An analysis said Nvidia could use expanding government investment in sovereign AI to reduce its dependence on large cloud companies. Demand for building AI infrastructure at the national level is emerging as a new growth driver. It could also help ease a revenue structure centred on a small number of customers.

CNBC reported on June 8 that U.S. investor Jim Cramer assessed that Nvidia's sovereign AI business could help address the customer concentration problem that investors have worried about.

The key is a shift in the customer mix. The market has raised the possibility that large cloud operators such as Amazon, Alphabet and Microsoft could reduce their reliance on Nvidia over the long term as they accelerate development of their own AI semiconductors. For Nvidia, the fact that a significant portion of revenue is concentrated among a small number of hyperscalers has been cited as a potential risk.

Cramer explained that Nvidia is pursuing a strategy to broaden its customer base to ease that structural burden. "Nvidia doesn't want to be held back by customers trying to lower their dependence on its products," he said. "It is moving in the direction of securing as many customers as possible," he said.

Sovereign AI sits at the centre of that strategy. Sovereign AI refers to efforts by governments to secure technology sovereignty by building AI infrastructure, data centres and computing capacity at home. As AI technology has recently emerged as a key element of national competitiveness, countries around the world are actively working to build their own AI ecosystems.

Cramer said Nvidia Chief Executive Jensen Huang (젠슨 황) has travelled around the world over the past year stressing the importance of sovereign AI.

Singapore, India, Japan, Germany, Switzerland, Taiwan, Israel, Qatar, the United Arab Emirates and Saudi Arabia are among countries known to be investing in AI infrastructure projects based on Nvidia technology. This is seen as an example showing Nvidia's customer base expanding beyond big tech companies to government and public projects.

The market has also continued to raise concerns about the profitability of hyperscalers' AI investment. Critics say huge costs are being poured into building large-scale AI infrastructure, but returns have yet to be sufficiently verified. Cramer said he keeps hearing that hyperscalers are losing money while buying Nvidia chips in large volumes, but added government-level purchases are different in nature.

He analysed that national AI infrastructure projects have different investment logic from private companies because the goal is to secure industrial competitiveness and strengthen technology sovereignty, rather than short-term profits.

This characteristic is also viewed as positive for Nvidia's business stability. Private companies can adjust purchase volumes depending on economic conditions or investment returns, but national projects are more likely to lead to relatively long-term and stable demand.

Cramer said sovereign AI currently accounts for about 14% of Nvidia's overall business. He projected that share could expand further if projects under way move into full-scale buildout. "Governments are buying Nvidia chips on a large scale for sovereign AI programmes," he said. "Even that demand alone will gradually reduce Nvidia's dependence on hyperscalers," he said.

The industry is watching how quickly Nvidia can expand its influence in the national AI infrastructure market as competition intensifies among large cloud companies developing their own AI chips, seeing it as a key variable that will determine future growth momentum.

Some forecasts also say sovereign AI could become Nvidia's new core revenue source as competition spreads globally to secure AI technology sovereignty.