Markets looking increasingly complacent

Cause for caution persists
It has been a difficult few months to navigate through these choppy markets in Japan, but in the end, technology and AI names proved to be a safe refuge against surging oil prices and the fear of supply disruptions caused by the latest Middle East conflict.

We are seeing early signs that shipping and logistical issues and the general rise in input costs, across the board, are threatening major cost overruns for the 3/27 term which could weigh on profit margins. With reportedly about a month left before actual fuel shortages hit immediate supplies to Asia and Africa, we think the clock is ticking for the Strait of Hormuz to be opened up for commerce or physical oil price could easily surpass recent highs and stay well above the indicated futures price, currently around $110.

Come what may, we believe this conflict with Iran has significantly redrawn the geopolitical map from its post-war order and one which has already shaken the foundation of OPEC following the UAE’s exit from the cartel last week. We think the outlook for the energy market looks to have been permanently altered.

Although near-term prices look set to spike if the current stalemate persists and the strait remains closed to traffic through the Persian Gulf, the crisis has arguably also pulled forward the peak in global oil demand which some argue has already passed. Industries are aggressively pivoting back toward coal, nuclear, and renewables, while spiking fuel costs have simultaneously revitalised global demand for EVs and hybrid vehicles.

These are some big changes worth pondering about and long/short opportunities they may bring. One name which is off our Japan remit but one we are very excited about is BYD (1211) which we believe could be seeing a market transformation, not unlike what we saw for the Japanese car makers back in mid-70s, after the first oil shock put their fuel-economic cars on the US map.

We think this latest oil shock has drowned out the inflationary wave we had been expecting from the AI capex boom. Indeed, we are even more concerned now that inflation will prove a far bigger problem to solve and interest rate cuts will be very difficult to justify by the US Federal Reserve, which is due to have a new chairman soon

BOJ’s continued inaction weighs on the yen
In Japan, the BOJ in our view remains years behind the curve in normalising its monetary policy mainly because it has not raised its base rate fast enough (from the current 0.75%), to help salvage the value of Japan’s currency and with it, the nation’s battered purchasing power. We remain critical both of the Takaichi government and the BOJ for not recognising Japan’s urgent need for a stronger yen policy and the much tighter monetary policy required to get us there.

Although the yen has held its own thus far since the intervention, we think its unilateral currency interventions will prove futile in changing the fundamental forces that have weighed on the Japanese currency. This would likely require a more hawkish monetary policy stance, which in our view, necessitates 50bps clips in interest rate hikes. This would shock the currency market and send a clear message that Japan policy makers are adopting a stronger yen policy change.

However, growing economic uncertainties stemming from the war are unlikely to allow BOJ too much room for being aggressive now. With yet another window of opportunity to raise rates squandered, we think it may become very difficult to shake off yen short sellers by interventions alone.

Ultimately, Japan’s central bank may be forced to act and raise rates should foreign currency traders choose to test the will of the government and continue to short the yen into MOF’s interventions above the dollar/yen level of 160, where they seem to have drawn their resistance line.

Growing delays in data centre completions becoming a major concern
To recall, we had become increasingly concerned about the immediate inflationary impact of AI spending since late last year and have argued that the market underestimates the impact of rising costs of utility bills, memory chips, CPUs, PCBs, gas turbines, optic fibre cables and everything else seeing bottlenecks from surging data centre demand where market prices seem totally inelastic and securing supplies remains a top priority.

We’ve argued that surging memory prices was the canary in the coal mine and this secular force will likely push up prices of almost anything it touches including wafer makers which we think are next in line, and should raise their 300mm wafer prices which is why we think Sumco (3436) still has huge upside despite its big rerating.

But looking at the bigger picture of AI, we have also become increasingly worried about the structural headwinds facing the AI boom at the physical level. We are seeing AI data centre project completions being increasingly pushed out as projects pile on, creating a massive logjam in the development pipeline. This isn’t just a matter of logistics but there is a growing dislike of these projects by local communities who view them as resource-draining monoliths.

Shortages of critical power and water—necessary for the cooling of high-density AI clusters—are becoming chronic. As utilities struggle to keep up with the load and local opposition hardens against the rising utility bills and the environmental footprint of these facilities, we suspect the timeline for the “next leg” of AI growth looks far more fragile than what has been priced in.

Delays in deliveries derived from the piling up of unfinished data centres seem a logical outcome in the months ahead and is the elephant in the room which could ultimately impact earnings visibility of the segment going into H2. But for now, investors seem more impressed with the growing backlogs and rising data centre capex and seemingly unconcerned about possible push-outs that began to hit half of all projects due for completion last year. That ratio is projected to rise even further this year and of the 12–16GW planned for 2026, only about 5GW is actually under construction.