Moving into H2 2026 – resilience over perfection

MARKETS like clean stories. Unfortunately, the world keeps offering messy realities.

As we enter the second half of 2026, corporate earnings have shown real resilience, with consensus earnings growth revised from around 12 per cent to closer to 29 per cent. This reflects companies walking into a challenging macro environment yet leaving with earnings still delivering durable growth.

Companies are proving more adaptable than investors initially feared. Margins have held up. Technology earnings remain powerful. Artificial intelligence spending has not collapsed under its own hype. Consumers have slowed in places but are not broken. Supply chains remain imperfect but no longer look as fragile as they did during the pandemic years.

That is the resilience part.

The perfection part is more uncomfortable. Markets have already priced in plenty of good news. Since the Iran war’s first official ceasefire on Apr 7, semiconductors surged 76 per cent in just 40 days before pulling back.

Investors still want the AI trade. They just no longer want to pay any price for it.

So our theme for the second half is simple: resilience over perfection.

That is not because we are dismissing opportunity. Earnings resilience gives markets support. But perfection pricing leaves little room for disappointment. When expectations are this high, even good results can receive a bad review. Markets are not grading effort. They are grading surprise.

The question is not simply whether growth can continue. It is whether portfolios can capture resilient earnings, while withstanding markets that may already expect too much.

The Fed backdrop also matters. The US Federal Reserve is now led by Kevin Warsh, appointed by US President Donald Trump. With this, the bar to hike rates is very high.

There was an unspoken dovish slant to the appointment, even if nobody wants to print that on a mug. With the 10-year US Treasury yield near mid-4 per cent levels, the 30-year Treasury testing 5 per cent, and 30-year mortgage rates around 6.5 per cent, policy is already restrictive. That is far from the 2.6 per cent mortgage rates seen after the pandemic.

The Fed can talk tough but it cannot be too hawkish without tightening financial conditions further.

AI moves from talking to doing

A major market theme remains as no surprise – AI. But the story is evolving.

The first phase of generative AI was about content creation, search, summarisation and coding assistance. Useful, yes. Transformational, possibly. Occasionally irritating, definitely.

The next phase is agentic AI. These systems can work towards goals, reason through tasks, interact with software and escalate exceptions. Put simply, most AI today talks. Agentic AI acts.

That matters because enterprises not only need better answers, they need lower costs, faster workflows and measurable return on investment. The opportunity is broader than models. It includes enterprise software, cloud platforms, cybersecurity, data platforms and workflow redesign.

Chips still matter. But chips need memory, networking, cooling, electricity and somewhere to sit. In the real world, silicon still needs substations.

The risk is that markets have moved faster than deployment. Demos are easy. Production systems need reliability, auditability, governance and clear ownership. Deployment is where PowerPoint meets the procurement department.

Return of the spare tyre

The second market theme is strategic re-industrialisation.

For decades, companies built supply chains around efficiency. Inventories were lean. Production was concentrated. Capital was cheap. The goal was simple: make everything as low-cost as possible, preferably just in time.

That model worked beautifully – until it did not.

Pandemic shortages, semiconductor bottlenecks, export controls and geopolitical shocks have changed market behaviour. Companies and governments now care more about resilience, redundancy and domestic capability.

“Just-in-case” is less elegant than “just-in-time”. But it is most useful when ports close, chips run short or energy routes become geopolitical poker chips.

This is not a short cyclical bounce. Governments want control over critical industries. Companies want supply chains that can survive disruption. Energy security is no longer just economic but also strategic.

Power grids, electrical equipment, nuclear, renewables, storage, automation and critical materials should benefit. Resilience is not free. It usually arrives with a larger invoice. But it also increases the value of assets with scarcity, pricing power and regulatory support.

China is not one trade

China also fits the resilience theme, but selectively.

The mistake is to treat “China exposure” as one tidy bucket. Onshore A-shares, especially growth indices such as ChiNext, are more exposed to hardware, industrial upgrading, semiconductors and advanced manufacturing. These sectors align with Beijing’s push for “new quality productive forces”.

Offshore H-shares look different. They remain dominated by Internet platforms, financials, telecoms and energy names. The platforms are no longer being valued simply as asset-light compounders. They are spending heavily on AI models, cloud infrastructure and chips amid cut-throat competition.

That matters because investors are questioning the payback period. H-share platforms are fighting for consumer wallets, while margins are diluted by capex and subsidies. We think A-shares house the upstream beneficiaries of that spending.

The initial public offering pipeline reinforces this divide. Hong Kong issuance is recovering, helped by A+H listings and potential American Depositary Receipts returns. Onshore reforms across Star Market and ChiNext are reopening pathways for semiconductors, AI and biotech companies, even before traditional profitability thresholds are met.

So, the China answer is not capitulation. It is precision. Existing H-share exposure can be held where valuations are depressed and earnings catch-up remains possible. Incremental capital should be deployed selectively into A-share growth leaders, where policy alignment and earnings visibility are stronger.

Bullish on trend, vigilant on valuation

The second half of 2026 offers opportunity. But it also demands discipline. We remain constructive, but not complacent.

The AI trade currently looks richly positioned, its momentum ironically prolonged by the looming US-Iran deal, though a period of consolidation is likely. But these dips should be bought selectively. Second-quarter results starting in July could show another round of solid earnings, with technology still leading.

But we would not be surprised that the earnings growth broadens beyond AI.

Finally, history also offers a useful reminder. The Nasdaq 100 has risen 175 per cent since ChatGPT was launched. But during the dotcom mania of the late 1990s, the Nasdaq 100 grew by around 10 times before peaking. That does not mean history must repeat itself. It means structural themes can run longer than sceptics expect, while still suffering brutal corrections along the way.

Portfolios should not be built for perfection. They should be built for endurance.

An elegant portfolio is nice. A resilient one is better. Investors do not need a sports car on a perfect highway. They need an off-road SUV with a full tank, a spare tyre and enough patience to buy the next dip.

The writer is head of investment strategy, UOB Private Bank