones with the deepest moats.

An operator who genuinely understands how freight consolidation works out of Guangzhou, how Vietnamese customs changes every year, and how to design a landed-cost model that holds across six destination markets, that knowledge does not transfer easily. The complexity is the barrier. And the complexity, at the moment, is only increasing.

Tariff changes. Carbon reporting requirements. Cross-border regulatory divergence. Every one of these adds another layer that brands need help navigating, and another reason to build toward a model where one intelligent, tech-enabled partner is accountable for all of it.

This is what the fourth-party logistics model, 4PL, exists to do. And it is a model that is still, genuinely, in its infancy in Asia.

What is a 4PL?

A fourth-party logistics provider, or 4PL, is an independent, non-asset-owning partner that designs, manages, and optimises your entire supply chain, coordinating freight forwarders, warehouses, carriers, and technology under one roof. Think of it less like a supplier and more like a control tower: one point of accountability for everything that moves.

Also Read: Startups driving AI automation, fintech, and accessibility gather at Echelon Singapore 2026

The unsexy bet is often the right one

I started building in logistics because I saw a problem I could not stop thinking about. Not because it was fashionable. Not because investors were excited. Frankly, most of them were not.

We bootstrapped to eight figures without a cent of external funding, which I mention not to brag but to make a point: the fundamentals of the problem were strong enough that we did not need anyone to believe in the vision before the market proved it. The demand was real. The inefficiency was real. The gap was real.

The next decade of e-commerce growth in Asia is going to be built on infrastructure. Not just digital infrastructure, but the physical and operational infrastructure that moves real products from real factories to real customers. The companies that build the intelligence layer on top of that, the platforms, the visibility tools, the coordination systems, those are the companies that will be quietly, unglamorously, extraordinarily valuable.

So yes. Freight. Supply chain. Logistics.

I promise it’s more interesting than it sounds.

—

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 Why the most boring industry in the world is quietly becoming a startup goldmine appeared first on e27.

Most discussions about cyber insurance in industrial sectors start from the wrong assumption. They treat insurance as a recovery tool that will somehow make a severe OT incident manageable after the fact. That is too comforting and too shallow. OT environments are not ordinary digital estates. Many security guides stress that these systems carry unique performance, reliability, and safety requirements, and that logic executing in OT has a direct effect on the physical world, including potential harm to people, the environment, equipment, and production.

That is why cyber insurance will not save OT in the way some boards hope it might. Any basic guide to cyber insurance describes cover mainly in terms of losses tied to IT systems and networks, along with incident management support. Put plainly, a policy may help pay for response, legal support, forensics, and parts of business interruption. It does not restore process integrity, rebuild operational judgement, or make a compromised plant safe to trust again.

OT is exactly where the limits show up

The limits of insurance become sharper in industrial settings because the real cost of failure is often operational, not merely financial. Unexpected outages in industrial processes are unacceptable, that outages often need to be planned days or weeks in advance, and that high availability requires exhaustive pre deployment testing.OT components often remain in service for 10 to 15 years, sometimes longer, and that change management is more demanding because software and firmware updates can require careful assessment and revalidation.

The insurance market itself has recognised that OT is not yet a fully mature underwriting domain. There is still a comparative lack of understanding and awareness of cyber physical risk, even as the potential for threats to bridge IT and OT is becoming more apparent. It means buyers should not assume the policy market has already solved how to price or absorb the full reality of industrial cyber exposure.

Where does insurance actually matter

It matters as an incentive mechanism.

Cyber insurance should not be viewed as a substitute for strong internal defences, but rather as a means to encourage better risk management practices. Insurance can support cyber risk management by improving quantification, providing access to expertise and crisis services, and encouraging risk reduction through premium pricing. This is the strategist’s lens that matters more. Insurance is most valuable when it changes organisational behaviour before the incident, not when it simply finances some of the damage afterwards.

Also Read: Fighting misinformation and cyberbullying against women in public sphere: Call for gender equality and online safety

That behavioural effect is already visible in underwriting logic. Coalition’s published guidance says insurers typically look for controls such as multi-factor authentication, training, tested backups, identity access management, and data classification before agreeing coverage, and that stronger controls can help firms secure more favourable rates. The market is large enough to influence buyer behaviour, and selective enough to shape which controls become non-negotiable.

