From the CIO’s Desk – Asia Insights February 2026

February delivered strong returns across much of Asia, but volatility, as ever, was waiting around the corner. The US-Israel and Iran conflict has injected fresh uncertainty, while the indiscriminate sell-off in software and consumer internet names has tested investor conviction. In this month’s commentary, we share our assessment of the Middle East conflict, explain why we believe the AI disruption narrative is overdone, and make the case for the old economy’s quiet resurgence. In times like these, the ability to separate signal from noise is what defines long-term performance.

Market Review

The MSCI All Country Asia Ex-Japan Index rose 8.21% (in USD terms¹) over the month of January. Relative to the rest of the region, South Korea and Thailand were the top performers, while China and Indonesia were the laggards. Sector-wise, IT and Industrials continued to lead performance, while Communication Services and Consumer Discretionary were the worst performers.

MSCI China declined 5.68% in February, pulling back amid subdued economic data and cautious sentiment ahead of the National People’s Congress (NPC). The official manufacturing Purchasing Managers’ Index (PMI) slipped to 49.0 (vs 49.3 in January), partly weighed by disruptions from the nine-day Lunar New Year holiday². However, the private Caixin/S&P Global PMI told a different story, surging to 52.1 (the highest since December 2020), suggesting stronger conditions among smaller, export-oriented manufacturers. Lunar New Year consumption showed mixed signals: tourism trips hit a record 596 million and spending reached RMB 803.5 billion (~USD 116 billion), both up 19% y/y.³ However, average spending per trip fell 0.2%, reflecting ongoing price sensitivity.

Indian equities returned 1.45% in February, modestly underperforming the broader region but maintaining a positive trajectory. The highlight of the month was the announced US-India trade agreement framework reducing US tariffs on Indian goods from 25% to 18%, while India committed to lower tariffs on US industrial and agricultural products. The deal also included a USD 500 billion purchase commitment over five years, improving trade visibility and reducing uncertainty. Domestically, the FY27 Union Budget maintained fiscal discipline with a 4.3% deficit target while increasing spending 7.7%. The RBI held rates at 5.25% with a neutral stance and raised its FY26 GDP growth estimate to 7.4%. Manufacturing PMI hit a four-month high of 56.9, signaling continued expansion from robust domestic demand.⁴

Korean equities surged 22.08% in February, extending their position as the region’s standout performer. The KOSPI broke above 5,500 to hit fresh record highs, powered by the AI-driven memory semiconductor supercycle. Samsung Electronics surpassed 1,000 trillion won in market capitalization for the first time, while SK Hynix continued to rally on the back of insatiable demand for high-bandwidth memory (HBM) chips from hyperscale AI data centers. Foreign buying was a significant tailwind, with overseas investors adding meaningfully to their positions, though late-month profit-taking saw a record single-session foreign sell-off. Retail investors, who had largely sat out the earlier stages of the rally, also began participating more actively, adding further momentum this month.

Taiwanese equities rallied 12.75% in February, with the TAIEX surging past 35,000 to a new all-time high, led by TSMC. Foreign institutional investors bought Taiwanese equities aggressively, recording their largest single-day purchases in over two decades. TSMC alone accounted for a significant portion of index gains, as continued demand for advanced logic chips and AI-related capital expenditure from US hyperscalers underpinned the rally. Broader tech names, including Delta Electronics and Hon Hai, also advanced on robust AI server demand.

Within ASEAN, Thailand was the month’s standout, with equities surging 20.57%, driven by a decisive election result on February 8 that saw Prime Minister Anutin Charnvirakul’s Bhumjaithai Party secure a strong mandate, boosting political stability. The rally was amplified by a surprise 25bp rate cut by the Bank of Thailand to 1.00%, which buoyed sentiment in rate-sensitive sectors including banks and property. The Philippines also delivered solid returns of 7.27%, supported by continued monetary easing and improved macro sentiment. Meanwhile, Indonesia was the region’s laggard at -1.14%, weighed down by persistent foreign outflows and subdued commodity sentiment.

Portfolio Commentary & Outlook

Navigating the Middle East Conflict

Note: This outlook was prepared as of March 8, 2026. Given the rapidly evolving nature of events, we will continue to update our views as the situation develops.

The US-Israel military operation against Iran has been the dominant risk event shaping global markets since early March. The conflict has since widened beyond its initial scope, with disruptions to air travel, energy logistics, and shipping in the Strait of Hormuz.

