AI capex explodes, fusion steps up, quantum gets tested, carbon rules get teeth

Alphabet guided to sharply higher 2026 capex, potentially around 2× 2025 levels, mostly for AI infrastructure build-out.

Spending focuses on hyperscale data centers, high-speed networking and AI accelerators to support Gemini, cloud and AI-enhanced search.

Google Cloud revenue growth is strong enough that management is comfortable with a multi-year AI build cycle.

AI is now an infrastructure build, not just a software story. The bottlenecks are electricity, land, cooling and fiber.

A handful of hyperscalers are driving a $500B-plus annual AI capex cycle.

Markets will increasingly judge AI players on return on capex (ROIC vs. spend), not just headline model launches.

Think in stack terms: chips, power equipment, substations, land, grid connections—not just model providers.

Map regulatory and local pushback risk around water use, emissions and zoning for large campuses.

Stress-test theses for utilization risk: can real workloads keep up with the infrastructure curve?

Wolfpack says a large share of IonQ’s 2022–24 revenue came from defense earmarks that have now been cut or restructured.

The report also questions IonQ’s treatment of bookings and its use of acquisitions to add non-core revenue streams.

IonQ maintains it has been transparent and argues the acquisitions support a long-term quantum strategy.

Shows how grant and defense-driven revenues can be confused with scalable commercial traction.

Highlights the need to break out bookings vs. funded backlog vs. recognized revenue in frontier tech.

Underlines that quantum computing is still pre-product-market-fit; stories can run ahead of hardware reality.

Separate R&D grants and pilots from recurring, usage-based enterprise spend when evaluating progress.

Treat complex acquisition roll-ups as a question: capability-building or just buying revenue?

For quantum overall, anchor on technical roadmaps, error correction progress and target use-cases, not just TAM slides

DOE and external reviewers approved Thea’s preconceptual design for its Helios fusion pilot plant.

The review is part of a staged DOE milestone program that pays companies as they hit specific technical and design gates.

Thea’s roadmap runs from Helios (pilot) toward a future Eos commercial plant in the 2030s.

Shifts fusion from “lab experiment” to early project-finance candidate with structured milestones.

Stellarator-based designs broaden technology diversity beyond tokamaks and compact concepts.

Milestone-based public funding offers a template for blended capital stacks in high-risk energy tech.

Fusion exposure remains binary and long-dated; instrument choice and position sizing are critical.

Near-term opportunity may be stronger in enabling supply chains: magnets, power electronics, precision components, digital twins.

Track which platforms develop credible offtake structures and regulatory paths, not just physics milestones.

India earmarked ~$2.2B for CCUS in power, steel, cement, refining and chemicals.

The EU approved its first CRCF methodologies for direct air capture with storage, BioCCS and biochar.

CRCF sets a formal EU standard for how durable carbon removals are measured, verified and certified.

India’s budget points toward funded demonstration plants in hard-to-abate sectors.

CRCF methodologies raise the bar for measurement, permanence and additionality in long-lived removals.

The combination of public money + high-integrity standards will influence other major jurisdictions.

In sectors like cement and steel, CCUS is becoming part of the license to operate, not optional upside.

CRCF may evolve into a reference standard for premium, durable removal credits.

Risk-return may be more balanced in infrastructure and tools—transport and storage networks, MRV systems, project developers—than in single-asset equity.