The venture capital world is awash with AI hype. Every fund claims to back the next frontier. Few can point to companies that have crossed from demo to dollars in under a year. Antler, the global early-stage VC with a growing Asia footprint, is making that claim and backing it with numbers.

Jussi Salovaara, co-founder and Managing Partner for Asia, sat down to defend the firm’s thesis on agentic AI, its “One Asia” platform spanning Korea, Japan, and Southeast Asia, and its controversial bet on China-outbound founders. He also confronts the hard questions: enterprise trust, deeptech timelines, talent wars with Samsung and Hyundai, and what happens to these startups if the AI spending bubble pops.

Also Read: Why Antler is going all-in on Japan’s earliest-stage founders

The answers are sharper and more candid than most VCs offer.

Edited excerpts:

You’re describing a shift from AI copilots to autonomous systems. But most enterprise buyers are still struggling to trust AI with basic decisions. Aren’t you getting ahead of reality?

The question assumes AI autonomy is binary. It isn’t. Think of your best manager training a new employee. With the right guidance, that employee can make basic decisions and handle well-defined responsibilities. AI is at a similar stage. Most modern models already have the logical reasoning needed for many business tasks. The real challenge is designing the right context, guardrails, and scope.

That’s exactly what we look for at Antler. We’re not backing companies claiming artificial general intelligence. We’re backing founders who identify a narrowly defined problem, codify domain expertise into AI systems, and enable reliable decisions within a carefully crafted scope.

The results speak for themselves. IndustrialMind.ai, founded by three ex-Tesla Gigafactory executives, built AI that replaces up to 80 per cent of repetitive engineering work. AppSecAI automatically writes, validates, and delivers security patches in 30 minutes at one-hundredth of the cost of manual processes. CONPA secured six-digit contracted revenue within three months of launch. These are commercial outcomes, not experiments.

What exactly counts as “meaningful commercial traction”? Is that a paying customer, a signed pilot, or something else?

Meaningful traction means contracted revenue, live ARR, or a very large qualified pipeline with documented ROI. We do not count free pilots or letters of intent.

To give specific examples: ChainShift secured six-figure contracted revenue within 10 months. i10x reached seven-digit annualised revenue in eight months. This pace is significantly faster than historical benchmarks for early-stage software, which often took 18 to 24 months to reach similar milestones.

Korea, Japan, and Southeast Asia have very different startup cultures and enterprise buyer behaviours. How does Antler actually operate as a unified “One Asia” platform in practice?

The starting points are genuinely different. Japan and Korea offer unmatched industrial depth, robotics expertise, and corporate R&D budgets. Southeast Asia offers a massive, mobile-first digital economy and an agile scale-up environment. Chinese founders bring frontier AI research talent and an execution intensity forged in the world’s most competitive technology market. These are not interchangeable, and we do not pretend they are.

What the Antler platform provides is a common outcome opportunity: building a global company. The friction appears in localisation, regulatory compliance, and enterprise sales cycles. We mitigate that with experienced, on-the-ground partners across our 27 global locations.

Global VCs like a16z, Sequoia, and Lightspeed are all doubling down on agentic AI. What does Antler genuinely offer an AI founder in Asia that they can’t get from a brand-name fund?

Several of those funds have backed companies we first invested in at inception; they operate at a different stage and serve a different need. What we bring beyond capital is a network of local partners with boots on the ground across 27 locations, embedded in the ecosystems where founders are expanding.

The most concrete expression of this is our Embark programme, a four-week immersion that bridges our strongest Asian portfolio companies into Silicon Valley, connecting them with US enterprise customers, investors, and operators. Twelve startups across Asia have gone through three Embark cohorts. Every single one has secured US traction. We build the infrastructure and systematic support to get founders to the stage where global funds are ready to write the next cheque.

In a press release, you mentioned backing “China-outbound entrepreneurship.” Given geopolitical tensions and scrutiny in Western markets, how do you assess those risks?

China has spent two decades producing some of the world’s most technically rigorous engineers and AI researchers. A growing number of those founders are choosing to build for global markets from day one. That combination of frontier technical training and genuine global ambition is rare, and it is concentrated in this cohort right now.