Our base case remains that the conflict is more likely to move toward a negotiated resolution in the near term rather than a prolonged military engagement. Multiple parties have strong incentives to de-escalate:

• From the US perspective, midterm election primaries are already underway, and the administration faces growing domestic pressure as both the human and fiscal costs mount. More importantly, the administration’s most likely off-ramp is economic pain. Markets initially assumed the conflict would be short-lived, and that the administration would, as it has before, “TACO” and pull back quickly. But with oil now above USD 110 per barrel and equity markets falling, the pressure is mounting. We believe that another 10-15% move higher in oil or a further 5-10% decline in markets could be the catalyst for the administration to declare that the operation has achieved its objectives and wind down military action. We saw this playbook during the tariff episodes where it was ultimately market and economic pain that forced the reversal. We expect a similar dynamic here.
• China, as the world’s largest importer of oil and gas, has a direct economic interest in restoring stability – Beijing has publicly called for an immediate ceasefire and a return to dialogue, while pressing Iran to ensure safe passage through the Strait of Hormuz.
• India’s oil supply is also being disrupted, adding another major economy to the coalition of voices urging resolution.
• GCC states, several of which have been struck by Iranian retaliatory fire, have publicly emphasized neutrality, condemned the attacks, and urged de-escalation through diplomacy. Aside from impacts to oil and gas and tourism, GCC states also face threats to food security while passage through the Strait of Hormuz remains limited.

We continue to monitor several signals that suggest further escalation may be contained:

• Early in the conflict, the UAE opened safe air corridors for commercial flights, a step that would typically not be taken if authorities expected the conflict to widen materially.
• China’s actions – calling for immediate ceasefire and pressing Iran on Hormuz – are consistent with the precedent set during the Venezuela situation, where both China and Russia expressed strong rhetoric but ultimately stayed out of direct military involvement.
• Finally, the majority of Iranian missiles launched have been intercepted, limiting the scale of physical damage. But according to certain press reports, the US’s interceptor stockpiles, having been used over the past few years, can be depleted quickly if the conflict is prolonged.

The key escalation risks we are watching include any retaliatory action by GCC nations beyond self-defense, which would represent a meaningful widening of the conflict. We are also monitoring for any indication that China or Russia may move beyond rhetoric to direct involvement – this remains a tail risk, though current signals suggest it is unlikely.

From a portfolio perspective, we have made no changes to our portfolio as a result of the conflict. The primary transmission mechanism to Asian markets is through oil prices, but this remains manageable given that oil reserves across the region are abundant, providing a buffer against near-term supply disruption. Any portfolio adjustments made in recent weeks, such as trimming Korean semiconductor positions, were driven by valuation discipline, not geopolitical considerations. However, the sharp correction across Asian equity markets in early March has created value in sectors such as industrials, financials, consumer discretionary, and technology, which we may look to take advantage of as clarity emerges.

The “SaaSpocalypse” and Why We Think Moats Matter More Than Models

Markets have taken a “shoot first, ask questions later” approach to AI disruption. Following the widely circulated Citrini Research report warning of a dystopian wave of AI-led disruption, a massive sell-off swept through software-as-a-service (SaaS) stocks and broader consumer internet names globally. The fear is twofold: 1) that agentic AI will render traditional aggregator apps obsolete (e.g. booking travel, hailing rides, ordering food) without needing a human to navigate an interface; and 2) that “vibe coding” and off-the-shelf large language models (LLMs) will make it trivially easy for enterprises to build in-house tools, eliminating the need for expensive third-party software.

We choose to ask questions first.

Can everything be disrupted? Perhaps not. We believe the current correction is a sort, not a crash. Markets are in the process of separating value-adding platforms with genuine moats from “thin” apps that merely provide an interface for data. That sorting will ultimately be healthy, but in the meantime, the indiscriminate sell-off has created significant dislocations. Here is why we think the market has overreacted.

Enterprise adoption takes time. Ask your own IT team how many systems they plan to replace with AI in the next 12-18 months. The answer is likely very few. There is a process that can’t be shortcut, including evaluation, vendor selection, testing, compliance review, and integration. If enterprises replaced their entire software stack overnight, businesses and economies would falter. As history consistently shows, the impact of technological disruption is overestimated in the near term and underestimated in the long term. Adoption will happen, but on the enterprise’s timeline, not the market’s.

