AI is a Decadal Opportunity – Identifying opportunities across the value chain
Artificial Intelligence (AI) is rapidly emerging as the defining technology of this decade. Just as electricity powered the industrial age and the internet transformed the digital era, Artificial Intelligence is set to reshape the global economy over the coming years.
We believe it represents a multi-year, transformative investment theme. Opportunities will span the entire value chain — from semiconductors and infrastructure to enterprise applications and security.
At WealthSpring, our objective is to help clients capture this opportunity across the entire AI value chain — not only in headline AI companies, but also in the critical enablers and adjacent technologies that will sustain long-term growth.
The AI value chain can be understood through five interconnected layers:
- Semiconductors (GPUs, high-bandwidth memory, networking)
- Infrastructure (hyperscalers, custom silicon, and advanced cooling systems)
- Foundation Models
- Agentic AI Frameworks
- Applications
Cutting across all layers are essential guardrails—security, governance, compliance, and privacy—that will determine the pace and sustainability of adoption.
Semiconductors – The Core Enabler
Annual investment in data centers is projected to grow from a few hundred billion dollars in the early 2020s to nearly $1 trillion annually by the late 2020s. This insatiable demand for compute is driven by three main factors:
- Model training has moved from self-learning earlier to reinforced/guided learning now, which is 100x more compute intensive.
- Addition of reasoning and multi-step problem solving at inference is once again 100x more compute intensive.
- Physical AI – while LLMs embody the laws of language (statistical patterns of words), Physical AI will embody the laws of physics, enabling robots, drones, and autonomous systems to perceive, plan, and act in the real world.
The semiconductor segment is at the heart of the AI revolution, powering compute, memory, and interconnect. Value creation extends beyond GPU leaders such as NVIDIA and AMD, to other recognized players driving AI silicon innovation, including Google (TPUs) and Broadcom. The broader ecosystem of critical enablers includes:
- Foundries – TSMC, Samsung Foundry, GlobalFoundries, SMIC, UMC
- Memory (HBM) – SK Hynix, Micron
- Networking – Arista Networks, Cisco, Broadcom
- EDA (Electronic Design Automation) – Cadence, Synopsys
- Equipment Providers – ASML, Applied Materials, Lam Research, KLA
Infrastructure – Scaling the AI Backbone
The infrastructure layer underpins the deployment of large-scale AI, ensuring models can be trained and run efficiently at scale. This spans hyperscalers, custom silicon, power, and thermal management:
- Hyperscalers: AWS, Microsoft Azure, Google Cloud, Alibaba Cloud — providing global compute infrastructure, data storage, and AI-specific services.
- Custom Silicon: Proprietary accelerators designed by hyperscalers to optimize AI workloads (e.g., Google TPUs, AWS Trainium, Microsoft Maia).
- Cooling Systems: Vertiv (VRT) and other liquid/immersion cooling specialists enabling high-density data center operations.
- Power: Small Modular Reactors (SMRs, e.g., NuScale) and advanced grid solutions to meet the exponential energy requirements of AI data centers.
- Networking & Interconnects: Arista, Cisco, and Broadcom driving high-bandwidth, low-latency data center fabrics.
The infrastructure layer represents the intersection of massive capital intensity and durable competitive advantage. Players that solve the challenges of power availability, advanced cooling, and ultra-low latency will capture disproportionate value as annual data center investments approach the trillion-dollar mark.
Data Engineering & MLOps
Foundation models are built on data. The entire ecosystem of data sourcing, ingestion, cleaning, labeling, and management is a critical investment area.
- Data players include Snowflake, Databricks, Scale AI,
- Vector database providers like Pinecone and Chroma.
Foundation Models – The Intelligence Layer
Foundation models are the central intelligence layer of the AI stack. These large-scale models — trained on massive datasets spanning text, images, audio, and code — form the basis for a broad range of applications. They represent the shift from narrow, task-specific systems to general-purpose intelligence platforms.
- Leading Providers: OpenAI, Anthropic, Google DeepMind, Meta, Mistral, Cohere.
- Regional & Vertical Players: Baidu (ERNIE), Huawei, AI21 Labs, Stability AI — developing models optimized for local languages or specialized domains.
- Capabilities Expansion: Moving beyond text into multimodal systems (text, vision, audio, video), enabling richer and more adaptive interfaces.
- Economic Implications: Foundation models are becoming critical infrastructure — powering applications across healthcare, finance, manufacturing, and defense.
Perspective: Leaders that achieve scale, data quality, and fine-tuning will not just shape competition but also capture platform-like economics across industries
Agentic AI Frameworks – Automating Workflows
Agentic AI represents the next frontier, moving beyond simple chat interfaces toward automation of complex workflows, decision-making, and orchestration across multiple systems. Key emerging providers include LangChain, LlamaIndex, AutoGPT, CrewAI, and OpenAI’s ecosystem tools, alongside enterprise orchestration layers such as Cognosys and Dust.
Applications Layer – Core Segments
- AI-Native Application Innovators Startups and platforms purpose-built for AI, enabling new categories of applications in customer support, marketing, productivity, and enterprise decision-making.
Examples include – Jasper (AI marketing), Copy.ai (content generation), Character.ai (conversational AI), GitHub Copilot (developer productivity), Palantir (enterprise AI platform for decision intelligence, workflow automation, and secure data integration — widely adopted in defense, government, and regulated industries)
- Enterprise ERP/CRM Incumbents Embedding AI Large-scale enterprise software platforms embedding AI and evolving into AI-enabled workflow automation hubs with security and compliance built in:
- CRM: Salesforce (Einstein AI), HubSpot, Zoho
- ERP / HCM: SAP (Joule AI), Oracle, Workday
- Service Management: ServiceNow (Now Assist)
By embedding AI with workflow automation, auditability, and governance, these platforms will drive mainstream enterprise adoption, while innovators like Palantir showcase how AI-native platforms can orchestrate secure, mission-critical workflows at scale.
- Edge AI: On-device/near-device inference for low-latency, privacy-sensitive, or offline use cases. Key domains include smartphones, wearables, industrial IoT, vehicles, and camera systems. This segment enables AI adoption in environments where connectivity, latency, or privacy make cloud-only deployment impractical.
These segments will ultimately determine how AI translates into enterprise productivity, user adoption, and monetization — the point at which investors see scaled returns.
System Integrators (SIs) for AI Use Cases
Global and regional systems integrators are playing a critical role in embedding AI into business workflows—tailoring models, frameworks, and infrastructure to industry-specific processes. Key players include Accenture, TCS, Infosys, Wipro, Deloitte, and Capgemini, alongside specialized AI-native integrators.
SIs will therefore be essential partners for scaling AI deployments, making them indirect beneficiaries of every wave of enterprise adoption.
Security – The Essential Guardrail
AI adoption cannot scale without robust security, governance, and compliance embedded across the stack. Key beneficiaries include:
- Cybersecurity Leaders: Palo Alto Networks, CrowdStrike, Zscaler, Fortinet
- Cloud-Native Security Platforms
- AI-Specific Security Startups
Security spend will scale in lockstep with AI adoption, making this one of the most durable and non-cyclical investment opportunities in the ecosystem.
Our Perspective
We view AI not as a standalone sector but as the nucleus of a broader technological super-cycle, with transformative effects across every industry. Our perspective recognizes that value creation will come not only from market leaders in AI infrastructure and applications, but also from the less obvious enablers—semiconductor innovators, workflow automation platforms, cybersecurity layers, and adjacent fields like blockchain and biotechnology. By taking this systemic view, we help investors capture both the headline opportunities and the hidden multipliers that drive long-term competitive advantage.