Also Read: Analysis: SEA’s June funding spike masks a narrow recovery in VC funding

The question we assess at the investment stage is simple: where is your customer, where is your data, and where is your team? If the answers point towards global ambition from inception, the geopolitical risk profile is fundamentally different from a company that started in China and is now trying to expand outward.

Several portfolio companies are in sectors with notoriously long commercialisation timelines. How does Antler’s inception-stage model align with deeptech?

Deeptech companies with long timelines are precisely where early conviction creates the most asymmetric returns. We help founders compress the timeline from lab to first enterprise deployment, then hand them off to the right capital partners to carry the journey forward.

At inception, we look for technical validation, strong IP protection, and the first commercial signal — a paid pilot, a joint development agreement, or a signed letter of intent. Korea’s conglomerates and Japan’s industrial corporates are among the most sophisticated early adopters of deeptech in Asia, and we work closely with those networks to connect our founders with the right enterprise partners.

What happens to these companies if the enterprise AI spending correction some analysts are warning about actually materialises?

A spending correction would actually accelerate the path for the companies we back. A correction is, by definition, a correction in spending on broad horizontal platforms, experimental tooling, and marginal productivity gains. When budgets tighten, enterprise buyers do not cut tools that are reducing their costs or generating their revenue.

Our founders create business value through genuine domain expertise, not generalist AI. Verixus Labs CEO Joel Kosmin holds an Oxford PhD in Molecular Genetics, has over a decade of research experience, and worked at AstraZeneca before building an AI-powered operating system for biomanufacturing. His platform delivers 61 per cent higher mammalian stem cell yields and 66 per cent fewer experiments compared to standard approaches. That is not a product that gets cut when AI budgets tighten. A correction would validate it.

AI talent in Asia is fiercely competed for by Samsung, Hyundai, and SoftBank-backed companies. How are early-stage founders competing for engineers without matching corporate salaries?

Early-stage founders compete on ownership, autonomy, and the chance to build category-defining technology from scratch. The best engineers are often frustrated by bureaucracy and slow deployment cycles inside large conglomerates.

Also Read: Antler invests US$5.6M across 14 AI startups with early commercial traction

The founders in our portfolio are the very talent those conglomerates want to hire. IndustrialMind.ai was founded by executives who led Tesla’s manufacturing AI transformation. Infron was founded by ex-Alibaba AI researchers who left one of the most well-resourced AI environments in the world. They didn’t leave because they couldn’t get corporate salaries. They left for equity, creative control, and the chance to define a category. That’s the story they tell every engineer they recruit, and it’s credible precisely because they made the same choice themselves.

The post Why Antler is backing Chinese founders building away from China appeared first on e27.

Qapita is an ESOP platform for startups through to listed companies, helping founders unlock the Power of Ownership for their stakeholders. Qapita’s focus aligns with a growing global trend: as startups stay private for longer, the complexity of managing cap tables, liquidity events, and investor reporting has created a surge in demand for various ESOP management tools. Powering over 2,400 clients globally, Qapita offers cap table, ESOP advisory, liquidity programmes, as well as valuation and financial reporting services tailored to meet the needs of both shareholders and employees.

Managing an ESOP within a Singapore-incorporated private company comes with a structural limitation that direct share issuance and traditional trusts don’t fully solve. Singapore limits private companies (Pte Ltd) to 50 shareholders, presenting a unique challenge for founders to manage this statutory restriction.

For a firm in Singapore, allowing employees (ex-employees and advisors) to exercise their options early may lead to additional admin, including but not limited to potentially crossing this 50-shareholder private company threshold sooner than expected. With many founders considering incorporating an entity or a holding company in Singapore, this signals a need for alternatives in share delivery solutions across the Southeast Asia region.

To address this, Qapita has recently launched ESOP SPV, a first-of-its-kind share delivery solution built specifically for Singapore-incorporated entities. This could be particularly useful for founders who have yet to set up their ESOP plan and want to incorporate an SPV from the start to ensure a clean cap table before future team expansion and fundraises.

