Nvidia’s Business Model Overhaul and China’s H200 Nod Converge to Reshape the Investment Case
Published on 07/13/2026 at 12:11 | Redaktion boerse-global.de
Nvidia is quietly rewriting the rules of its own industry. The chipmaker, long known for selling high-priced processors to the world’s largest cloud operators, has begun financing its customers instead. And while that transformation unfolds, Beijing has just pried open a sliver of the Chinese market for its H200 chips. The combination is creating an investment narrative that goes far beyond the usual supply-demand math.
The shift began on 1 July 2026, when Nvidia formally adopted an “AI Factory” strategy. Rather than merely shipping silicon to hyperscalers, the company is now using its own balance sheet to extend credit lines to smaller cloud providers and to structure revenue-sharing agreements. It is effectively becoming a merchant banker to the global AI infrastructure buildout. Some 40,000 Grace-Blackwell units are already humming inside newly constructed “AI campuses” that Nvidia helped finance.
The logic is straightforward. The hyperscale operators have largely completed their initial buildouts, and the next wave of growth lies with mid-tier providers and startups that cannot afford the up-front cost of tens of thousands of accelerators. Nvidia’s $4.3 trillion market capitalisation gives it the firepower to bridge that gap. In return, it locks in recurring demand rather than depending on one-off purchase orders.
The bet carries obvious risk: Nvidia now shoulders the credit risk of its customers, not just the risk of order cancellations. But it also aligns with a deeper industry shift. The company calls this “the year of inference” – the moment when the bulk of data-centre compute pivots from training large models to real-time, production-grade inference. Analysts estimate that inference now accounts for 80 to 90 percent of total AI workload.
Should investors sell immediately? Or is it worth buying Nvidia?
That pivot dovetails with Nvidia’s next-generation architecture. The Vera-Rubin platform, successor to Blackwell, is due to ramp in the second half of 2026. It is optimised for power efficiency rather than raw performance, and features a closed-loop water cooling system designed to cut on-site water consumption to near zero – an explicit response to the energy and water constraints plaguing large data centres.
Amid this strategic re-engineering, a geopolitical window has opened. China has granted permission for select domestic AI companies – Alibaba, ByteDance and DeepSeek among them – to purchase a limited quantity of Nvidia’s H200 chips. The move reverses some of Beijing’s earlier restrictions on US technology. Nvidia’s official forecasts have not included any data-centre revenue from China, so any demand that materialises would be additive.
The stock closed the most recent week at €184.60, a gain of nearly 8 percent over the prior seven days. Even after that bounce, it remains 8.84 percent below the 52-week high of €202.50 set on 14 May. The 50-day moving average sits at €181.22, with the 200-day average lagging at €164.78 – a configuration that suggests the market is still calibrating the right valuation for a company in transition. The relative strength index of 58.6 leaves room for further upside before the stock becomes overbought.
Goldman Sachs has flagged the valuation as unusually compressed. The current forward price-to-earnings ratio of 21.7 is close to the S&P 500 average and well below Nvidia’s own five-year mean of 72. Analysts’ average price target of €264.16 implies potential upside of 43 percent from current levels. That optimism rests not only on the new business model but also on the prospect of “sovereign AI” projects in countries such as India, Brazil and the United States.
Nvidia at a turning point? This analysis reveals what investors need to know now.
Yet the road is not without potholes. A report from SemiAnalysis earlier this month claimed that Nvidia’s Kyber NVL144 rack, part of the Rubin-Ultra platform, is facing a delay of more than twelve months due to manufacturing issues with the midplane board. Nvidia forcefully denied the report, telling Bloomberg that the entire roadmap remained intact and that existing Rubin systems are already in full production, with shipments to eight cloud customers scheduled for the autumn. The episode underscores how sensitive sentiment has become to any whisper of execution risk.
What makes the current moment so intriguing is the convergence of these three forces: a business model that is transforming Nvidia from supplier to financial partner, a Chinese market that is cautiously reopening, and a valuation that has retreated to levels rarely seen in the company’s recent history. Whether the stock can sustain its recovery depends less on abstract AI enthusiasm and more on a single, tangible question: can Nvidia continue to prove, quarter after quarter, that the world’s AI infrastructure still needs precisely what it builds?
Ad
Nvidia Stock: New Analysis - 13 July
Fresh Nvidia information released. What's the impact for investors? Our latest independent report examines recent figures and market trends.
Disclaimer regarding our articles: No investment advice, no buy or sell recommendation. Information on prices, companies, and markets is provided without guarantee; changes are possible at any time. Stock market transactions can lead to substantial losses. Our articles are created and reviewed in whole or in part automatically with the support of AI.
