Intel’s $20 billion equity raise and Nvidia’s up-to-$105 billion Ohio data center commitment show that AI is moving from a software race into a balance-sheet race, where long-term infrastructure obligations matter as much as reported capital expenditures.
The AI boom is no longer defined only by who has the best model. It is increasingly defined by who can finance the chips, power, data centers, and long-term obligations behind the model.
Intel has decided to significantly increase the amount of capital it plans to raise for business development. Initially, the company announced plans to issue shares worth approximately $15 billion, but ultimately, the target was raised to $20 billion. Taking into account the right of investors to purchase additional shares worth another $3 billion within 30 days, the total amount raised could reach $23 billion. This has become an important signal for investors. Intel is ready to invest significantly more of its own resources in restoring its manufacturing business and expects to turn contract chip manufacturing into one of its key sources of future profits. Thus, the company could have a chance to return to the Dow Jones Industrial Average.
Such a large‑scale placement may indicate that Intel is already seeing sufficiently strong demand for the services of its contract manufacturing division, Intel Foundry Services. The company will be able to use the funds raised to expand and modernize the production facilities needed to serve external customers. An additional positive factor was Intel CEO Lip-Bu Tan’s participation in the placement, which investors may view as an indirect sign of his confidence in the company’s prospects.
Intel’s main bet is on its 18A manufacturing technology. If the technology proves competitive enough for the company to win orders from leading chip developers, the current multibillion‑dollar investments could lay the groundwork for a significant shift in the company’s financial structure. Intel’s contract business is expected to break even in the fourth quarter of 2027, and in 2028, it could begin contributing positively to the company’s overall profitability.
Demand from major technology corporations looks particularly promising. Potential Intel clients include Apple, as well as cloud giants such as Google and Amazon. This concerns not only the production of processors themselves but also the advanced packaging of custom chips for data centers using EMIB technology. According to GF Securities, Intel’s revenue from these services could grow from approximately $1.1 billion in 2027 to $7 billion in 2028. If the forecast proves accurate, the growth rate in this area could be strong enough to significantly reshape the financial picture of the entire company.
At the same time, Intel has to operate in an extremely capital‑intensive industry, where competition for computing power and production resources is becoming increasingly fierce. Notably, as Intel expands its production plans, the largest AI developers continue to increase their commitments to infrastructure suppliers.
The most striking example is Nvidia, which has agreed to provide funding for the construction of a large data center in Ohio for OpenAI, with a total investment of up to $105 billion. The facility will have a power capacity of up to 8 GW, and the first 800 MW are expected to be operational by 2028. Meanwhile, Nvidia expects that OpenAI’s total equipment purchases could generate around $600 billion in revenue for the company by 2030.
Formally, Nvidia denies that this arrangement amounts to circular financing: the company claims that OpenAI will independently pay for the rental of computing power. However, the very structure of such transactions shows how heavily the AI industry depends on external capital to sustain its growth.
And this is where the main financial risk of the current investment cycle emerges. The total liabilities of U.S. technology companies associated with the construction of AI infrastructure have already exceeded $3 trillion. For comparison, reported capital expenditures of the largest industry players in the previous fiscal year amounted to about $600 billion.
At the same time, a significant portion of future spending is not reflected in the companies’ current capital expenditures. This includes long‑term data center leases, loan obligations, guarantees, and other agreements with payments spread over several years. For example, Alphabet’s total commitments of this kind have already reached $811 billion, compared with $332 billion just three months ago.
More insight into AI-related spending and the broader state of the technology industry could come from Nvidia’s earnings, scheduled for August 26, 2026. Ahead of the report, investors remain cautious, with U.S. stock futures — including Nasdaq, S&P 500 and Dow Jones futures — slipping on Monday, August 24.
Thus, Intel is now betting on growing demand for chip production and packaging, while Nvidia and other tech giants are betting on the continued expansion of AI infrastructure. As long as this model works, expectations of future demand allow companies to raise more capital and launch new projects. But for investors, it is becoming increasingly important to consider not only how much money companies are willing to spend on AI, but also whether future revenues will be sufficient to justify the trillions of dollars in accumulated liabilities. This question may turn out to be the main challenge for the entire industry in the coming years.