Also Read: From perk to power: Rethinking ESOPs in the modern talent economy

A Special Purpose Vehicle (SPV) acts as an alternative share delivery method that consolidates shareholder names in a single entity. When employees exercise, they become shareholders of the SPV instead of the company directly, encouraging tangible employee ownership in a flexible yet compliant manner.

Here’s how Qapita’s ESOP SPV works

A Singapore Private Company (Pte Ltd) is set up to hold shares. Employees hold shares in the SPV proportionate to their allocation. Only the SPV appears on the cap table. Here are some of the key features of Qapita’s new and improved product:

• Clean cap table from day one: A single SPV entry is cleaner for investors than a list of employee names. Employees stay consolidated in a single SPV entity, which simplifies due diligence, cap table documentation, and future fundraising rounds.
• Flexibility on employee share exercises: Employees can exercise more regularly without the company crossing the 50-shareholder limit. This lets them act when it’s most tax-efficient — rather than waiting for a liquidity event. As employees become shareholders of the SPV instead of the company directly, encouraging share exercises can allow them to feel a sense of ownership.
• Perfect middle ground between direct share issuance and trusts: ESOP trusts require a licensed trustee, ongoing fees, and greater regulatory overhead. For a startup, an SPV delivers the same structural benefit at a fraction of the cost.

To sum up, ESOP SPVs are best suited for early-to-growth-stage Singapore-incorporated startups with up to 50 ESOP participants. As the SPV allows employees to exercise their options and participate as shareholders through a structured vehicle, this reduces administrative hassle, maintains a clean, investor-ready cap table, and results in potential tax-saving opportunities for employees.

Ultimately, the right ESOP structure depends on your goals, your team size, and how you want employees to engage with their equity. Qapita can help you figure out what works, including implementation, structural, and taxation considerations for your employees.

Discover how an ESOP SPV compounds future benefits, giving employees real ownership while keeping your cap table investor-ready, at a fraction of the cost of a trust. Whether you’re setting up a new ESOP plan or already have an existing programme, Qapita’s advisory team can help you evaluate whether an SPV is the right structure for your startup and set it up end-to-end.

Learn more here: https://www.qapita.com/sg/companies/equity-compensation-advisory/spv

—

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

The e27 team produced this article in partnership with Qapita.

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: Qapita

The post Qapita launches ESOP SPV for Singapore-incorporated entities appeared first on e27.

Singapore-based travel rewards startup HeyMax has launched HeyMax First, a membership product that gives users access to miles before they have earned them, in a bid to reshape how frequent travellers think about redemption.

The product allows members to draw down up to one million Max Miles upfront and use them for flight or hotel redemptions through partner loyalty programmes. Members then earn the miles back over time through spending on the HeyMax app. The company said first-year membership fees will be waived for users who sign up during the launch period.

Also Read: What travel tech can look like for the travel industry’s revival

HeyMax said members must pay a reclaimable access fee to unlock the upfront miles. That fee is returned as users earn back the miles through future spending. The company said there is no deadline, penalty, or minimum activity requirement for members to complete the earn-back process.

The proposition is straightforward: instead of spending for years to accumulate enough points for a premium redemption, users can take the trip first and rebuild their balance later. The harder question is whether enough users will change their behaviour, and whether the model can be sustained without becoming a liability-heavy rewards scheme.

Reversing the loyalty sequence

Traditional airline loyalty programmes rely on a simple sequence: spend, earn, redeem. That structure gives airlines, banks and merchants years to manage liability, expiry, breakage and devaluation. HeyMax First reverses the order by moving redemption to the front of the customer journey.

“For so many years, loyalty programmes have asked travellers to do the same thing: spend first, wait years, and hope your miles are still worth something when you finally have enough,” said Joe Lu, CEO and co-founder of HeyMax. “HeyMax First reverses that. We front you the miles, you take the trip you’ve been putting off, and you earn them back on your own schedule.”

The product targets a clear consumer frustration. Premium award flights often require large mileage balances, and casual travellers may struggle to accumulate enough points before programmes change redemption rates or impose new restrictions. In Southeast Asia, where cross-border travel is frequent but incomes and credit card penetration vary widely by market, the ability to access miles earlier could appeal to younger professionals and aspirational leisure travellers.