The underwriting conversation should be different

The problem is that too many cyber insurance conversations still start with general IT hygiene and stop there. For industrial operators, that is not enough. The more serious opportunity is to use underwriting as a forcing function for a narrower set of OT relevant controls that genuinely reduce consequence.

A complete and accurate asset inventory is critical for managing OT risk, and that inventory data should include vendors, model numbers, firmware, operating systems, and software versions so vulnerabilities can be identified and tracked. It is also explicit that network segmentation and isolation help enforce security policies and control access to sensitive components, and that remote access should be provided only when justified, limited to business need, and supported by stronger safeguards. Tested backups are described as critical to recovery, with verification for reliability and integrity where technically possible. These are not theoretical controls. They are the foundations of whether an industrial site can contain, understand, and recover from a cyber event.

This is where insurance can become useful as a behavioural lever. If insurers and brokers start asking tougher OT questions around definitive asset inventory, segmented network zones, controlled vendor access, restoration testing, and evidence of recovery readiness, they will do more than screen risk. They will change internal priorities. Teams that struggle to win budget for resilience work often find that the conversation changes once underwriting, renewal, deductibles, or coverage conditions enter the room. That is not because insurance is replacing the engineering discipline. It is because insurance creates a commercial consequence for postponing it.

The market can also influence procurement

One of the most underused levers in OT security is procurement pressure. That is where cyber insurance could become more strategically useful over the next few years.

Operators should prioritise products and manufacturers that follow secure by design principles, and highlight issues such as logging, authentication, data protection, secure defaults, and established vulnerability management processes. That matters because insurers cannot underwrite away poor product design, but they can make weak procurement choices more visible and more expensive.

Also Read: Thailand’s cybersecurity boom has a weak core

A strategist should see the implications immediately. If policy terms, engineering standards, and procurement expectations all start pointing in the same direction, the market begins to reward firms that buy more defensible systems in the first place. That is far more valuable than arguing about claims after a major event. It shifts the conversation from “will this be covered” to “should we be accepting this exposure at all”.

What measurable risk reduction is

The weakness in many cyber insurance discussions is that they stop at broad hygiene language. Boards are told to improve resilience, but not how to tell whether risk is genuinely moving. 

In practice, a measurable reduction in OT should look less like policy paperwork and more like observable proof. Can the operator show a current inventory of critical OT assets and software versions? Can it demonstrate that high consequence zones are segmented and that permitted flows are understood? Can it prove that remote access is limited, approved, and capable of being disconnected quickly? Can it show that backups, images, and configuration states are actually restorable? Those are the sorts of measures that shorten recovery, reduce uncertainty, and make underwriting more meaningful. 

The strategist’s conclusion

Cyber insurance will not rescue OT from poor architecture, weak product choices, or years of deferred resilience work. The market itself has acknowledged limits around systemic events and around understanding cyber-physical exposure. But that does not make insurance irrelevant. It makes its real value clearer.

Its best role is to alter incentives.

It can force boards to treat OT risk as financially visible. It can force security teams to translate technical gaps into underwriting consequences. It can force operations leaders to evidence controls that otherwise remain assumed rather than proven. It can force procurement teams to take secure-by-design claims more seriously. Used that way, insurance becomes less a comfort blanket and more a discipline mechanism.

—

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 Cyber insurance won’t save OT, but it can change behaviour appeared first on e27.

I know what you’re thinking (pun intended). But this is not a sensational fantasy about mind reading. It is an extrapolation from real advances in AI, brain-computer interfaces (BCIs), and neuroscience. As systems for decoding neural signals and translating thought-related activity into digital output continue to improve, the question is no longer just what they can do on their own, but whether they could one day be networked, and to what extent. Although BCIs are no longer a novel concept, if such devices could communicate directly with one another, they might give rise to technological telepathy.

