Nvidia reaches 96.2 billion dollars in quarterly revenues. Analysis of Jensen Huang, the AI Factory ecosystem, and the 500 billion fund for infrastructure.

Can a company bill 100 billion dollars in three months?

The numbers published last Wednesday do not belong to a sovereign wealth fund or a state conglomerate, but to a single chip manufacturer. Nvidia closed the second quarter with revenues of 96.2 billion dollars, a figure that nearly doubles the results of the same period last year (+106%). Wall Street had predicted 92 billion; the company beat expectations, projecting revenues of 108 billion dollars for the third quarter.

The market reacted with a stock rise of nearly 4%, despite the company having already reached a capitalization of 5 trillion dollars, becoming the most valuable company in the world. It is not just a matter of hardware sales, but of scale speed. The Guardian report highlights how Nvidia's results are now considered a proxy for the entire artificial intelligence industry: if Nvidia accelerates, the entire sector is perceived as healthy.

The moment when compute becomes profit

For years artificial intelligence was seen as a cost center, an experimentation laboratory where tech giants burned billions in research and development without an immediate return. That cycle seems to have broken. Jensen Huang defined this moment as the industry's inflection, stating that 'AI has reached its inflection point. It’s doing useful work. Its tokens are productive and profitable'.

The key metric to understand this transition is datacenter revenue, which in the second quarter reached 89 billion dollars, marking a growth of 117% compared to the previous year. When Huang states that 'now, compute is revenue', he is shifting the paradigm: computing power is no longer a support utility, but the direct engine of revenue generation for businesses. AI is no longer just promising efficiency, it is producing monetizable output.

Jensen Huang and the illusion of dependence

Nvidia's CEO does not present himself as a simple component supplier, but as the architect of a new industrial ecosystem. While large customers — the so-called hyperscalers — attempt to develop proprietary chips to reduce dependence on Santa Clara, Huang maintains a position of extreme confidence. His strategy is based on the conviction that hardware is only one part of a more complex puzzle.

Huang defines Nvidia as a 'platform – an entire AI factory platform that spans the entire AI lifecycle'. The goal is to make Nvidia infrastructure omnipresent and provider-agnostic, emphasizing that its technology 'is in every cloud'. In this sense, the illusion of dependence lies in the fact that, even if a customer built a chip similar in performance, they would have to rebuild the entire software stack and the management ecosystem that Nvidia has already standardized globally.

Why do customers keep buying despite proprietary chips?

Strategic analysis suggests that Nvidia's competitive advantage no longer resides exclusively in the technical superiority of the single chip, but in vertical integration. Companies continue to buy Nvidia hardware because the company offers a complete solution covering the entire AI lifecycle. Replacing a chip is possible; replacing an 'AI Factory' that coordinates data, compute, and distribution across every cloud is an operation with very high operational risk.

Business analysis: For an entrepreneur, this scenario indicates that technological lock-in has shifted from the product to the platform. Nvidia's strategy is to create an industrial standard so pervasive that the cost of migrating to proprietary chips exceeds the economic savings derived from internal production. Huang's confidence, declaring '100% confidence that our technology will continue to be extraordinary for them', rests on Nvidia's ability to evolve faster than customers can copy.

500 billion dollars to lock down the ecosystem

To sustain this growth and prevent financial bottlenecks, Nvidia has orchestrated an unprecedented financial engineering operation. It collaborated with leading investment firms to create a 500 billion dollar capital allocation pool. This move serves to ensure that infrastructure expansion is not slowed by customers' lack of liquidity or the prohibitive costs of building datacenters.

Beyond financial capital, the company is investing in massive physical assets. A concrete example is the commitment to develop an 8-gigawatt datacenter plant in Ohio. This investment signals that Nvidia is not just selling chips, but is actively shaping the energy and physical infrastructure necessary for next-generation AI.

Capital flows can be summarized as follows:

• Quarterly revenues: 96.2 billion dollars (beating estimates of 92 billion).
• Earnings per share: 2.22 dollars (against the predicted 2.09).
• Strategic capital pool: 500 billion dollars for ecosystem expansion.
• Q3 revenue target: 108 billion dollars.

Is the infrastructure bubble ready to burst?

Despite record numbers, a critical reading of the data reveals gray areas. The chip sector has faced pressure in recent weeks due to growing doubts about the ability of massive investments in AI infrastructure to generate adequate returns in the long run. If the companies buying Nvidia chips fail to transform 'token productivity' into net profits, demand could drop sharply.

The risk is that we are facing a cycle of overinvestment. If the AI Factory does not produce real economic value for the end user, the 500 billion dollar capital pool could become toxic debt or an unfruitful investment. Business Chief reports how Nvidia is seen as the driving force, but this also means that any slowdown in its growth would be interpreted as the signal of the end of the 'golden age'.

To monitor the possible formation of a bubble, the verifiable indicator to observe is the ratio between cloud providers' CAPEX spending and the growth of B2B AI service revenues: if the former continue to rise while the latter stagnate for two consecutive quarters, the risk of correction becomes systemic.

The AI Factory monopoly and European strategic autonomy

For Italian and European entrepreneurship, Nvidia's rise to a 'global platform' poses a problem of technological sovereignty. If the entire EU AI infrastructure rests on a proprietary US platform, strategic autonomy becomes nominal. The adoption of closed standards makes European companies dependent not only on Santa Clara's prices, but also on the strategic decisions of a single CEO.

In a regulatory context defined by the AI Act and NIS2 directives on cybersecurity, the concentration of computing power in a single provider creates a point of systemic vulnerability. For Italian companies, the challenge is not to compete in chip production, but to diversify AI implementation, preventing the entire corporate value chain from being tied to a single ecosystem. Dependence on Nvidia is today a productivity accelerator, but tomorrow it could become an insurmountable competitive constraint if open or sovereign alternatives are not developed at the European level.

FAQ

---

Sources: Theguardian, Rocketnews, Businesschief ·