HeyMax says its miles transfer on a one-to-one basis to more than 20 airline and hotel loyalty programmes, giving users access to over 70 airlines across major global alliances. The company did not disclose the full commercial terms behind HeyMax First, including how it prices the access fee, manages redemption risk, or accounts for miles advanced to members.

A crowded rewards battlefield

HeyMax was founded in 2023 by four former Meta engineers. The company raised US$11 million in Series A funding in January 2026, led by Peak XV Partners, and has since expanded beyond Singapore into Hong Kong. It plans to enter Japan, Taiwan and Australia by the end of 2026.

The startup operates in a market that sits at the intersection of travel, fintech, commerce and loyalty. In Southeast Asia, rewards have become a customer acquisition tool for banks, e-wallets, superapps, airlines and cashback platforms. GrabRewards, ShopBack, Kris+, AirAsia MOVE and Cathay’s Asia Miles all compete in adjacent ways for consumer attention and transaction volume.

Also Read: HeyMax acquires Hong Kong’s krip to supercharge Asia loyalty rewards expansion

Globally, companies such as Bilt Rewards in the US have shown that non-traditional spending categories can be converted into travel rewards at scale. Points-search and redemption platforms such as Point.me and AwardWallet have also built businesses around the complexity of airline loyalty. HeyMax is taking a different route: it is not merely helping users optimise existing points, but advancing future rewards against expected spending.

That distinction is crucial. Loyalty programmes are balance-sheet businesses as much as marketing tools. Miles have real cost, and redemption-heavy users can be expensive if they do not generate sufficient follow-on activity. HeyMax First will likely depend on three things: a broad merchant network, repeat spending behaviour, and careful control of who receives upfront miles and how much.

The company says users can earn Max Miles from more than 800 merchants globally. That merchant base gives HeyMax a starting point, but the model’s durability will depend on whether members concentrate more of their everyday spending inside the app after taking an upfront redemption.

Why Southeast Asia is a relevant testbed

Southeast Asia is a logical market for this type of experiment. The region’s digital economy has grown rapidly, with Google, Temasek and Bain estimating gross merchandise value at US$263 billion in 2024. Online travel has also rebounded sharply since the pandemic, with consumers increasingly comfortable booking flights, hotels and experiences through digital platforms.

At the same time, the region remains fragmented. Loyalty behaviour differs across Singapore, Indonesia, Thailand, Vietnam, Malaysia and the Philippines. Payment methods vary, airline networks are uneven, and regulatory approaches to consumer credit, stored value and rewards liabilities are not uniform. A rewards product that looks simple to the user may require careful structuring behind the scenes.

Singapore gives HeyMax a useful launch market. It has high card penetration, heavy outbound travel demand, and consumers who are familiar with airline miles and bank reward points. But regional expansion will not be automatic. In larger Southeast Asian markets, the company would face stronger localisation demands, lower average spending power, and competition from entrenched wallets and superapps.

The product also arrives at a time when airlines and banks are becoming more protective of loyalty economics. Frequent flyer programmes have become valuable assets, and carriers routinely adjust redemption charts, fuel surcharges and partner availability. If HeyMax positions itself as a flexible layer across multiple programmes, it may benefit from consumer frustration with single-airline schemes. But it will also remain exposed to changes imposed by those same partners.

The test ahead

HeyMax First is an ambitious attempt to repackage loyalty around immediacy rather than delayed gratification. The company is betting that access to premium travel today will motivate users to route future spending through its platform tomorrow.

That may resonate with travellers who dislike the uncertainty of waiting years to redeem points. It may also appeal to consumers who see travel as a priority but do not have enough miles or credit card spend to reach business-class thresholds quickly.

Also Read: HeyMax hits US$6M revenue milestone, eyes Asia Pacific expansion

Still, the product will need to prove that its earn-later structure is not just attractive at launch, but economically repeatable. Waiving first-year membership fees should reduce friction, but the key metric will be post-redemption engagement: whether members continue spending after they have taken the trip.

