Jensen Huang's Case for 70% Growth and His Response to Round-Trip Deal Concerns
Jensen Huang projects 70% revenue growth and pushes back on round-trip deal allegations.
NVIDIA CEO Jensen Huang projects roughly 70% revenue growth next year, backed by surging AI infrastructure demand, rapid GPU architecture iteration, and broad positioning across the AI value chain. He directly pushed back on round-trip deal concerns, arguing customer purchases reflect genuine compute needs. NVIDIA benefits from sovereign AI buildouts, the CUDA software moat, and the inference market boom — though customer concentration and uncertain AI monetization timelines remain real risks.
NVIDIA's Ambition: ~70% Revenue Growth Next Year
At the peak of the AI chip boom, NVIDIA CEO Jensen Huang has thrown out a number that left Wall Street breathless: the company expects to deliver roughly 70% revenue growth next year. For a tech giant already valued in the trillions, that kind of growth rate is nearly unprecedented.
Huang's core argument is straightforward — NVIDIA "has its finger in every pie." That phrase neatly captures the company's current strategic footprint: from data center training chips to inference accelerators, from cloud hyperscalers to sovereign AI initiatives, from enterprise applications to consumer graphics, NVIDIA has embedded itself into virtually every critical link of the AI value chain.
The Core Engines Behind 70% Growth
Underpinning this optimistic outlook is the relentless global appetite for AI infrastructure. As large language model parameter counts continue to balloon and inference demand explodes, compute has become the new oil. Huang sees the year ahead as "a year of plenty" — a signal that demand for AI compute is nowhere near its ceiling.
From a technology cadence perspective, NVIDIA has compressed its GPU architecture refresh cycle to roughly once a year. The Blackwell GPU series is ramping into production on schedule, and a increasingly clear product roadmap gives the company sustained pricing power and a durable moat around its market share.
The Round-Trip Deal Allegations — and Huang's Rebuttal
Notably, Huang took care to emphasize that NVIDIA's transactions are not "circular deals." This clarification wasn't made in a vacuum — it directly addresses one of the most pointed criticisms circulating about the AI supply chain.
Where the Round-Trip Deal Concerns Come From
The concern goes like this: NVIDIA invests in or backs certain AI startups and cloud providers, and those same companies then turn around and purchase large volumes of NVIDIA chips. Critics worry that this pattern — where capital and purchase orders circulate between the same parties — could artificially inflate NVIDIA's revenue figures and obscure whether real end-market demand actually exists.
In other words, if NVIDIA is using its own capital to "cultivate" customers who then spend that capital buying NVIDIA products, the sustainability of that revenue stream is legitimately in question. In an environment where AI bubble fears are already simmering, this critique carries real bite.
How Huang Defends NVIDIA
Huang maintains that transactions between NVIDIA and its partners are grounded in genuine commercial need — not engineered capital loops. His core logic: customers buy GPUs because they actually need compute to power their AI businesses, not simply because they've received investment from NVIDIA.
Ultimately, the persuasiveness of this defense will be tested by time and earnings data. If downstream AI applications continue generating real commercial returns, the compute demand chain is healthy. If monetization downstream keeps getting deferred, the round-trip deal doubts won't fully go away.
The Opportunities and Risks Behind NVIDIA's Growth Story
Three Structural Tailwinds
Zooming out, NVIDIA's optimism isn't without foundation. Three structural factors stand out:
First, the sovereign AI buildout wave. Major economies worldwide are racing to build AI infrastructure, with governments and large enterprises ramping up compute investment. This gives NVIDIA a vast incremental market that no longer depends solely on a handful of hyperscale cloud providers.
Second, the software moat of the CUDA ecosystem. NVIDIA's competitive advantage isn't just hardware performance — it's the deep software lock-in created by the CUDA ecosystem. Even if competitors close the gap on hardware specs, dislodging developer dependence on CUDA remains an enormous challenge.
Third, the inference market explosion. As AI applications shift at scale from training into inference deployment, inference compute is emerging as a powerful new growth driver, opening up a significantly larger addressable market for NVIDIA.
Risks That Can't Be Ignored
That said, a 70% growth target implies sky-high market expectations. Any earnings miss, however modest, could trigger a sharp valuation correction. Beyond that, elevated customer concentration, supply chain capacity constraints, and uncertainty around the pace of real-world AI adoption all remain material variables in NVIDIA's growth narrative.
At its core, the round-trip deal concern reflects a deeper investor anxiety about AI's return-on-investment timeline. When vast amounts of capital pour into compute buildouts, investors naturally ask: when will these investments generate positive cash flows, and how?
A Boom Narrative Still Needs Real Demand to Back It Up
The picture Huang paints is one of relentlessly surging compute demand, with the 70% revenue growth forecast as its quantitative expression. For investors and industry observers, the critical question is whether this growth cycle is driven by real, sustainable end-market demand — or whether it is, at least in part, amplified by capital flows forming circular loops.
Either way, NVIDIA remains at the undisputed center of the AI wave. Every public statement Huang makes speaks not only to NVIDIA's own fate, but reflects the broader health and trajectory of the entire AI industry. The upcoming earnings season will be the most direct test of that confidence.
Related articles
How the CUDA Ecosystem Keeps the A100 in Service for a Decade: A Deep Dive into NVIDIA's Moat
How NVIDIA's CUDA software ecosystem keeps the A100 GPU mission-capable for nearly a decade — and transforms GPU compute into a rentable, durable, financeable asset.
LangChain + MCP: From Core Concepts to Agent Tool Calling in Practice
Learn how LangChain and MCP work together — covering LLM tool calling, Agent architecture, and conversation history management to build real-world AI applications.
Probabilistic Machine Learning: Why It's the Cornerstone to Unlocking the ML Black Box
Without probability theory, ML is always a black box. This article explores why probabilistic foundations are essential for understanding machine learning algorithms, Bayes' theorem, MLE, and more.