Investors should build long-term positions in nuclear energy (specifically Small Modular Reactors) and natural gas infrastructure to solve data center power constraints ahead of the critical 2030 to 2032 bottleneck.

Reduce supply-chain risk by investing in domestic foundries and alternative specialized silicon accelerators to diversify away from single dependencies like Taiwan Semiconductor Manufacturing Company (TSMC).

Target computational biology and pharma AI platforms such as Chai Discovery, which are actively generating multi-million-dollar enterprise revenues by accelerating pharmaceutical drug discovery.

Prepare for near-term commercial breakthroughs in physical AI and semi-humanoid robotics by tracking innovators like Sunday Robotics ahead of consumer beta launches scheduled for late this year into early next year.

Allocate toward high-conviction vertical enterprise AI software, led by platforms like Harvey, that demonstrate strong enterprise pricing power by automating complex workflows in legal, accounting, and compliance sectors.

AI Energy & Compute Infrastructure

• Energy and compute capacity are identified as the primary physical bottlenecks limiting the long-term growth of artificial intelligence.
  • Hyperscale data center operators do not expect sufficient power solutions to move the needle at scale before 2030 to 2032.
  • Essential energy inputs include natural gas as transition power and nuclear energy (specifically Small Modular Reactors or SMRs) to provide baseload clean energy.
  • Regulatory approval, public alignment, and physical supply chain constraints (labor, materials, and specialized components) remain major risk factors slowing deployment.
  • The concept of compute independence is emerging as a critical national security and economic priority, requiring diversification from single-point dependencies such as Taiwan Semiconductor Manufacturing Company (TSMC).

Takeaways

• Long-term investors should monitor energy generation assets—particularly nuclear power developers, natural gas infrastructure, and alternative grid solutions—that can deliver dedicated power to hyperscale data centers.
• Supply chain localization and domestic semiconductor manufacturing capacity are positioned to receive ongoing private capital and policy support.

Chai Discovery (Computational Biology & Pharma AI)

• AI models applied to biology are shifting from speculative research into revenue-generating software platforms.
  • The traditional biotechnology software model historically struggled to capture venture returns without owning drug assets.
  • Chai Discovery has demonstrated commercial traction by securing large enterprise contracts (such as $10 million deals) with top-10 global pharmaceutical companies to accelerate the R&D and drug discovery process.
  • While physical lab validation and regulatory approvals remain inherent bottlenecks, model-driven biological research is expected to accelerate the timeline for medical discoveries.

Takeaways

• Computational biology and AI-driven life sciences software are transitioning into viable high-growth enterprise markets, driven by immediate efficiency gains in pharmaceutical research and development.

Sunday Robotics (Physical AI & Semi-Humanoid Robotics)

• Breakthroughs in data collection methods and modern AI models are rapidly solving generalization and robustness challenges in physical robotics.
  • Founded by Stanford researchers Tony Zhao and Chang Chi, Sunday Robotics focuses on cost-effective data collection integrated with full-stack hardware and software manufacturing.
  • Development cycles have compressed significantly, with consumer beta testing for general, semi-humanoid domestic robots targeted for the end of this year or early next year.

Takeaways

• Physical robotics is moving from an open-ended research phase to near-term commercial application, offering new exposure opportunities in real-world automation and consumer hardware.

Harvey (Vertical Enterprise AI & Legal Tech)

• Large language models are well-suited for document-heavy, text-structured professional fields such as legal services.
  • Harvey applies AI retrieval and generation to legal precedents, contract reviews, and complex workflows like mergers and acquisitions (M&A) due diligence.
  • The software moves beyond basic administrative tasks toward automating substantial portions of complex corporate legal work.

Takeaways

• Vertical AI applications targeting highly structured, language-intensive professions (law, accounting, compliance) continue to exhibit strong enterprise willingness to pay and rapid product-market fit.

Alternative Chip Architectures & Semiconductor Hardware

• Venture capital dynamics around semiconductor and hardware investments have fundamentally improved compared to historical downturns.
  • Extreme buyer demand for AI compute accelerators and supply chain redundancy is reducing commercialization risk for new chip designers.
  • Hyperscalers and large enterprises are actively seeking alternatives to standard GPU architectures and single-source foundry pipelines to mitigate supply shortages.

Takeaways

• Demand for customized compute, specialized silicon accelerators, and power-efficient chip architectures provides a favorable backdrop for hardware innovators addressing the global compute shortage.