Ten years into building, the annual planning cycle stops being useful. A year is too short a unit to learn anything from. It’s long enough to feel like progress and short enough to hide the fact that the money isn’t made evenly across time. It shows up in bursts, during the stretches when conditions favour you, and it gets defended during the stretches when they don’t.

Most operators know this and rarely say it plainly, because it sounds like admitting the good years were luck. They weren’t luck. They were timing, and timing is a decision made years before the moment it pays off.

You can’t cause a cycle, but you can force your position in it

Andrew Carnegie’s steel business went through two price collapses, in the 1870s and again in the 1890s. Competitors did the obvious thing both times: cut production, lay off crews, wait for demand to return. Carnegie ran his mills at a loss instead, because construction costs were cheap and rivals were selling assets at fire-sale prices to survive. His instruction to a subordinate was “small profits and large sales” while everyone else retrenched. He wasn’t predicting the recovery. He was buying it in advance, at a discount, while the rest of the industry was too scared to spend.

Toyota’s position going into 1973 is the cleaner example of catching a shift rather than buying one. American demand ran on large-displacement engines, and Japanese compacts held a rounding error of US market share. The 1973 oil embargo changed the math on fuel cost overnight, and Japanese import share in the US moved from roughly 9 percent in 1976 to 21 percent by 1980. Toyota didn’t cause the oil shock. It had spent the prior decade building a fuel-efficient car for a market that didn’t want one yet, so when the market changed, the product was already on the lot.

Also Read: 15 Thai AI companies betting on products, not hype

Neither company controlled the macro event. Both controlled whether they were structurally ready the moment it hit, in cash, in capacity, in product that already existed rather than product still waiting to be built. That’s the actual answer to whether a tailwind can be forced: the conditions can’t be forced, but the position relative to them can.

Built to survive the gap between tailwinds

Corning is the case for doing this more than once. It’s a 170-year-old glass manufacturer that ran on Pyrex and CorningWare through most of the twentieth century, invented low-loss optical fibre in 1970 decades before the internet needed it, rode the fiber boom of the late 1990s, absorbed the 2001 fibre bust without gutting its glass science team, sold off Pyrex in 1998 to focus entirely on advanced glass, and turned a shelved forty-year-old formula into Gorilla Glass in 2007 after a call from Steve Jobs. Corning didn’t get one cycle right. It built a company that could survive the years between cycles, so it was still standing when the next one arrived.

Samsung’s memory chip business runs the same logic on a shorter clock. In 2008, when the financial crisis hit and every other DRAM maker cut capital spending to preserve cash, Samsung increased it, and repeated the move in the 2012 and 2019 downturns. Competitors treated the downturns as something to survive. Samsung treated them as the one window where capacity was cheap and competitors were retreating, which is a structurally different decision, and it’s a large part of why Samsung still leads the category.

Also Read: Your product is not your startup

None of these four were guessing about macro timing. Each decided, years ahead of the shift, what kind of company it wanted to be caught being when conditions turned, then built the balance sheet, product, or manufacturing base to match before the turn happened.

Conclusion

The planning horizon matters more than the plan itself. A company that budgets in single years will always be reacting to the season it’s already in. A company that plans in multi-year cycles can spend the quiet years on the unglamorous work: building capacity, buying distressed assets, shipping a product nobody’s asking for yet, so it isn’t scrambling to catch up when conditions turn favourable.

The harder question isn’t whether the next favorable window is coming. It always is. It’s whether the quiet years get spent building something ready to catch it, or just something that survived long enough to still be there when it shows up.

—

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 You can’t force a tailwind, you can force your readiness for one appeared first on e27.

The next phase of digital payments may not be defined by faster checkouts or cheaper transfers, but by a more uncomfortable question: who or what is being trusted to move money?

As businesses begin experimenting with AI agents that can search for suppliers, compare prices, negotiate terms, initiate payments, and reconcile invoices, the old assumptions around financial control start to fray. A human no longer clicks every button. A finance team may not manually approve every step. In some cases, software will act on behalf of a company, within rules set in advance.

Also Read: The next AI payments boom may happen in the back office

That shift sits at the centre of “Beyond Automation: Defining Agentic Global Payments”, a report by Sunrate and Mastercard. Its central argument is straightforward: automation alone is not enough. If AI agents are to handle commercial decisions involving millions of US dollars, companies need more than speed. They need accountability.

