AI Infrastructure
Portfolio
Q1 2026 · January – March 2026
Macro & Market Context
For the full macro and market environment commentary, please refer to the Emit Capital Global Opportunities Portfolio Q1 2026 Quarterly Report.
That report covers the Q1 Iran conflict, global oil shock, sovereign bond market analysis, the Emit Nexus investment thesis, and the full cross-asset review.
The key macro developments that shaped portfolio outcomes in Q1 2026, the Iran conflict and Strait of Hormuz closure, the global oil price shock, the synchronised repricing of sovereign yield curves across the US, Germany and Japan, and the central bank pivot from accommodation to hawkish restraint are covered in full in the Global Opportunities Q1 2026 Quarterly Report. This supplement focuses on the portfolio-specific performance, positioning, and outlook for the AI Infrastructure Portfolio.
Portfolio Performance
Performance Summary — AUD Returns to 31 March 2026
Swipe to view all periods
| 1 Mth | 3 Mth | 6 Mth | SI (Total) | |
|---|---|---|---|---|
| AI Infrastructure Portfolio | +1.4% | +0.4% | +5.7% | +31.0% |
| MSCI ACWI (Benchmark) | −1.8% | −10.2% | −8.7% | +5.3% |
| Active Return | +3.2% | +10.5% | +14.4% | +25.7% |
Performance is net of fees. Based on the aggregation of all managed accounts. Individual account performance may vary.
The AI Infrastructure Portfolio delivered +0.4% (AUD) for Q1 2026, outperforming the MSCI ACWI by +10.5 percentage points — a benchmark that fell −10.2% as broad equity markets sold off on the Iran oil shock and heightened AI cost scrutiny. Vertiv Holdings was the standout contributor at +311 bps, re-rating sharply as the Iran energy crisis elevated the value of mission-critical power infrastructure for AI workloads. Marvell Technology contributed +190 bps, driven by accelerating custom ASIC demand from hyperscalers. Micron Technology delivered +171 bps as HBM demand visibility improved materially. Primoris Services (+67 bps) and Evolv Technologies (+37 bps) rounded out the top five.
The primary detractor was Microsoft at −166 bps, reflecting AI cost scrutiny and multiple compression in the highest-valuation mega-cap names. Meta (−40 bps), Celestica (−35 bps), and Quanta Services (−30 bps) also detracted. The options overlay was critical in limiting the Q1 drawdown during March's severe risk-off episode.
Growth of $100,000. Approximate inception NAV series used for illustrative purposes.
Portfolio Analytics
Top 5 Contributors — Q1 2026
| # | Holding | Sector | BPS Contrib. |
|---|---|---|---|
| 1 | Vertiv Holdings | Industrials | +311 bps |
| 2 | Marvell Technology | Technology | +190 bps |
| 3 | Micron Technology | Technology | +171 bps |
| 4 | Primoris Services | Industrials | +67 bps |
| 5 | Evolv Technologies | Technology | +37 bps |
Top 5 Holdings — Portfolio Weight
| # | Holding | Sector | Port Wt % |
|---|---|---|---|
| 1 | Marvell Technology | Technology | 9.0% |
| 2 | Micron Technology | Technology | 8.8% |
| 3 | Meta Platforms | Technology | 7.5% |
| 4 | Primoris Services | Industrials | 7.5% |
| 5 | Celestica | Technology | 7.3% |
AI Infrastructure Investment Thesis
The AI Infrastructure Portfolio is the purest expression of Emit Capital's conviction that artificial intelligence infrastructure represents the most compelling structural investment opportunity of the current decade. Unlike the Global Opportunities Portfolio, which balances the AI infrastructure thesis across four regional sub-portfolios, the AI Infrastructure Portfolio concentrates entirely on the companies enabling, building, and powering the AI economy.
Q1 2026 reinforced the thesis on two fronts simultaneously: the hyperscaler capex cycle continued to accelerate beyond even optimistic expectations, and the Iran-driven energy price shock elevated the criticality of reliable, clean power for AI operations from a commercial consideration to a strategic and geopolitical one. The power constraint is now the defining bottleneck for the entire AI economy — and that constraint sits squarely within the investment universe of this portfolio.
