Nvidia stock soared over the last few years but has cooled off lately as the market wonders if competitors can catch up. The company became a powerhouse by selling the essential chips needed to run modern artificial intelligence, creating a huge lead that is hard for others to copy. Now, rivals are trying to build their own alternatives to chip away at that success.
What does it do?
Nvidia is a hypergrowth business that earns money by designing and selling high-performance computing platforms that combine specialized chips with a deep layer of software. Money flows primarily from large cloud providers, consumer internet companies, and governments that need massive amounts of computing power to train and run artificial intelligence. Customers pay upfront for hardware like the H100 or Blackwell GPUs and increasingly pay recurring fees for software services that help them manage complex AI workloads. The company’s "CUDA" software creates a cycle where developers write code specifically for Nvidia hardware, which in turn makes that hardware more valuable to the businesses that hire those developers.
Where does revenue come from?
The vast majority of revenue comes from Data Center chips, which now account for 92% of the total business. The company recently moved to two market platforms: Data Center, which includes hyperscale cloud providers and enterprise AI factories, and Edge Computing, which covers gaming PCs, workstations, robotics, and autonomous vehicles. Geographically, revenue is highly concentrated among large global tech firms, though the company is seeing rapid growth in "Sovereign AI" as individual countries build their own domestic computing hubs.
Revenue Breakdown
Revenue by Geography
Who are its customers?
Nvidia serves a handful of massive cloud providers alongside a growing base of enterprise customers and 200 million gamers. In its most recent quarter, Data Center revenue reached $75.2 billion, primarily driven by hyperscale cloud companies and large consumer internet firms. While the company does not name its largest clients in every release, public filings show that a small number of "Big Tech" firms represent a significant portion of total demand. Beyond the giants, Nvidia is expanding into industrial and healthcare sectors, while its Edge Computing segment brought in $6.4 billion in the latest quarter from a mix of PC manufacturers and automotive partners like Hyundai and BYD.
What gives it staying power?
Nvidia's staying power comes from its CUDA software platform, which has become the universal language for AI development over the last 15 years. Because most AI models are built on CUDA, switching to a competitor's chip requires rewriting massive amounts of code, creating high switching costs.
Where is it headed?
Nvidia is headed toward becoming the foundational platform for "Agentic AI," where AI systems act as autonomous employees that can reason and execute tasks. Management is betting that this shift will require an even larger buildout of "AI factories" as companies move from simple data storage to active AI production.
Nvidia is growing at a rate rarely seen in companies of this scale, with revenue up 85% year-over-year. In the most recent quarter, revenue reached $81.6 billion, and profit grew even faster as the company maintained incredible pricing power on its high-end chips.
Cash generation is exceptional, with free cash flow reaching $96.68 billion in the last fiscal year. This massive cash engine allows the company to fund its own research and development while simultaneously returning billions to shareholders through buybacks.
The balance sheet is fortress-like, with $13.2 billion in cash and very little debt. With a debt-to-equity ratio of just 0.07x, the company has no financial constraints and can easily outspend any competitor to maintain its technical lead.
Nvidia is one of the most profitable and fastest-growing large-cap businesses in history.
The Data Center business is producing record results, with revenue jumping 92% to $75.2 billion in a single quarter. This surge is driven by the massive demand for AI training and the early rollout of the new Blackwell architecture, which is being adopted by every major cloud provider.
Customer concentration is the primary risk, as a few large tech companies account for a huge portion of the current spending. If those firms decide to slow their AI investments or successfully build their own internal chips, Nvidia's growth could decelerate quickly.
The AI hardware market is approximately $200 billion today and is growing at roughly 35% annually, putting it on a path to exceed $500 billion within the next three years. This is an exceptional industry because performance is far more important than price, giving the market leader significant pricing power. Nvidia is the undisputed leader in this market, controlling the vast majority of the high-end chips used for AI training. Its position is so dominant that it currently defines the pace of innovation for the entire technology sector.
The market for AI chips is intensely competitive but currently dominated by a single player because the barriers to entry are extremely high. Developing a competing chip takes years and billions of dollars, and even a successful chip must overcome the software barrier that locks developers into the leader's ecosystem. One sentence on what this means for long-term pricing power. High barriers to entry mean the current leader can maintain high margins as long as its technical lead remains significant.