The report calls this missing infrastructure the “Trust Layer”, a framework that allows businesses to verify an agent’s identity, understand its authority, and trace what it has done. In other words, the future of payments will not just depend on whether AI can act intelligently. It will depend on whether organisations can prove that these actions were authorised, limited, and auditable.

From Know Your Customer to Know Your Agent

For the past decade, much of fintech has been shaped by Know Your Customer (KYC) rules. Banks, payment companies, and fintech startups have built systems to verify that users are who they say they are, screen them for risk, and monitor suspicious activity.

Agentic commerce introduces a new layer of complexity. If an AI agent places an order, books travel, pays a supplier, or moves funds across borders, the payment ecosystem needs to know more than the identity of the company behind it. It must also understand the identity and authority of the agent itself.

This is where “Know Your Agent”, or KYA, comes in.

KYA is not simply a branding exercise. It points to a practical set of controls: verifying an AI agent, defining what it is allowed to do, recording the intent behind a transaction, and ensuring that actions remain within commercial and policy boundaries. An agent authorised to buy office supplies, for example, should not be able to approve a large foreign exchange transfer. A procurement agent with a US$10,000 spending limit should not be able to split payments to bypass that limit.

Also Read: From chatbots to payment agents: AI’s next role in SEA commerce

For Southeast Asia, where many companies already operate across fragmented markets, currencies, payment methods, and compliance regimes, this matters. A regional startup may have suppliers in Vietnam, customers in Indonesia, finance operations in Singapore, and banking relationships across several jurisdictions. Adding autonomous agents into that mix without governance could create a risk environment that is difficult to monitor.

The three pillars of the Trust Layer

The report breaks the Trust Layer into three broad pillars.

The first is credential protection. In today’s payment systems, tokenisation is already used to replace sensitive card or account details with secure digital tokens. In an agentic payments environment, this becomes even more important. AI agents should not be passing around raw card numbers, bank credentials, or account information. If those agents are compromised, the damage could be significant.

The second pillar is intent capture. This means securely transmitting the user’s budget, preferences, constraints, and instructions along with the transaction. In human terms, it is the difference between saying “buy the cheapest ticket” and “buy a refundable economy ticket under US$700, departing after 7pm, with no overnight layover”. For businesses, intent capture allows systems to determine whether an agent acted in line with approved instructions.

The third pillar is KYA and governance. This is the architecture that verifies the agent’s identity and sets strict permission boundaries. It includes authentication, policy enforcement, audit trails, and the ability to revoke or modify permissions when needed.

These controls may sound technical, but their commercial importance is simple. Businesses cannot delegate financial decisions to agents if they cannot later explain what happened, why it happened, and whether it was allowed.

Why many AI projects do not make it past pilots

The urgency is not theoretical. According to Gartner, at least 50 per cent of AI projects were abandoned last year after the proof-of-concept stage. The reasons included poor data readiness, high costs, and a lack of risk control.

That last point is particularly relevant for payments. In many companies, AI pilots are still treated as productivity experiments. Teams test whether a model can draft emails, summarise documents, or automate customer support. Payments are different. A bad recommendation may waste time. A bad transaction may move real money, breach regulations, or damage a company’s relationship with banks and suppliers.

Also Read: The scarcity mindset is killing creativity, not AI

For Southeast Asian startups, this creates both a warning and an opening. The warning is that building a clever agent is not enough. A product that can automate procurement or treasury workflows may impress in a demo, but enterprise customers will ask harder questions before deploying it in live payment flows.

Who approved this transaction? What data did the agent use? Can the company prove that the payment matched its internal policy? Can a bank or payment service provider trace the chain of authorisation? What happens if the agent is tricked by fraudulent instructions?

The opportunity lies in answering those questions better than competitors. The strongest companies in this space may not be the ones with the most sophisticated AI interface, but those that combine automation with controls that banks, CFOs, auditors, and regulators can trust.

Why the ecosystem matters

A Trust Layer cannot be built by a single startup in isolation. Agentic payments will require coordination across banks, card networks, payment service providers, enterprise software platforms, and regulators.

This is where established networks such as Mastercard are likely to play a significant role. Card networks already sit across large parts of the payment ecosystem and have experience with tokenisation, identity standards, fraud management, and dispute processes. Extending governed, traceable tokenisation into autonomous payment flows is a logical next step.

Payment service providers and cross-border platforms also matter, particularly in Southeast Asia. The region’s businesses often deal with multi-currency payments, varied settlement timelines, and uneven levels of banking infrastructure. If AI agents are to operate across borders, they will need infrastructure that can translate business intent into compliant payment execution across different markets.