From private thought to usable input

Thoughts have historically remained private by default, becoming shareable only when forced through speech, writing, gesture, or code, each of which introduces delay, tension, and translation between intention and expression. Technological telepathy becomes consequential not because machines can literally read minds in a science-fiction sense, but because computing is progressively collapsing that gap, as BCIs, silent-speech systems, neural decoding models, and generative AI converge into a communications stack in which cognition itself becomes a usable input, suggesting that future networks may connect minds rather than merely devices or identities.

A longer lineage of brain-machine translation

This trajectory does not originate with ATR, AlterEgo, Neuralink, or the current “AI summer,” but extends through a longer history of rendering the brain legible to machines, beginning with Hans Berger’s EEG and its demonstration of non-invasive neural capture, continuing through José Delgado’s stimulation experiments, Alvin Lucier’s ”Music for Solo Performer” and its use of EEG for artistic control, Eberhard Fetz’s work on learned modulation of neural firing, and Jacques Vidal’s articulation of “brain-computer communication,” later made publicly tangible through BrainGate’s cursor control for paralysed patients and Kevin Warwick’s experiments in technological telepathy, all of which situate this internet of minds as a continuation rather than a rupture.

Institutional drivers and military interest

This history cuts across neuroscience, engineering, military funding, performance art, and public spectacle, with DARPA embedded as part of the field’s institutional structure, particularly through programs such as N3 that target high-performance neural interfaces for human-machine teaming, active cyber defence systems, and control of unmanned aerial vehicles, although no credible public evidence supports operational BCI-to-BCI communication in military or covert use.

Also Read: AI is irrevocably changing the tech landscape, and you are going to need a new map

Science fiction as conceptual groundwork

Science fiction anticipated the conceptual and social implications well before technical feasibility, as seen in Alfred Bester’s “The Demolished Man“ and its treatment of telepathy and social order, William Gibson’s “Neuromancer“ and “Johnny Mnemonic“ and their linking of neural systems to networked computation, Iain M. Banks’s neural lace, Ramez Naam’s infrastructural treatment of networked cognition, and Isaac Asimov’s “Foundation,“ alongside concepts such as hive mind theory and consciousness field theory, which framed expectations even as technological telepathy itself remains grounded in engineering rather than speculative human abilities.

What technological telepathy actually is

In practical terms, technological telepathy does not involve full extraction of continuous private thought, but instead consists of narrower capabilities such as decoding constrained visual categories from brain activity during sleep, reconstructing partial features of perceived or imagined images, or inferring silently articulated words from neuromuscular signals in the face and jaw, which are distinct but collectively indicate that communication technologies are moving upstream toward earlier stages of cognition.

In practical terms, current and near-term systems rely on specific device classes rather than abstract “mind reading,” including implanted electrode arrays that record neural firing directly from the cortex, non-invasive headsets based on EEG or functional imaging that capture aggregate brain activity, endovascular interfaces that access signals via blood vessels, and wearable EMG sensors placed on the face or jaw to detect subvocal speech, all of which produce partial, task-specific signals that must be decoded and interpreted through software, meaning that what is transmitted is not raw thought but a constrained, device-mediated representation of selected aspects of cognition.

The layered communication stack

A clearer understanding emerges when treated as a layered system, beginning with capture through electrodes, imaging systems, or wearable sensors that detect neural or neuromuscular activity, followed by decoding via machine-learning models that map signals to probable words, intentions, percepts, or categories, then mediation through software that filters noise, ranks interpretations, predicts continuations, corrects errors, and structures ambiguous biological signals into coherent output, and finally transmission to devices, other individuals, or networks, with AI functioning as the intermediary that translates between biological activity and digital meaning rather than enabling direct thought transfer.

From data to the “internet of minds”

Within this architecture, cognition-related data becomes processable in ways analogous to other network data while remaining qualitatively closer to thought itself, introducing the possibility that privacy breaches occur prior to completed expression, and implying that the dominant model will resemble cognitively mediated client-server communication rather than direct peer-to-peer telepathy.

Also Read: How fintech in Asia is enabling and making education affordable for everyone

Philosophical constraints on “pure thought”

Philosophical objections of Neuralink-like research, such as Slavoj Žižek’s 2020 talk at the University of Winnipeg, emphasise that conceptual thought does not exist independently of language, challenging the notion that “pure thought” can be transmitted without distortion and reframing linguistic imperfection as intrinsic to meaning rather than an obstacle to be eliminated.

