Nvidia and ServiceNow: Infrastructure Rails for the AI S-Curve

The AI boom is no longer a speculative narrative. It is evolving into a durable, measurable global investment cycle, and that shift defines the core investment thesis. Research firm Gartner expects worldwide AI spending to grow 44% year over year to $2.5 trillion in 2026. This isn't sentiment-driven capital chasing a trend; it's productive enterprise and government capex tied to real workloads. The distinction is critical. Unlike cryptocurrencies, where demand can surge and fade with changing regulation or risk appetite, AI capital expenditure is increasingly anchored to deploying models that generate tangible business value.

This creates a paradigm shift. AI demand is moving from infrequent, capital-intensive training to recurring, operational inference-deploying models in real products and workflows. This makes the demand stickier and more predictable once systems are built. The result is a multi-year infrastructure build-out, not a fleeting asset class. In this new paradigm, companies are not just software vendors; they are fundamental rails.

Nvidia is the dominant compute layer for this new infrastructure. Its AI-optimized chips, networking, and software solutions are the essential fuel. The consensus estimate for hyperscaler capital spending on AI initiatives is projected to reach $527 billion in 2026. Nvidia's revenue visibility for its next-generation systems extends beyond $500 billion through 2026, a testament to the scale and longevity of this cycle. The company is building the fundamental hardware platform for the AI S-curve.

ServiceNow represents a key software platform layer for enterprise AI operations. Its cloud-based workflow automation is being rapidly augmented by AI, creating a durable revenue stream. The company's AI product suite, NOW Assist, surpassed $600 million in annual contract value last quarter and is expected to reach over $1 billion in 2026. With a 98% customer renewal rate and a contracted backlog of $28.2 billion, ServiceNow offers the recurring cash flows and visibility that speculative assets lack. It is building the operational software layer that runs on the AI infrastructure rails.

Together, Nvidia and ServiceNow exemplify the shift from speculative to sustainable. They are not betting on a price bubble; they are building the foundational layers of the next technological paradigm.

Nvidia: Scaling the Compute S-Curve

Nvidia is not just riding the AI S-curve; it is engineering the track. The company's latest quarterly report shows the infrastructure build-out is accelerating, not plateauing. Revenue hit a record $57.0 billion for the third quarter of fiscal 2026, a 62% jump from the year before. The data center segment, which powers the AI boom, grew even faster, up 66% year-over-year to $51.2 billion. This isn't just growth; it's exponential scaling, with Jensen Huang noting that compute demand is "accelerating and compounding across training and inference."

The key to sustaining this pace is lowering the cost of operation. Nvidia's new Rubin platform is designed for that exact purpose. By harnessing extreme codesign across its six new chips, Rubin targets a 10x reduction in inference token cost compared to its predecessor. This is a critical move. As AI shifts from rare training to constant inference, the cost per operation becomes the bottleneck for mainstream adoption. Rubin's ability to slash that cost accelerates deployment across industries and locks customers into a platform where the economics of running AI models become irresistible.

Financially, Nvidia is in a position to fund the next leap. The company returned $37.0 billion to shareholders in the first nine months of fiscal 2026 through buybacks and dividends. With $62.2 billion remaining under its repurchase authorization, it maintains immense firepower. This capital strength is essential for the next S-curve transition, where the company must continue investing in R&D and capacity to stay ahead of the exponential demand curve. The Rubin launch, with its focus on agentic AI and reasoning, is the next chapter in that build-out, ensuring Nvidia remains the fundamental compute layer for the paradigm shift.

ServiceNow: The Enterprise AI Workflow Layer

ServiceNow is operationalizing the AI paradigm shift within the enterprise, moving from workflow automation to an AI control tower. The company's latest quarter showed the durable growth engine is accelerating. Subscription revenue grew 21% year-over-year to $3.47 billion, with total revenue up 20.5%. More importantly, its contracted backlog is expanding at a robust pace. Current remaining performance obligations, which represent revenue to be recognized in the next 12 months, grew 25% year-over-year to $12.85 billion. This is the financial bedrock of the S-curve: a growing, predictable stream of future cash flows that signals deep enterprise adoption.

The company is actively scaling its AI product suite, Now Assist, which is becoming the core of its agentic workflow strategy. The platform's net new annual contract value more than doubled last quarter. ServiceNow is embedding this capability into critical business operations, as seen in its new partnership with Fiserv. The two companies are committing to transform commerce and financial services by embedding AI directly into operational workflows. This is a concrete example of how the software layer is being used to run the new paradigm, moving AI from experimentation to production.

Financially, ServiceNow is in a strong position to fund this expansion. The board recently authorized an additional $5 billion under its share repurchase program. This move, coupled with a planned $2 billion accelerated buyback, signals high confidence in the company's durable growth and cash generation. It's a vote of confidence from management that the current trajectory of high organic growth and margin expansion is sustainable. For investors, ServiceNow represents a play on the sticky, high-margin software layer that will manage the complex orchestration of AI agents across the enterprise.

Catalysts, Risks, and What to Watch

The infrastructure thesis for Nvidia and ServiceNow is now in the execution phase. The forward view hinges on a few key catalysts, significant risks, and specific metrics that will validate the sustainability of the AI S-curve.

The primary catalyst is the commercial rollout of Nvidia's Rubin platform. This is not a speculative product launch but a fundamental infrastructure upgrade designed to accelerate mainstream AI adoption. The platform's 10x reduction in inference token cost directly addresses the economic bottleneck for widespread enterprise deployment. Its scaling across hyperscaler data centers and partnerships with companies like Microsoft and CoreWeave will be the real-world test. For ServiceNow, the catalyst is the scaling of enterprise AI adoption beyond early adopters. The company's Now Assist net new annual contract value more than doubled last quarter, but the next phase is embedding this capability across industries. The partnership with Fiserv to transform commerce is a blueprint; broader replication will prove the model's durability.

The main risk to the thesis is a slowdown in hyperscaler capital spending. While the consensus estimate for 2026 is a massive $527 billion, this cycle is still maturing. Any deceleration in this capex would directly pressure Nvidia's revenue visibility and could ripple through its ecosystem. For ServiceNow, the risk is competitive erosion in the enterprise workflow layer. As AI becomes table stakes, rivals may pressure its 98% customer renewal rate and contracted backlog growth. The company's high-margin, sticky software model is a moat, but it is not impervious.

Investors should watch three specific metrics to gauge the sustainability of the S-curve. First, quarterly growth rates for both companies must remain robust. Nvidia's 62% year-over-year revenue growth is a high bar; any deceleration would signal market saturation or competitive pressure. ServiceNow's 21% subscription revenue growth and expanding current remaining performance obligations are the indicators of its workflow layer's strength. Second, guidance updates from both firms will reveal their confidence in the multi-year cycle. Third, and most importantly, any shifts in the AI spending cycle itself-whether from enterprise budgeting or hyperscaler planning-will be the leading indicator of the paradigm's health. The bottom line is that the thesis is built on exponential adoption. The coming quarters will show if the rails are being laid fast enough to keep pace.