Regulators will also have to catch up. Many existing rules assume a human actor at key decision points. Agentic systems challenge that assumption. Over time, authorities may need clearer standards on agent identity, liability, consent, auditability, and operational resilience.

Trust as a competitive advantage

The rise of AI agents in payments is often framed as a story about efficiency. There is truth in that. Agents could reduce manual work, speed up reconciliation, and help businesses optimise costs across suppliers and currencies.

But efficiency will not be the deciding factor if companies fear losing control.

Also Read: The next AI payments boom may happen in the back office

The more important race is to build systems where autonomy does not mean opacity. Businesses will need to know not only that an agent completed a task, but that it did so within defined limits. Banks will need confidence that transactions are legitimate. Payment networks will need ways to trace credentials and intent. Regulators will need evidence that responsibility has not disappeared into a black box.

For Southeast Asia’s startup ecosystem, the message is clear. The next wave of payments innovation will not be won by speed alone. It will be won by companies that can make AI agents accountable.

The future belongs not just to agents that are smart enough to act, but to systems that are safe enough to trust.

The post From KYC to KYA: how AI agents are reshaping payment risk appeared first on e27.

The post The Kospi enters a bull market while crypto consolidates: Where did the risk appetite go appeared first on e27.

Singapore’s biotech sector has long had the ingredients of a serious life sciences hub: strong universities, public research funding, hospital networks, and a government keen to pull high-value industries into the city-state. What it has had less of is a steady stream of venture-backed drug developers built to compete internationally from day one.

K2 Therapeutics is trying to fit that gap.

The Singapore-based biotechnology company has raised US$50 million in seed financing from MPM BioImpact, the US investment firm that founded the company in 2024. The round has already helped K2 expand its pipeline to eight therapeutic programmes across multiple modalities, including antibody-drug conjugates and T-cell engagers, with assets ranging from pre-clinical candidates to clinical-stage programmes.

Also Read: The 27 SEA biotech firms betting on cells, fermentation, and code

That is a wide starting point for a newly formed biotech company. In drug development, “pre-clinical” typically means a therapy is still being tested in the lab or in animal studies, while “clinical” means it has entered human trials. The jump between those stages is where many young biotechs struggle, as costs rise sharply and scientific promise meets regulatory, safety and manufacturing realities.

K2’s model is to identify drug candidates internationally and advance them through in-house development. In practice, that means the company is not beginning solely as a discovery lab. It is looking globally for promising assets, then using capital and development expertise to move them through the difficult middle stretch of biotech: from candidate selection to human proof-of-concept, and potentially towards commercial partnerships or approvals.

A platform built around global asset sourcing

K2 Therapeutics said it expects to grow its portfolio further through asset acquisition, capital deployment and development. The approach reflects a broader shift in biotech financing, where investors increasingly back teams that can source overlooked or underdeveloped science globally, rather than relying on a single internal platform.

This matters in Southeast Asia because the region’s biotech ecosystem is still young compared with those in the US, Europe, China, South Korea and Japan. Singapore has strong research capabilities and hosts major pharmaceutical manufacturing and regional headquarters operations, but building venture-scale therapeutic companies remains difficult. Drug development takes years, requires specialised talent, and depends on access to sophisticated clinical, regulatory and manufacturing infrastructure.

A US$50 million seed round gives K2 unusually deep early backing by regional standards. Seed rounds in software can be used to build a product and test the market. In biotech, that money is often spent on experiments, toxicology studies, manufacturing preparation, regulatory filings and early clinical work before any revenue is in sight. The scale of K2’s financing signals that MPM BioImpact is not treating the company as a small exploratory bet, but as a vehicle to assemble and advance a serious therapeutic pipeline.

MPM BioImpact manages more than US$3.5 billion in assets and has a long history of forming and financing biotechnology companies. Its decision to found K2 in Singapore is also notable at a time when global life sciences investors are looking beyond the traditional Boston-San Francisco axis for scientific talent, clinical access and new deal flow.

A CEO with commercial experience

Alongside the financing, K2 has appointed Ying Huang as CEO. Huang was previously chief executive and a board member of Legend Biotech, where he oversaw the development and commercialisation of cell therapies and the company’s expansion to more than 3,000 employees. Before Legend, he was head of biotechnology equity research at Bank of America Merrill Lynch.

That mix of operating and capital markets experience is important for a company like K2. Biotech CEOs are not only expected to understand the science; they must also raise large amounts of capital, prioritise programmes, manage clinical risk, negotiate with pharmaceutical partners, and explain complex data to investors and regulators.

