Indonesia stocks lower at close of trade; IDX Composite Index down 1.10%

The Jakarta IDX Composite fell 1.10% on the session. Top gainer Nusantara Almazia jumped 34.72% while Hotel Fitra International dropped 15.00% and Asiaplast hit a three‑year low, down 13.79% to 250.00. WTI crude for May rose 1.58% to $95.97/bbl and Brent for June gained 1.82% to $103.74; USD/IDR strengthened 0.54% to 16,972.10, indicating commodity-driven moves amid mixed local equity flows.

Analysis

A Middle East geopolitical shock that lifts energy volatility and stresses EM FX flows reorders capital: near-term cash allocators favor liquid safety and commodity plays, but capital-intensive AI compute cycles are relatively inelastic. Higher energy costs increase total cost of ownership for large-scale training runs, creating demand for denser, more energy-efficient server platforms — a structural advantage for vendors who can pair GPU supply with efficient chassis and services.

Supply-side constraints (GPU allocations, export controls, logistics) are the most immediate gating factor for compute hardware revenue — not end-market demand. If OEMs like SMCI can secure multi-quarter GPU delivery slots, they convert backlog to high-margin aftermarket and integration revenue; if GPU supply tightens further or export curbs widen, order fills and margins get pushed out, creating asymmetric execution risk over 3-9 months.

Adtech platforms (APP) face a different transmission: ad budgets are cyclical and earliest to be cut in a growth scare, especially where EM monetization suffers from local currency weakness. However, AI-driven yield improvements (better targeting, lower waste CPM) can offset budget declines over 1-4 quarters if product adoption is fast; the pacing of that adoption — and FX-driven ARPU compression in EM geographies — is the key variable.

Consensus will likely overweight commodities and EM cyclicals; the contrarian read is that compute capex and ad-yield improvements keep a two-speed market where AI-capable infrastructure providers capture outsized cashflows despite macro noise. Position sizing should reflect that asymmetry: overweight suppliers with secured GPU access and energy-efficiency differentiation, caution on pure ad-revenue exposed names until ad yields demonstrate stability.

Key Decisions for Investors

• SMCI — Tactical long (1.5% NAV): Buy SMCI stock on a pullback of ~8-12% or on any confirmation of multi-quarter GPU allocations. Target +50-80% in 6–12 months if backlog converts; hard stop -20% from entry. Alternate: buy a 12-month call spread (buy ATM, sell +35% strike) to cap cost and target 3:1 upside/downside.
• APP — Selective long (1.0% NAV): Initiate a smaller position in APP, sizing to volatility (1% NAV) and use options to time exposure (6–9 month calls). Target +30–40% in 3–6 months if AI-driven CPMs/ARPU stabilize; use -18% stop on the equity or cut options if monthly ad-revenue prints miss expectations.
• Execution hedge — Protect compute exposure: For SMCI positions, buy 3–6 month put protection equal to ~30% notional (or purchase collar) to insulate against sudden GPU export-control headlines or a demand shock. If oil/FX-driven EM outflows accelerate, trim APP first and keep SMCI exposure via hedged option structures.