AMD is the most direct threat, offering high-performance chips that are often more affordable, though they lack the same software ecosystem. Intel is attempting a comeback by embedding AI capabilities into every PC and server chip, while Broadcom is helping cloud giants like Google and Amazon build their own custom AI silicon. The most dangerous threat comes from "Big Tech" customers like Google and Amazon building their own internal chips to reduce their dependence on Nvidia.
Nvidia is currently holding its ground and even gaining share in the high-end market as its Blackwell chips set a new performance ceiling that rivals have yet to match. Its Data Center revenue grew 92% this year, far outstripping the growth of its nearest competitors.
The primary source of protection is a combination of specialized hardware and the CUDA software platform that makes that hardware easy for developers to use. Nvidia's moat is built on switching costs: millions of developers have spent a decade building AI tools that only run on Nvidia's code. This creates a massive hurdle for any rival chipmaker, as customers would have to rewrite their entire software stack to switch.
The company's 63% ROIC and 75% gross margins are proof that this advantage is real and durable. These are not the numbers of a company in a commodity price war: they reflect a business that has effectively decoupled its pricing from its costs. The combination of triple-digit ROE and accelerating revenue growth proves that Nvidia's lead is currently widening rather than eroding.
The moat is strengthening because the move to "Agentic AI" and complex AI factories requires even tighter integration between hardware and software. The most important signal is the record-breaking adoption rate of the new Blackwell platform before it has even fully shipped.
Beat revenue and EPS targets for 5+ consecutive quarters with triple-digit growth.
Authorized $80B in new buybacks and increased dividend by 2,400% this quarter.
Founder CEO holds a multibillion-dollar stake, representing the vast majority of his wealth.
Capital Allocation Track Record
Jen-Hsun Huang is one of the most effective visionary leaders in technology, having correctly predicted the shift to accelerated computing a decade before it happened. His strategic judgment is evident in the creation of CUDA, which transformed Nvidia from a gaming company into the most important infrastructure provider in the world. The company's ability to execute on triple-digit growth while maintaining 75% gross margins proves that management has exceptional control over its complex global supply chain.
The primary governance risk is the high degree of key-person dependence on Huang, who has been the singular driving force behind the company's strategy since its founding. While the company has a deep bench of experienced executives like CFO Colette Kress, the vision and technical direction are uniquely tied to the founder. Investors should monitor succession planning, though the current team has proven it can handle the most complex infrastructure expansion in history without major operational misses.
We expect revenue to grow from $214B in FY2026 to $1020B in FY2031 (~37% CAGR), with EPS growing from $4.69 to $25.26 (~40% CAGR). The massive build-out of AI data centers and the transition to Blackwell-based systems drive sustained hardware demand. High-margin software services and the premium pricing of specialized AI chips allow the company to maintain exceptional profitability. EPS grows faster than revenue because the company uses its significant free cash flow to aggressively repurchase shares. Operating margin expected to reach ~62% by FY2031.
Agentic AI creates a massive new wave of chip demand. If AI moves from chatbots to autonomous agents that work 24/7, the need for computing power will multiply far beyond today's training requirements.
Sovereign AI hubs expand the customer base to nations. Individual countries building domestic AI infrastructure reduces Nvidia's dependence on a few US cloud giants and opens a global market.
Software services become a high-margin recurring revenue stream. By selling software to manage AI factories, Nvidia can generate stable, recurring profits that are less cyclical than chip sales.
Large cloud customers successfully transition to internal chips. If Google, Amazon, and Meta move a large portion of their workloads to their own custom silicon, Nvidia's primary growth engine could stall.
Geopolitical restrictions further limit access to the China market. Expanded US export controls could permanently cut off a significant portion of global demand that Nvidia cannot easily replace elsewhere.
An "AI Winter" where businesses pause spending to prove ROI. If companies stop seeing immediate value from their AI investments, they may delay the next multi-billion dollar hardware upgrade cycle.