Technical limits and partial decoding

Technical constraints remain substantial, particularly in the form of invasiveness, as high-performance BCIs often depend on implants placed in or near the brain, introducing surgical risk, long-term maintenance, and questions of removal and bodily autonomy, while less invasive approaches such as wearables or endovascular interfaces shift these tradeoffs without removing them, as illustrated by ATR and Yukiyasu Kamitani’s lab, whose dream decoding studies demonstrated category-level prediction of dream content under tightly constrained conditions rather than full reconstruction, thereby establishing partial permeability of internally generated experience without generalisability.

Silent speech and the boundary of intent

Alternative approaches, such as AlterEgo, focus on silent speech, relying on intentional subvocalisation and neuromuscular detection to create a clearer boundary between private thought and transmitted output, although current limitations in surface EMG prevent reliable decoding of inner-speech phonetic content, reinforcing that existing systems detect controlled signals rather than unrestricted cognition.

The fragility of intentionality boundaries

This boundary of intent, while conceptually important, remains technically and institutionally fragile, as systems may expand the definition of “intended” signals through software updates, model retraining, or error correction, and as movement toward decoding imagery, semantic content, and prelinguistic intention further complicates distinctions between thinking, rehearsing, and transmitting, with proposed safeguards such as learned cognitive protocols or mental “keys” likely to erode under pressures for efficiency and usability.

From research frontier to platform economy

The transition from research to commercialisation, evident in companies such as Neuralink, Synchron, and Paradromics, reframes neurotechnology as infrastructure rather than experiment, introducing business models that range from high-cost clinical hardware reimbursed through healthcare systems to platform-based software and institutional deployment in workplaces, defence settings, or insurer-managed care, and elevating neural data as a potentially valuable resource due to its proximity to intention before action, preference before declaration, and friction before expression.

Also Read: Beyond Silicon Valley dreams: Why Southeast Asia is rewriting the rules of tech for good

Expression as a hybrid artifact

In this context, expression becomes a joint product, as systems infer, correct, rank, and autocomplete outputs derived from incomplete signals, producing hybrid artefacts that blur the boundary between user intention and machine contribution, thereby shifting the problem from privacy alone to questions of authorship and authenticity, since output may already reflect negotiation between human cognition and computational prediction.

Why consent is insufficient

Consent, traditionally understood as sufficient ethical grounding, becomes inadequate when users do not stand outside the systems shaping their expression, requiring structural governance mechanisms such as purpose limitation, auditable mediation, rights of refusal, prohibitions on employer coercion, protected clinical boundaries, and legal remedies when machine output is misattributed to the user as fully self-authored.

Feedback loops and cognitive adaptation

Feedback dynamics further complicate the system, as users adapt to decoder behaviour, anticipate system completions, and potentially reshape their own cognitive patterns to improve legibility and control, creating a feedback loop in which thought is partially oriented toward machine interpretability and generating ambiguity in responsibility when outputs reflect inferred rather than explicitly intended meaning.

Assistive promise and differential cognitive citizenship

While assistive applications for paralysis, severe motor impairment, or speech loss remain one of the strongest justifications, variability in neural and neuromuscular signals introduces differential cognitive citizenship, in which some individuals are more easily legible to systems than others due to anatomy, fatigue, stress, medication, injury, learning history, or neurotype, producing structured inequalities in performance, correction burden, and access.

Regulation, geography, and legibility inequality

These inequalities intersect with broader regulatory and geopolitical conditions, as jurisdictions that prioritise speed, scale, or strategic advantage may normalise less reliable systems more quickly, while Chile’s neurorights turn and the OECD’s neurotechnology governance work represent efforts to establish rights-based and standards-based constraints before large-scale commercialisation hardens, highlighting that cognitive interfaces will be shaped by states, healthcare systems, defence institutions, and major technology firms with divergent regulatory approaches.