Also Read: Singapore’s Biobot Surgical raises US$15.6M to take prostate-care robot global

“By combining global asset sourcing with experienced development leadership,” K2 can rapidly advance differentiated therapeutic candidates with the potential to address significant unmet medical needs, Huang said.

The quote is measured, but it captures the thesis. K2 is betting that the bottleneck in biotech is not only invention. It is also execution: knowing which assets deserve capital, which should be stopped early, and how to move the strongest candidates through a highly regulated system.

Why ADCs and T-cell engagers are attracting attention

Among K2’s notable programmes are antibody-drug conjugates, or ADCs, and T-cell engagers. Both areas have drawn intense investor and pharmaceutical interest globally.

ADCs are often described as targeted cancer therapies. They combine an antibody, which seeks out specific markers on diseased cells, with a toxic payload designed to kill those cells more precisely than traditional chemotherapy. The field has seen several major acquisitions and licensing deals in recent years as drugmakers race to build oncology pipelines.

T-cell engagers work differently. They are designed to bring immune cells, particularly T cells, into close contact with cancer cells so the immune system can attack them. The idea is powerful, though developing safe and effective T-cell engager therapies can be scientifically and clinically challenging.

K2 has not disclosed the specific diseases targeted by its eight programmes, nor the terms of any asset acquisitions or licensing arrangements. That leaves key questions unanswered: how differentiated the candidates are, how much clinical data already exists, and how K2 will decide which assets deserve priority.

The competitive field

K2 Therapeutics will be entering a crowded global race. In Asia, companies such as China’s Akeso, Kelun-Biotech, DualityBio and RemeGen have drawn attention for antibody-based oncology drugs and ADC pipelines. Singapore has also produced antibody and oncology-focused biotechs such as Hummingbird Bioscience, while larger global players including Genmab, BioNTech, AstraZeneca, Gilead and Daiichi Sankyo are investing heavily in next-generation cancer therapies.

The competition is not only for patients or market share. It is also for assets, clinical trial sites, scientific talent, manufacturing capacity and partnership attention from big pharma. For a Singapore-based biotech, that means regional credibility alone will not be enough. K2 will need to show that its pipeline can stand up to global scientific scrutiny.

Still, Singapore offers some advantages. Its regulatory environment is considered predictable, its biomedical research base is deep for a country of its size, and its position in Southeast Asia gives companies a regional operating base close to diverse patient populations. For founders and investors, the challenge is turning those strengths into globally competitive drug development companies rather than regional outposts for multinational pharma.

K2’s US$50 million seed financing is therefore more than another funding announcement. It is a test of whether Singapore can host the next generation of biotech companies that are not merely doing research, but assembling, developing and potentially commercialising therapies for global markets.

Also Read: From lab to factory floor: ChemT nets US$4M to make cell therapies easier to manufacture

For now, K2 Therapeutics has capital, a sizeable early pipeline and a CEO who has taken advanced therapies from development into commercial scale. The harder part begins next: proving that the assets it has gathered can survive the long, expensive and unforgiving path from promising science to medicines that patients can actually use.

The post K2 Therapeutics raises US$50M to build global biotech pipeline from Singapore appeared first on e27.

One of the laziest ambitions in strategy is the desire to own the whole stack.

It sounds bold in leadership meetings. It sounds defensible in investor conversations. It sounds like control. If we own more of the value chain, more of the customer relationship, more of the workflow, more of the economics, then surely we are building a stronger position.

Often, we are doing the opposite.

In many markets, trying to own everything is not a sign of strength. It is a sign that the firm has not yet understood where leverage actually lives. Ownership expands surface area. It increases execution burden. It drags the company into activities where it may have no real advantage. It creates cost, complexity, and management sprawl. Worst of all, it can distract leaders from the far more important question. Which part of this system truly matters enough that others will keep orienting around us, even if we do not own the rest.

That is where strategic chokepoints come in.

The strongest positions often sit between assets, not on top of them

A surprising amount of strategic thinking still assumes power sits with the party that owns the most assets. More infrastructure, more products, more distribution, more channels, more touchpoints. The image is imperial. The larger footprint must mean the stronger position.

Real markets are often organised differently.

Some of the most durable positions sit not with the actor that owns everything, but with the actor that sits at the point where different things have to come together. The place where supply meets verification. The place where data becomes decision. The place where activity becomes auditable. The place where users become billable. The place where risk becomes governable. The place where systems that do not naturally speak to one another must suddenly agree.