The GPU and AI accelerator build-out is the most visible wave. Marvell Technology (the portfolio's largest holding at 9.0%, contributing +190 bps in Q1) and KLA Corp anchor the compute exposure. High Bandwidth Memory from Micron Technology, the second-largest holding at 8.8% and a +171 bps contributor, remains structurally supply-constrained as HBM adoption accelerates at every major hyperscaler. Advanced networking fabrics connecting GPU clusters are a non-discretionary spend, directly benefiting Marvell's custom ASIC and networking division as hyperscaler design wins compound.
Where compute demand meets the physical world. Vertiv Holdings, the portfolio's top Q1 contributor at +311 bps, provides power distribution and thermal management solutions that are mission-critical for AI data centre deployments as GPU power density continues rising. Primoris Services (+67 bps) and Quanta Services provide EPC contractor exposure to the grid and data centre infrastructure cycle. Amazon has guided to over $200 billion in AI infrastructure capex. Data centres now account for over 70% of new large-load interconnection requests to US regional grid operators.
Microsoft and Meta, each held at approximately 7.2–7.5% portfolio weight, represent the most advanced monetisation of the AI application layer, though both detracted in Q1 (−166 bps and −40 bps respectively) as the market repriced AI cost efficiency risk. Evolv Technologies (+37 bps) is an early-stage AI-driven physical security platform demonstrating how AI is penetrating non-traditional enterprise verticals with real commercial traction.
Portfolio Positioning & Outlook
The AI Infrastructure Portfolio is positioned to capitalise on structural investment themes within its universe, supported by the Emit Nexus derivatives overlay which provides active downside risk management and tactical flexibility across market regimes.
Key Positioning Themes — Q2 2026
- Maintain Vertiv as a high-conviction position. The +311 bps Q1 contribution validates the power distribution and thermal management thesis; as AI data centre power density continues to increase, Vertiv's mission-critical role deepens.
- Continue overweighting Marvell Technology (+190 bps) and Micron Technology (+171 bps). Both confirmed the thesis in Q1. Marvell's custom ASIC pipeline and Micron's HBM position are structurally supply-constrained multi-year opportunities.
- Reassess Microsoft (−166 bps) and Meta (−40 bps) sizing. Both face AI cost scrutiny and valuation compression headwinds unlikely to resolve quickly; selective trimming on any Q2 bounce would improve portfolio quality.
- Add to Primoris Services (+67 bps) and Quanta Services (−30 bps) selectively. Both are EPC infrastructure contractors directly exposed to the AI data centre and grid electrification capex cycle, with Quanta's Q1 dip representing a potential entry point.
- Use options overlay to manage hyperscaler earnings risk in Q2. With MRVL, MU, and NVDA all reporting, position-level hedges will protect accumulated gains while maintaining structural upside.
Portfolio Facts & Structure
| Name | Emit Capital AI Infrastructure Portfolio |
| Structure | Separately Managed Account (SMA) |
| Investor Type | Wholesale & Sophisticated |
| Asset Class | Global Listed Equities — AI & Technology |
| Regions | Global (thematic) |
| Benchmark | MSCI ACWI |
| Strategy Inception | 31 May 2025 |
| Holdings | 15–25 positions |
| Cash Range | 5–15% |
| Min. Investment | A$250,000 or equivalent |
| Pricing | Daily |
| Prime Broker | Interactive Brokers (IBKR) |
| Currency | AUD / USD |
| Allocation in Global Opps | 25% of Global Opportunities Portfolio |
| Manager | Emit Capital Asset Management Pty Ltd |
| AFSL / ABN | 551084 | ABN 57 652 326 237 |
This portfolio is one of five strategies managed by Emit Capital Asset Management. For the full firm overview and Global Opportunities Quarterly Report, visit emitcapitalam.com.