Ownership, control, and contested infrastructure

Cognitive infrastructure is therefore unlikely to be uniformly owned, with centralisation more likely in hardware, clinical deployment, and large-scale inference systems, while mediation layers, user-facing software, and potentially open models remain more contestable, shifting the central political question from ownership of discrete thought-data to governance of the channel through which thought becomes public, legible, and actionable.

Also Read: The foundation of Southeast Asia’s tech future

Conditions for legitimate development

Legitimate development would require constrained deployment, local processing by default where feasible, strict separation between therapeutic and productivity uses, independent auditing of intent-detection and mediation systems, meaningful user oversight and contestability, and enforceable rights to refuse cognitive monitoring without loss of work, care, insurance, or civic participation, recognising that technological telepathy simultaneously compresses the distance between thought and communication while inserting additional computational mediation between them.

Near-term reality: low-bandwidth cognition

In near-term scenarios, the most achievable outputs remain limited to affective state, emotional valence, stress, calm, urgency, attentional load, and simple assent or refusal rather than full semantic language, although even these low-bandwidth signals may carry operational value in domains such as military coordination, justice, or entertainment.

Conclusion: authorship under mediation

The progression from networks connecting machines to those connecting identities suggests that an internet of minds would connect cognition itself to computation at unprecedented proximity, leaving unresolved the central question of whether thought, once mediated, inferred, and transmitted through such systems, can remain meaningfully one’s own.

—

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 Technological telepathy: Is an ”internet of minds” possible? appeared first on e27.

The global humanoid robotics industry is fragmenting into two distinct ecosystems pursuing fundamentally different scaling strategies: China’s deployment-led approach prioritising rapid manufacturing scale and real-world learning, versus North America and Europe’s AI-first methodology betting that foundation models and vision-language systems will determine long-term competitive advantage.

This strategic bifurcation carries profound implications for technology trajectories, supply chain configurations, and ultimately, which regions capture value as the market matures.

Also Read: The humanoid robot economy is no longer science fiction

According to “Humanoid robots 2026” by Roland Berger, these contrasting approaches reflect different resource endowments, institutional capabilities, and strategic philosophies about how complex technologies scale. Neither path guarantees success; each offers distinct advantages and risks. Still, the divergence increasingly shapes ecosystem development, reducing cross-regional interoperability and creating parallel technology stacks unlikely to converge.

The scale differential is striking: China’s estimated 15,000 units produced in 2025 exceed North America’s output by a factor of 30 and dwarf EMEA’s production by more than 150 times. Yet North American companies command nearly equivalent total funding (US$3.8 billion versus US$4.1 billion), reflecting higher capital intensity per company and a greater emphasis on software development, which requires substantial AI infrastructure investment rather than manufacturing capacity.

China’s manufacturing flywheel: scale drives data, data improves AI, AI enables deployment

China’s strategic approach prioritises getting robots into real-world environments quickly, accepting initially limited capabilities in exchange for operational data and manufacturing experience. This deployment-first methodology draws on the nation’s established strengths in hardware manufacturing, rapid iteration cycles, and vertically integrated supply chains that can absorb early-stage demand volatility.

The 39 identified Chinese startup OEMs documented by Roland Berger pursue targeted applications in entertainment, logistics, and basic manufacturing — environments with structured workflows, repetitive tasks, and controlled conditions where current AI capabilities prove sufficient. Rather than waiting for human-level general intelligence, Chinese developers optimise for specific contexts, accumulating deployment experience and operational data whilst building manufacturing infrastructure.

This approach constructs a powerful flywheel: manufacturing scale reduces unit costs, making robots accessible to more deployment environments; deployments generate operational data that improve AI capabilities; improved AI enables robots to handle more complex tasks, expanding the addressable market; market expansion drives additional manufacturing scale. If this flywheel accelerates successfully, China could establish compounding advantages that are difficult for rivals to overcome, despite superior foundational AI research capabilities concentrated in Western institutions.

The industrial policy dimension reinforces private sector initiatives. China’s “Robot+” strategy, articulated in the 14th Five-Year Plan for Robotics Industry Development, establishes explicit targets for humanoid robot development with governmental support spanning R&D funding, pilot deployment programmes, and procurement preferences. Provincial and municipal governments offer additional incentives (subsidies, tax benefits, and land allocations), creating supportive ecosystem conditions for rapid scaling.