B2B disruption will come faster than B2C. This distinction matters. In our own business, we are already going to our vendors, asking how they are leveraging AI, and requesting that they share some of their productivity gains. That dynamic, where procurement teams demand AI-driven efficiencies from their suppliers, will accelerate disruption in B2B software. B2C, however, is a different story. Consumer-facing businesses that have spent years fine-tuning their operations, managing queries, returns, last-mile logistics, dispute resolution, etc., have built deeply embedded processes that are far harder to replicate. When there is a physical back end involved, operational sophistication is the moat.

Physical assets are being completely discounted. The sell-off has ignored the real-world infrastructure that underpins many consumer internet companies, including their fulfilment centers, warehouses, delivery fleets, and logistics networks. Regardless of how AI might reshape the front-end of consumer discovery, goods still need to move from point A to point B, returns still need to be processed, and service issues still need to be resolved. Companies with these tangible assets will endure.

Differentiation will come from taste and real-world presence, not just technology. The “shoot first” approach assumes that everything will be commoditized. But in a world where software interfaces may converge, differentiation will increasingly come from taste – consumers’ preferences for quality, brand, experience, and trust. AI may help consumers find the best options that match their preferences, but companies still need to deliver goods and services that stand apart. Moreover, it is naïve to assume that consumers will do everything online. The brands that truly understand their customers know this. Shopping malls, for example, have reinvented themselves for Gen Z, blending experiential retail with social and lifestyle elements that can’t be replicated online. The companies best positioned to thrive have an omnichannel presence – tech savvy companies with local interfaces, local presence, and local servicing capabilities to meet customers both offline and online. Another example is ride-hailing, where the online experience needs to match the offline execution. Established players like Grab have over a decade of driver and rider data, trust ratings, and operational refinement that are extremely difficult to replicate, regardless of how sophisticated a new AI-powered entrant may be.

Incumbents are rarely truly disrupted. There is an implicit assumption in the sell-off that established companies are too slow to adapt. But if we look into history, in the consumer internet space, it is remarkably rare for an incumbent with a deep moat to be truly displaced. The only clear example we can point to is ByteDance’s Douyin, where a genuinely superior algorithm disrupted incumbents. Beyond that, incumbents typically adapt, acquire, or simply out-execute. Moreover, many Chinese consumer internet companies have survived periods of far more intense competitive pressure, including regulatory crackdowns and subsidy wars, and have emerged stronger. To assume they will be asleep at the helm while AI disruption unfolds is to dismiss their track record. Smart incumbents will use LLMs to innovate, and those with strong balance sheets can acquire disruptive AI startups or build solutions in-house, much as Microsoft did by rapidly incorporating features and acquiring strategic assets.

AI will expand and retreat, like Uber did. A useful analogy is Uber’s global expansion: the company expanded aggressively into numerous markets where it faced entrenched local competitors with superior regulatory relationships, cultural understanding, and willingness to compete on local terms. Many of these markets generated minimal revenue while consuming disproportionate management resources and creating regulatory liabilities, leading the company to ultimately retreat. We think AI disruption will follow a similar pattern. It will expand into every category and across every market, but not every segment will support rapid AI adoption. We will see it retreat from areas where the unit economics, regulatory environment, or consumer behavior do not align. Companies with deep domain expertise, that can identify and address clear pain points, will continue to provide value.

What about jobs?

The more significant issue, as alluded to in the Citrini report, is AI’s potential impact on employment. As we first wrote about in our September commentary, we believe AI-driven job displacement – particularly in white-collar, entry-level, and middle-management roles – is a real and valid concern. But we do not expect large-scale job losses to materialize within the next one to two years. The more likely timeline is four to five years or beyond, as enterprise adoption cycles and organizational change management takes time.

What we also haven’t fully baked in is the likelihood of government intervention. If job displacement begins to materialize at scale, policymakers will almost certainly respond – whether through regulation, retraining programs, or fiscal measures – to minimize the pain of disruption.

In the near term, we see offsetting forces. New business formation in the US is surging, fueled by AI and LLMs that are dramatically reducing the cost and complexity of launching a company. As these firms scale, they will hire for sales, operations, customer success, and roles we have not yet imagined – underscoring that AI is more likely to strengthen than disrupt the labor market over the next couple of years. The gold rush in AI applications is real: capital that was previously in crypto is now being redirected into AI app development, adding further momentum to the startup ecosystem.