Below is our estimate of current and future fair value, with detailed reasoning and assumptions. Fair value is a judgment, not a fact, and other analysts will likely land on different numbers. Use it as one data point in your research, and apply your own discretion in any investing decision.
We use a Normalized P/E approach based on mid-cycle earnings power rather than the current hyper-growth print. It fits NVIDIA because the company is currently seeing a massive earnings and margin expansion that triggered our pre-check; using a normalized base (price divided by average cycle earnings) prevents over-extrapolating a potentially temporary infrastructure boom into a terminal valuation.
Next year's projected EPS of $8.94 multiplied by a 35x normalized multiple gives a per-share fair value of $313. This 35x multiple sits between Apple (36x) and high-growth semi peers like Broadcom (30x), which is justified by NVIDIA’s near-monopoly in AI training and its superior return on invested capital (63%). We use the deterministic engine's FY2027 EPS of $8.94 as our primary base because it aligns with the start of the Blackwell architecture revenue ramp, which represents the current mid-cycle baseline.
Cross-checked with a peer-anchored Forward P/E (FY2027 EPS $8.94 × 36x peer average), we get $322 — within 3% of our $313 answer, confirming the result. While the deterministic engine's 5-year DCF produces a much higher fair value of $546, that model captures the full platform maturity through FY2031. Our Normalized P/E approach is more conservative, focusing on the immediate value of the current cycle and accounting for the higher volatility inherent in semiconductor infrastructure build-outs compared to pure software businesses.
We're assuming a 35x normalized P/E multiple is appropriate for the company's new platform-dominant profile. While historically volatile, NVIDIA's 90% share of the data center accelerator market and high switching costs for its CUDA (Compute Unified Device Architecture) software justify a premium over traditional semiconductor peers like AMD.
We're assuming mid-cycle earnings power of roughly $9.00 per share, which bridges current hyper-growth with inevitable tapering. This figure represents the average projected earnings across the next three years of the AI infrastructure build-out, providing a more stable anchor than the current high-growth quarterly prints that can be distorted by supply-chain timing.
We're assuming the "Blackwell" hardware cycle maintains NVIDIA's performance lead through at least FY2028. Current MLPerf benchmarks show NVIDIA maintaining a significant advantage in training and inference, which is the primary driver for the multi-year upgrade cycle currently underway at global data centers.
The biggest risk is a "digestion period" where hyperscalers pause new orders to integrate their massive current capacity, triggering a sharp cyclical correction. This would likely pull mid-cycle EPS estimates down from $9.00 toward $6.50, knocking roughly $85 off the per-share fair value as the multiple compresses to historical norms. Watch for "purchase commitments" in the 10-Q to drop below $15 billion as an early warning signal of cooling demand.
Bear case ($185): Hyperscaler capital expenditure (AI infrastructure spending) drops by more than 15% year-over-year in FY2028; or Non-NVIDIA custom silicon (like OpenAI’s Jalapeño) captures more than 20% of the training market share.
Bull case ($480): Software and services revenue (CUDA/NVIDIA AI Enterprise) grows to represent 12% of the total revenue mix; or Operating margins remain above 60% through the end of the Blackwell Ultra hardware cycle in FY2029.
Clearthesis wrote this report from 40 sources, including SEC filings, industry research, and recent news.
How did you like this thesis?
Your feedback helps us make reports better for you
© 2026 Clearthesis.ai · Report generated on July 8, 2026
This is an AI-generated analysis for informational purposes only and does not constitute financial advice. Data and analysis may not reflect recent developments if viewed significantly after the generation date. Always conduct your own due diligence before making any investment decisions.
The market is leaning bullish because Nvidia owns the essential software and hardware stack that every global company must use to build artificial intelligence. Nvidia's software platform, CUDA, forces developers to rely on its chips for building AI systems. This creates a massive moat that keeps customers buying their specialized hardware and protects their pricing power.
Skeptics think that a wave of internal chip development by major tech customers will eventually erode Nvidia's dominance. As industry leaders like OpenAI and SpaceX begin designing their own custom chips to run their AI models, Nvidia faces the long-term risk of losing its biggest and most profitable customers.