Supply chain integration provides additional advantages. China’s electronics and mechanical manufacturing ecosystems supply components for consumer electronics, automotive, and industrial automation globally. This established base enables humanoid developers to source actuators, sensors, structural components, and compute modules domestically with shorter lead times and tighter integration than developers dependent on cross-border supply chains.

Western AI-first strategy: software advantages create defensible moats

North American and European ecosystems pursue fundamentally different competitive positioning, treating humanoid robotics as an AI problem requiring cutting-edge machine learning capabilities rather than primarily a manufacturing challenge. This software-first approach bets that long-term competitive advantage will emerge from foundation models, vision-language systems, and proprietary training datasets, enabling robust autonomy in unstructured environments, capabilities that manufacturing scale alone cannot replicate.

Also Read: The real battle in humanoid robotics is about data, not hardware

The capital intensity reflects this philosophy. North American companies typically allocate more funding per startup than their Chinese counterparts, consistent with their need for substantial computational resources, AI talent, and extended R&D timescales. Leading Western humanoid developers increasingly position themselves as AI companies that happen to build robots, rather than robotics companies incorporating AI, a subtle but significant strategic distinction.

Western developers emphasise generalisation, creating robots capable of learning new tasks with minimal task-specific programming, over optimisation for predefined workflows. This ambition requires more sophisticated AI architectures, larger training datasets, and longer development timescales before initial deployment. The approach reflects confidence that superior AI capabilities will ultimately overcome China’s manufacturing scale advantages once Western robots demonstrate human-comparable adaptability.

Academic and corporate AI research ecosystems in North America and Europe provide a competitive advantage in foundational capabilities. Universities and research institutions in these regions publish disproportionately in top-tier AI conferences and journals; technology companies operate cutting-edge AI infrastructure; and talent concentrations in hubs like the San Francisco Bay Area, Seattle, Boston, London, and Zurich create network effects that accelerate innovation. These advantages are particularly important for frontier AI development, which requires deep expertise and significant computational resources.

Strategic divergence: How two paths will shape the future of humanoid robotics

The emerging split in the global humanoid robotics industry — a deployment-led, manufacturing-first path in China versus an AI-first, research-driven trajectory in North America and Europe — is more than a strategic curiosity. It is the formation of two distinct ecosystems that will shape how capabilities evolve, where value is captured, and how quickly robots become an ordinary part of economic life.

Each path plays to regional strengths and carries different risk–reward profiles. China’s scale-first model accelerates real-world learning, drives down unit costs, and can produce rapid market adoption in structured applications. The Western AI-centric approach aims for generality and long-term defensibility through advanced models and software expertise, accepting slower initial deployment in exchange for potentially larger payoffs if foundational AI breakthroughs deliver human-comparable adaptation.

Practical implications to watch:

• Supply chains and standards will bifurcate, making interoperability and component sourcing more complex.
• Market segmentation will deepen: high-volume, task-specific deployments versus lower-volume, highly capable generalists.
• Policy and industrial policy will matter: procurement, subsidies, and regulation can amplify regional advantages.
• Investment patterns will reflect these dynamics: capital flows into manufacturing scale in China and compute- and talent-intensive R&D in the West.

Ultimately, the market’s outcome won’t be a simple winner-takes-all. Instead, expect parallel value chains to coexist and compete: one optimised for cost-effective, immediate utility; the other for general-purpose intelligence and adaptability. The most consequential question for industry leaders and policymakers is not which approach is intrinsically superior today, but which ecosystem can convert its early advantages into durable, compounding strengths, through data, standards, talent, and access to markets.

Also Read: Why robotic hands could make or break the humanoid industry

Whichever path proves more successful, the near-term fragmentation will shape product design, regulation, and commercial strategy for years to come. That fragmentation is not merely a technological divergence; it is the unfolding of a geopolitical and industrial contest whose outcomes will determine how and by whom robots are woven into the fabric of everyday life.

===

The post China builds robot armies while the West chases robot brains appeared first on e27.