Chart. Weekly business formation in the US is rising rapidly, likely driven by AI

Source: US Census Bureau, Shikhara Investment Management, February 2026

Staying focused on the rotation

As we outlined in last month’s commentary, we maintain that the defining story of this year is the rotation from AI enablers to AI adopters. The early signs are already visible. Nvidia, despite delivering another blockbuster quarter, is seeing its share price plateau. It becomes increasingly difficult for companies at such extraordinary market capitalizations to compound at past levels. Meanwhile, questions remain about whether hyperscaler capex will actually be executed at the levels estimated for the coming years. A significant portion of planned AI infrastructure investment is expected to be funded through private markets, but with private assets now marked down due to software exposure, the ability of private credit to support the required capex is uncertain. The “AI enabler” trade is becoming stretched, and when the market recognizes that it has been too dismissive of the adopter side of the equation, a correction will follow.

This is exactly the environment in which our investment approach comes into its own. We are investing in “last man standing” companies – businesses with deep moats, proprietary data, fulfilment and logistics infrastructure, and demonstrated value-add that cannot be easily replicated by an AI agent or a new entrant. These are the companies that will not only survive the market’s sorting but will emerge stronger from it.

Revenge of the Old Economy: The upcoming manufacturing renaissance

This is a theme we have been building conviction on for over a year. In May 2025, we identified what we called “The Coming Investment Cycle” – the consequence of 17 years of limited capital investment since the Global Financial Crisis, during which the rise of capital-light tech platforms captured the lion’s share of market attention and capital allocation. US companies had been spending overwhelmingly on software rather than physical equipment, while the age of US capital stock reached its oldest level in post-war history. In our June half-yearly report, we formalized this as one of three core investment themes about the multi-year infrastructure and manufacturing capex.

That thesis is now playing out in real time, and the AI revolution is accelerating it. AI is not just a software story – it is profoundly infrastructure-heavy. The buildout demands physical construction: concrete, steel, cabling, transformers, and the skilled workers to install them. Every new gigawatt of AI compute capacity necessitates an ecosystem of railroads, power grids, and logistics networks to support it. In any technological revolution, there are picks-and-shovels beneficiaries, and this cycle is no different.

Meanwhile, the narrative around AI and jobs has been almost exclusively focused on white-collar displacement. But a closer look at labor markets in the US and Europe reveals a very different story for blue-collar workers. The construction industry in the US faces a monthly shortfall of approximately 350,000 workers, with 94% of firms reporting difficulty filling positions.⁵ The EU also reports labor shortages in 98% of occupation categories.⁶ Critically, this is not cyclical – it is structural, driven by the convergence of an aging workforce (over 21% of US construction workers are now 55 or older), a depleted vocational training pipeline, immigration constraints, and a cultural bias toward university education that has diverted generations away from skilled trades.⁷ The combined US-EU blue-collar workforce deficit could widen to 4-6 million unfilled positions by 2030.⁸

This has several important implications. First, it provides a powerful counterpoint to the AI job displacement narrative. While AI may displace certain white-collar roles over the medium term, blue-collar labor is becoming scarcer, and automation in physically complex sectors like construction and skilled trades remains years away from making a meaningful impact. Second, wage growth in blue-collar occupations is already outpacing white-collar, a trend that is likely to persist and benefit categories catering to this workforce. Third, companies that own physical infrastructure (e.g. the fulfilment centers, logistics networks, and industrial assets that underpin the real economy) are being undervalued by a market fixated on software disruption.

Given this, we maintain a significant exposure to Asian industrial champions in our portfolio, including in shipbuilding, power equipment, defense, and manufacturing – companies that are natural partners in the global rebuild. Our investment approach is deliberately unconstrained, spanning the growth-value spectrum to capture the best opportunities across Asian markets. While we focus on innovative, high-growth companies, we equally pursue value where competitive dynamics are shifting in ways the market has yet to price in. As of today, despite strong performance over the past year, the portfolio trades at just 13.4x two-year forward P/E, while estimated 3-5 year earnings growth sits at 21.3%. Importantly, not every position is firing at once. Our one-year attribution analysis shows that roughly half of securities in the portfolio returned less than 25% last year. We view these not as underperformers, but as dry powder – competitively robust businesses with identifiable catalysts for growth acceleration or value unlocking in the quarters ahead. This balance between contributors and coiled springs is by design, and it is how we build a portfolio that compound through cycles, and not just in them.