A chokepoint is where uncertainty has to be resolved

The clearest way to identify a real chokepoint is to stop asking where activity happens and start asking where uncertainty must be settled before activity can continue.

That is the deeper strategic move.

In many markets, the most valuable position is not at the point of creation or consumption. It is at the point of resolution. The place where someone has to decide whether identity is real, whether payment can be trusted, whether compliance is sufficient, whether a model output is acceptable, whether a supplier is approved, whether risk is within tolerance, whether a transaction can be recorded as final, whether a failure can be recovered without chaos.

Also Read: Why Southeast Asian startups should stop treating Europe as one market

Those moments are strategically rich because they are not optional. The surrounding market can innovate, fragment, diversify, and compete aggressively, but when it reaches a point where uncertainty must be converted into confidence, somebody has to perform that function.

Whoever performs it well can become disproportionately powerful.

Leverage is usually designed at the point where others need certainty

The original strategic instinct behind many great businesses is not, how do we own more. It is, how do we become the answer at the moment others need certainty faster than they can create it themselves.

That is a much more intelligent design question.

A strategic chokepoint can emerge around trust. It can emerge around technical compatibility. It can emerge around data custody. It can emerge around regulatory interpretation. It can emerge around reconciliation, recovery, settlement, or proof. What matters is not the category name. What matters is whether others start depending on that point to turn ambiguity into action.

This is why the best chokepoints often feel smaller than the markets they influence. They are concentrated. They do not need to carry the whole weight of the system. They only need to sit at the moment where the system cannot proceed safely, credibly, or efficiently without them.

Once that happens, leverage follows almost naturally.

The weak version of this idea is bottlenecking, the strong version is coordination

Not every chokepoint is strategically healthy. Some are little more than bottlenecks. They create friction without adding enough legitimate value. They slow the system down, tax it, or trap participants through inconvenience rather than through necessity. Those positions may produce short term leverage, but they also invite resentment, workaround behaviour, regulation, or eventual displacement.

The stronger version of a chokepoint is different. It improves coordination.

A legitimate chokepoint does not merely obstruct passage. It makes passage safer, faster, more intelligible, more governable, or more trusted. It reduces transaction cost. It lowers institutional anxiety. It gives multiple participants a shared basis on which to act. It helps the market function at a level of scale or complexity that would otherwise be difficult to sustain.

That is why the best strategic chokepoints are not experienced as pure extraction. They are experienced as useful compression. They narrow the system at the exact place where narrowing is valuable.

This is also why they last. Participants may not enjoy dependence, but they will tolerate it when the alternative is disorder.

Designing a chokepoint means designing a habit in the market

A useful way to think about strategic leverage is that the company is not simply building a product or service. It is trying to build a habit in the market.

Not a consumer habit in the narrow behavioural sense, but a systemic habit. A repeated pattern in which others begin to assume that before they proceed, they should pass through this layer. Before a model is trusted, it must be reviewed here. Before a vendor is activated, it must be cleared here. Before value is counted, it must be recorded here. Before a workflow scales, it must connect here.

That habit is what turns a useful position into a durable one.

Also Read: The myth of the neutral stack: Why SEA startups can no longer sit on the fence

The deeper point is that leverage compounds when the market starts organising itself around your existence without having to be forced. Once institutions begin embedding you into policy, process, reporting, integration design, or internal governance, the relationship is no longer just commercial. It becomes operational and cognitive. You are no longer merely chosen. You are expected.

That is the point at which designing a chokepoint starts to look less like product expansion and more like market architecture.

The danger is becoming so powerful that you weaken your own legitimacy

The more important a chokepoint becomes, the greater the temptation to overuse it. Companies start increasing take rates, privileging their own offers unfairly, reducing transparency, or changing rules in ways that maximise extraction at the expense of trust. That is usually the beginning of strategic decay, even if the financial effects take time to show.

A chokepoint remains durable only while participants believe the power attached to it is being exercised in a way that preserves the health of the broader system. Once that belief breaks, market actors start building alternatives, regulators become more interested, and internal defenders inside customer organisations become less willing to protect the relationship.

This is why the strongest chokepoints are governed, not merely exploited.

They carry a burden of stewardship. The company at the centre has to act in ways that keep the market willing to route through it. That means predictability, fairness, quality control, and enough restraint that dependence does not start to feel intolerable.

In other words, the position has to remain useful enough to stay legitimate.

—

The post Strategic chokepoints: Designing leverage without owning everything appeared first on e27.