Two narratives are competing for attention in markets this week — and both have direct implications for how sophisticated investors should be positioning capital right now.

The first: Bitcoin outperformed every traditional safe haven during the opening weeks of the U.S.-Israeli strikes on Iran. The second: agentic AI systems — autonomous agents capable of planning, executing, and adapting trading decisions without human intervention — have moved from pilot programs into production at the world's largest capital markets institutions.

These stories are connected. And at DKP, we think their intersection defines the most important capital allocation opportunity of 2026.

Bitcoin's War Test: The Numbers That Matter

Since geopolitical tensions escalated on February 28, 2026, Bitcoin has risen approximately +7% while gold fell -2%, the S&P 500 dropped -1%, and Brent crude surged +45% on supply disruption fears. BTC is currently trading around $72,300–$73,900.

More telling than the price action is the source of demand. U.S. spot Bitcoin ETFs recorded $1.1 billion in net inflows during the initial escalation window (March 2–4) and an additional $750 million in a single day on March 13 as volatility spiked. These are not retail panic-buyers — this is institutional capital making deliberate portfolio decisions.

What makes this historically significant is that it represents the first major geopolitical stress test of Bitcoin's maturing institutional base. In past crises, BTC and gold sold off in tandem. This time, Bitcoin held its gains while gold dropped as much as 10% in a single session (with gold volatility spiking to 44%, its highest since 2008).

The nuance matters: Bitcoin is not yet a pure safe haven. During the sharpest risk-off moments, it dropped alongside equities. But the recovery speed and institutional bid represent a structural shift that was not present in 2022 or 2023. BTC outperformed gold by 9 percentage points — the first time it has done so during a major geopolitical crisis.

Quant Signals in the Current Regime

For quantitative traders, the current environment has distinct characteristics that favour specific strategy types:

Volatility regime: BTC implied volatility at 39%, gold at 44%, VIX below 25 after a spike. We are in a high-volatility-high-dispersion environment — historically excellent for statistical arbitrage and mean-reversion strategies, but punishing for naive momentum models.

Funding rates: Perpetual futures funding rates have been elevated through the geopolitical premium period, creating attractive carry opportunities for delta-neutral funding arbitrage desks. Positions capturing this spread have seen annualised returns in the 18–28% range during the past three weeks.

Cross-exchange dispersion: Geopolitical-driven volume spikes create brief but persistent price discrepancies across exchanges. Our models have identified execution windows of 180–400ms on major pairs — wider than typical, providing better risk-adjusted capture rates.

Macro regime overlay: With no Fed rate cut expected (99% probability of pause at the March 18–19 FOMC), and PPI printing at +0.7% versus +0.3% forecast, the macro backdrop argues for reduced net long exposure in risk assets while maintaining crypto-specific positioning where geopolitical premium remains a real driver.

The Agentic AI Inflection Point

While markets focused on Iran, something equally significant happened at NVIDIA's GTC 2026 conference in mid-March: KX — the analytics platform used by the world's largest banks and hedge funds — launched two production-ready agentic AI blueprints for capital markets.

The first: an AI Research Assistant that compresses earnings analysis and filing review from hours to minutes, already deployed at RBC Capital Markets. The second: an AI Trading Signal Agent that discovers, validates, and monitors trading signals in real time with sub-second responsiveness — with governed, auditable outputs designed to meet institutional risk requirements.

This matters because it represents the transition from AI as a tool to AI as an actor. Agentic systems don't wait for instructions. They monitor, decide, and execute — within defined parameters — continuously. McKinsey estimates early deployments are generating 3–5% productivity gains, with projections scaling to 10%+ enterprise-wide. BCG puts the total addressable opportunity in AI-driven financial services at $200 billion.

For context: 42% of enterprises are now running agentic AI in production environments, with capital markets leading adoption due to the data intensity and latency requirements of the sector. The "alpha paradox" — the collapsing signal-to-noise ratio as traditional quant signals get arbitraged away — is precisely what Temporal AI and agentic trading agents are designed to solve.

What This Means for Capital Allocation in Q2 2026

At DKP, we are adjusting our framework across three dimensions in response to current conditions:

1. Crypto allocation: geopolitical premium is real, but time-limited. We are maintaining BTC exposure in the $68K–$76K range, treating $70K–$71K as a structurally important support level. If FOMC remains dovish, a move toward $75K–$80K resistance is plausible. We are not adding leverage here — the geopolitical premium compresses rapidly when crisis news normalises.

2. Quant strategies: favour dispersion over direction. In the current regime, statistical arbitrage, funding rate carry, and cross-exchange arbitrage outperform directional momentum strategies. We are allocating incrementally to these non-directional return streams while reducing gross directional exposure.

3. AI infrastructure: the compounding edge is being built now. The institutions deploying agentic AI systems today will operate with structural advantages that compound over years. For private capital, the opportunity is to partner with — or invest alongside — AI-native strategies before this capability becomes commoditised. At DKP, we are deploying proprietary agentic research and signal validation tools into our own process. The compounding begins here.

The Convergence Thesis

The long-term thesis we are building toward is simple: crypto markets are maturing into institutionally-viable asset classes at precisely the moment that AI-native trading infrastructure is becoming production-ready. The convergence of these two trends creates a multi-year alpha generation window for managers who can operate competently at both layers.

Prediction markets already anticipate the crypto ecosystem growing to a $100 trillion total addressable market over the next decade — not primarily through price appreciation of existing assets, but through the absorption of global payments, financial services, and capital markets infrastructure onto programmable settlement rails.

We believe this is directionally correct, and we are positioning accordingly — with the discipline, risk controls, and AI-augmented infrastructure to navigate the volatility that lies between here and there.