For now, HeyMax has put a sharp twist on a familiar category. In a region where travel demand is rising, rewards are becoming more competitive, and consumers are increasingly willing to try fintech-led alternatives, the company has chosen a high-risk, high-attention way to stand out.

The post HeyMax First offers upfront miles, but the economics will face a real-world test appeared first on e27.

The post The first mover myth: Why being first rarely means winning appeared first on e27.

Every AI conversation seems to begin with the same question: what can AI do better than humans?

It is an understandable question since AI can now analyse information quickly, summarise long reports, generate first drafts, support customer service, and automate tasks that used to take hours. For many companies, the appeal is immediate. If a tool can help a team work faster, reduce repetitive work, and make better use of existing resources, it is difficult to ignore.

But I think there is another question we should be asking more often, especially in Southeast Asia: Who actually gets to benefit from this shift?

The current AI conversation often assumes that everyone starts from the same place. It assumes that workers have time to experiment with tools, businesses have budgets for training, and communities have equal access to digital infrastructure. In reality, the gap between those who are ready for AI and those who are not is still very visible.

This is where the discussion about human value becomes more interesting. The issue is not simply whether AI will replace certain tasks. It is whether we are building an AI economy where more people can meaningfully participate.

Human value is changing, but it is not disappearing

Much of the anxiety surrounding AI comes from the belief that machines are replacing human value. I understand where that concern comes from, but I do not think it tells the full story.

For a long time, many professional skills were built around access to information. People were valued for how quickly they could research, organise knowledge, analyse trends, or turn information into a useful output. Those skills still matter, but AI has changed the baseline. A first draft, a summary, or a basic analysis is no longer as difficult to produce as it once was. That does not mean human value has disappeared. It means the source of value is moving.

In an AI-enabled workplace, the people who stand out are often not the ones who can simply produce the most output. They are the ones who can ask better questions, understand context, make sound judgments, and connect technology to real human needs.

Also Read: The accordion effect: How AI follows the rhythm of expansion and compression

AI can generate a list of ideas, but it cannot always know which idea is right for a specific market, community, or moment. It can analyse patterns, but it does not carry the lived experience needed to understand why people behave the way they do. It can help optimise a process, but humans still need to decide what kind of outcome is worth optimising for.

This is why I do not see the future of work as a simple story of humans versus machines. It is more likely to become a story of who can use machines with enough judgment, empathy, and responsibility.

The real divide is access, not interest

In Southeast Asia, interest in AI is not the problem. Many people and businesses are curious about it. However, the harder question is whether they have the same opportunity to learn, test, and apply it.

The World Economic Forum’s Future of Jobs 2025 coverage on Southeast Asia notes that digital skills are becoming more important for companies across the region, but many employers still see significant gaps. Upskilling and reskilling are becoming priorities because the pace of change is already affecting what businesses need from their teams.

This matches what many of us are seeing on the ground. Larger companies can invest in AI tools, internal training, consultants, and structured experimentation. Smaller companies often have to make do with limited time, limited budget, and limited guidance.

For workers, the difference can be just as stark. Someone in a major city with strong internet access, an English-language education, and exposure to global tools may find it easier to learn AI. While a frontline worker, informal worker, or small business owner in a less connected area may not have the same starting point.

The risk is that AI becomes another layer of advantage for people and organisations that already have access to capital, infrastructure, and education.

Southeast Asia needs inclusive AI growth, not just faster AI adoption

The region’s digital economy is still growing quickly. The e-Conomy SEA 2025 report says Southeast Asia’s digital economy has grown from US$40 billion in GMV a decade ago to more than US$300 billion in 2025.

Indonesia is a useful example of why inclusion matters in this conversation. MDI Ventures’ recent white paper, Catalysing Digital Resilience and Sustainable Growth: Advancing Inclusive Innovation and AI-Driven Impact Across Indonesia’s Digital Economy, notes that the country has around 65 million MSMEs, contributing 60.5 per cent to GDP and absorbing 96.5 per cent of the national workforce. It also points out that Indonesia’s digital economy is projected to reach between US$180 billion and US$340 billion by 2030, while many small businesses still face challenges in financing access, digital infrastructure, cybersecurity, and AI readiness.

