Artificial intelligence is no longer a single theme. It has become a broad technology ecosystem that touches more of the global economy each year. That ecosystem spans graphics processing units (GPUs), high-bandwidth memory, cloud infrastructure, data pipelines, observability platforms, enterprise software and security layers.
Each part of the AI stack is scaling at a different pace. Each has its own constraints. And each contributes differently to the companies positioned to benefit from AI’s growth.
The November 2025 rebalance of the
WisdomTree Artificial Intelligence and Innovation Index—which the
WisdomTree Artificial Intelligence and Innovation Fund (WTAI) seeks to track—reflects this reality.1
As we approach the end of 2025, the rebalance can be best understood through five core pillars shaping today’s AI landscape.
Pillar 1: Compute Scale Remains the Dominant Bottleneck
If AI had a unifying message in 2024–2025, it would be this: compute is scarce, and the shortage is real. Companies building data centers today are not speculating on future demand—they are responding to visible, overwhelming customer needs.2
As a result, the largest weights in the strategy remain concentrated in the compute layer.
NVIDIA: Conviction That Remains
NVIDIA continues to be one of the largest exposures, with a weight of 4.50%. The rationale is straightforward:
- AI compute demand continues to outpace supply.
- NVIDIA maintains unmatched ecosystem depth, spanning CUDA,3 networking, inference optimization and full-stack system design.
- Early Blackwell demand signals suggest GPU scarcity will remain a pricing and margin tailwind.
This is not a case of “sticking with what worked.” Rather, each new data point reinforces NVIDIA’s ability to extend and compound its competitive advantage.
Hyperscaler Infrastructure: Alphabet, Amazon and Oracle
Training and deploying large-scale AI models requires AI-optimized cloud infrastructure, inference endpoints and enterprise-grade deployment capabilities.
That is why the following hyperscalers remain among the top weights:
- Alphabet (GOOGL): 5.00%
- Amazon (AMZN): 4.00%
- Oracle (ORCL): 4.00%
Each provides leverage to a different dimension of AI infrastructure:
- Google: large-scale training and inference, continued Gemini model improvements, AI-native distribution and proprietary tensor processing units (TPUs).
- Amazon: AWS’s foundational role in enterprise compute, model-hosting economics and an accelerating mix of AI workloads.
- Oracle: rapid AI infrastructure deployments, emerging operating leverage and a growing revenue pipeline tied to model hosting.
Oracle has faced increased market scrutiny in late 2025 around debt financing and customer profitability timelines. For informed AI investors, this is unsurprising. The real question is whether model scaling continues. If AI becomes merely “slightly better software,” the thesis weakens. If scaling remains intact—as recent Gemini results suggest—it remains early days for hyperscaler infrastructure.4
Pillar 2: Memory and Networking Matter More Than Ever
As models grow larger and more complex, GPUs are no longer the sole constraint. Memory and networking have become structural bottlenecks.
Modern AI workloads depend on:
- DRAM and HBM: ultra-fast working memory that feeds GPUs in real time.
- NAND: persistent storage for training data, checkpoints, vector databases and model files.
AI cannot scale without both. Each new NVIDIA architecture requires more memory per chip, not less.5
Supply constraints in HBM and NAND are now delaying system deployments across the industry.
Key suppliers include SK Hynix (3.58%), Micron (2.66%) and Samsung (1.50%).
Networking: The Quiet Enabler of Scale
Training large models requires thousands of GPUs to communicate as a single system. That coordination extends across servers, racks and entire data centers, demanding ultra-low latency and massive bandwidth.
This infrastructure is delivered through high-performance switches, optical interconnects and specialized cabling—areas led by companies such as Broadcom (3.25%), Marvell (2.00%), Credo (1.00%) and Astera Labs (1.14%). As AI scales, so does demand for these technologies.
Figure 1 illustrates how these infrastructure pillars are reflected in the top 10 index weights.
One notable inclusion is Meta Platforms. While Meta uses AI primarily to enhance existing businesses—advertising and content recommendation—we believe investors may underappreciate how powerful this application is. These remain some of the most profitable digital business models, and differentiated AI capabilities continue to strengthen them.
Pillar 3: Enterprise AI Adoption Is Now Real
For years, enterprise AI was dominated by demos and limited productivity gains. In 2025, that changed. Enterprises are now seeing measurable revenue uplift from AI-enabled workflows, copilots and data-layer integration.
ServiceNow (2.75%)
ServiceNow’s Now Assist platform is transforming enterprise workflows into automated, AI-driven systems, delivering tangible productivity and cost benefits.6
Atlassian (1.50%, Addition)
Developer workflows are rapidly integrating AI copilots, debugging tools and code recommendations. As AI generates a growing share of new code,7 Atlassian’s position within development pipelines gives it early leverage to AI-driven productivity.
Shopify (0.50%, Addition)
AI is reshaping merchandising, logistics and storefront optimization. Shopify sits at the center of this shift, with a data flywheel that enables AI-driven insights for merchants as consumer search and purchasing behavior evolves.
Snowflake (2.00%)
Snowflake remains foundational to enterprise data pipelines, vector databases and AI data layers. Its native LLM integrations position it well as adoption accelerates.
Salesforce (1.00%)
Agentforce and Data Cloud embed autonomous agents directly into existing enterprise workflows, aligning AI adoption with systems companies already rely on.
Enterprise AI is no longer experimental—it is a budget line item. The rebalance reflects where that spending is flowing.
Pillar 4: Cybersecurity Is a First-Order AI Theme
As AI scales, the cybersecurity threat landscape expands. AI increases attack surfaces, accelerates threats and adds complexity to cloud security.
CrowdStrike (2.00%)
CrowdStrike’s AI-native threat detection and strong net retention make it a critical security platform in the AI era.8
Datadog (1.50%)
AI workloads increase system complexity and observability demands. Datadog is essential for operating AI systems reliably at scale.
Palo Alto Networks (2.50%)
Palo Alto integrates cloud security, networking visibility and AI governance, making it increasingly relevant as enterprise AI deployments grow.
AI adoption and cybersecurity demand rise together. This is a structural relationship, not a cyclical one.
Pillar 5: Avoiding the Noise
The final pillar reflects discipline. Some companies invest heavily in AI but lack direct revenue linkage, face intensifying competition or trade well ahead of fundamentals. The strategy adjusts exposure accordingly.
Tesla is the most notable removal. While deeply involved in AI, its valuation continues to embed a significant narrative premium relative to near-term fundamentals.
The guiding philosophy is simple: WTAI increasingly focuses on measurable AI adoption, infrastructure investment and revenue leverage—rather than speculative future narratives.
Closing Thought: AI Is Spreading, but Conviction Is Sharpening
AI is no longer a theme—it is a system. And systems reward precision.
The November 2025 rebalance reflects that mindset. It focuses on:
- where capital must flow,
- which companies have real leverage to that spend, and
- maintaining discipline around valuation, durability and adoption.
AI is entering a more operational, infrastructural and enterprise-driven phase. The current expression of WTAI is designed to meet that moment—with clarity, conviction and five pillars at its core.
Enjoyed this article? Sign up for our newsletter to receive regular insights and stay connected.