Macro intelligence

Trends

Every 5 days we zoom out: which themes are attracting capital, which have catalysts, which have turned into FOMO. Research themes are free here. Early-curve actionable ideas sit behind Pro.

Market regime

Where are we now?

The complete absence of these themes from X/Twitter discourse despite presumably stable macro conditions suggests capital remains concentrated in established narratives (likely AI infrastructure mega-caps and rate-sensitive trades). This creates unusual opportunity for early positioning in overlooked verticals before institutional rotation begins.

Updated August 5, 2026

Summary

August 2026's opportunity set reveals a striking pattern: every single validated theme sits in pre-formation narrative stage with zero FinTwit momentum, suggesting we're capturing genuine pre-consensus alpha rather than chasing existing rotations. The complete absence of Perplexity validation data across all 20 themes is concerning and suggests either a technical failure or that these themes are so early-stage they lack web presence. Given the uniform Grok validation showing 'pure_signal' status, we're either looking at an extraordinary confluence of undiscovered opportunities or a systematic blind spot in current market attention. Sector concentration is extreme with defense_tech and ai_compute dominating actionable themes—a quality concern that suggests Stage 1 generation may have been too narrow.

🔒

2 actionable ideas are behind Pro

Our strongest early-curve themes come in the Pro newsletter — including specific instruments and buy zones.

Premium · coming soon

Research Timing: early

Directed energy weapons for drone defense entering mass production

Laser weapon systems are reaching the cost-per-shot advantage needed to defeat cheap drone swarms economically, with BlueHalo and Epirus positioned as pure-plays. However, no sub-$5B publicly traded pure-play exists, limiting actionable exposure to private markets or large-cap defense contractors.

Why now

+Cost per laser shot dropping below $10 vs $100K+ for missile interceptors
+BlueHalo's LOCUST system achieving 60-second continuous fire capability
+Israel's Iron Beam entering operational deployment as proof of concept

Risks

−Atmospheric conditions limiting laser effectiveness in real combat
−Power requirements proving prohibitive for mobile deployment
−Kinetic interceptors remaining preferred for high-value target defense

Research Timing: early

Synthetic data generation platforms for LLM training hitting cost parity with human-labeled data

Synthetic data costs are approaching the $0.001/sample threshold where they become cheaper than human labeling at $0.10+/sample, potentially disrupting the entire AI training pipeline. Pure-plays like Mostly AI and Gretel.ai remain private, limiting public market exposure.

Why now

+GPT-5 class models requiring 10 trillion+ tokens driving synthetic data demand
+Human labeling costs rising 30% YoY due to quality requirements
+Mostly AI announcing $0.0015/sample pricing for structured data in July 2026

Risks

−Model collapse from recursive synthetic training proving insurmountable
−Regulatory scrutiny on synthetic data quality for regulated industries
−Open-source solutions commoditizing synthetic generation

Watch Timing: early

Direct lithium extraction from geothermal brines in California's Salton Sea

DLE technology at Salton Sea could provide 40% of US lithium demand, but technical challenges and permitting delays keep pushing commercial timeline. Controlled Thermal Resources leads but remains private.

Why now

+GM's $650M investment in Controlled Thermal Resources validating DLE potential
+Lithium prices stabilizing at $15,000/ton making DLE economics viable
+California fast-tracking permits for Lithium Valley initiative

Risks

−Technical challenges with brine composition delaying commercial viability
−Environmental concerns about Salton Sea ecosystem impacts
−Lithium price collapse making DLE uneconomic vs traditional extraction

Watch Timing: early

Edge AI inference chips achieving 1 TOPS per dollar

Edge inference economics reaching smartphone integration threshold, but Hailo and Mythic remain private while established semis dominate the narrative. The 1 TOPS/dollar metric enables AI processing in sub-$200 devices.

Why now

+Hailo-8 achieving 26 TOPS at $25 chip cost crossing the threshold
+Smartphone makers mandating on-device AI for privacy and latency
+Edge inference market projected at $45B by 2028 per IDC

Risks

−Cloud inference remaining dominant due to model size requirements
−Apple/Google building proprietary chips cutting out third parties
−Power efficiency not improving enough for always-on edge AI

Past briefings

August 30, 2026

12 themes · Tactical satellite ISR proliferation — Space Force TacSRT contracts flowing to commercial small-sats

Where are we now?

Summary

2 actionable ideas are behind Pro

Directed energy weapons for drone defense entering mass production

Synthetic data generation platforms for LLM training hitting cost parity with human-labeled data

Direct lithium extraction from geothermal brines in California's Salton Sea

Edge AI inference chips achieving 1 TOPS per dollar

Past briefings

Before you go —get our 10 best stocks, free.

Before you go —
get our 10 best stocks, free.