Also Read: Singapore, AI, and the rise of emotional outsourcing

That context matters because Indonesia’s digital economy cannot be considered truly strong if its smaller businesses are left behind. Growth may happen at the top, but resilience depends on whether the broader business ecosystem can participate.

This is where AI should be seen as more than a productivity tool. If applied well, it can support better credit scoring, improve access to digital financial services, strengthen cybersecurity, and help small businesses operate with more confidence. But these benefits will not spread automatically. They need infrastructure, trust, relevant products, and patient ecosystem-building.

The MDI white paper makes this point indirectly through its focus on impact capital, digital trust, AI, cybersecurity, and inclusive digital infrastructure. Its portfolio examples, including Amartha, Qoala, Privy, and CYFIRMA, show how technology can support access, protection, identity, and trust within the wider digital economy.

We should also think about how people learn

There is another part of this shift that deserves more attention. As companies automate more entry-level tasks, we may accidentally weaken the pathways that help people build experience.

Many junior roles are built on tasks that are not glamorous but are deeply educational. Writing meeting notes, preparing research, drafting reports, checking details, and supporting senior colleagues are often how people learn how an industry works. These tasks teach judgment slowly. They expose people to context, mistakes, client expectations, and decision-making.

If AI takes over too much of that early work without a replacement learning path, companies may solve one efficiency problem while creating a future talent problem.

This is why the talent conversation should not stop at whether people know how to use AI tools. The deeper question is how quickly people can keep learning as the nature of work changes. LinkedIn estimates that 70 per cent of the skills used in most jobs will change by 2030, while PwC’s 2025 Global AI Jobs Barometer found that workers with AI skills command a 56 per cent wage premium. This suggests that AI is not simply reducing the value of human talent. It is raising the value of people who can keep adapting.

For organisations, the risk is that workers who already have access to training, tools, and experimentation time will move further ahead, while those without that access fall behind. This does not mean companies should avoid automation. It means they need to be more intentional about learning.

If AI handles the first draft, junior employees still need to learn how to evaluate that draft. If AI summarises research, people still need to learn how to question the source, spot missing context, and decide what matters. If AI supports execution, teams still need to teach accountability, communication, and ethical judgment.

AI can speed up work, but it should not remove the process through which people become thoughtful professionals.

Great talent now looks different

This also changes what we should look for in talent. A few years ago, the strongest candidate might have been the person with the most polished technical skills or the most impressive credentials. Those things still have value, but they are no longer enough on their own.

Also Read: AI slop is a strategy problem, not a content problem

In an AI-enabled environment, I would pay closer attention to curiosity, adaptability, clarity of thinking, and the ability to work with ambiguity. I would also look for people who know how to use AI without outsourcing their judgment to it.

That last part matters. There is a difference between someone who uses AI to think better and someone who uses AI to avoid thinking. The first person becomes more capable. The second person becomes more dependent.

This is why AI literacy should not be treated as a narrow technical skill. It is becoming part of how people communicate, analyse, make decisions, and build trust. The strongest professionals will be those who can combine technological fluency with human understanding.

The future of AI should be measured by who gets included

Many businesses are asking how AI can help them do more with fewer people. That is a practical question, and it will not disappear.

But I hope more leaders also ask a broader question: how can AI help more people contribute?

That question leads to a different set of priorities. It pushes organisations to invest in training beyond senior teams. It encourages businesses to think about frontline workers, small merchants, regional entrepreneurs, and communities that may not be first in line for new technology.

Southeast Asia’s future growth will depend not only on how quickly AI is adopted, but on how widely its benefits are shared. If smaller businesses, young workers, and underserved communities are left behind, the digital economy may become more advanced without becoming more resilient. That would be a loss for everyone.

In the end, the most important human contribution in an AI-powered world may not be competing with machines. It may be making sure the future we build with them still works for more humans.

—

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

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27.

Join us on WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected.

The post Human value in the AI era is not what most people think appeared first on e27.