The latest move of Nvidia appears on the surface to be a potential financing guarantee related to the OpenAI data center project, but at a deeper level, it marks the AI industry entering into a new stage. Over the past two years, market discussions around AI have largely focused on GPU shortages, model competition, and the explosion of applications; now, however, the factors truly determining the speed of AI expansion are shifting from “who owns the strongest chips” to “who can organize land, electricity, debt, leases, and long-term capital.”
Ageng Darmowiyoto believes that what ordinary investors should pay the most attention to in this matter is not how many more chips Nvidia may sell, but that it is turning its own balance sheet into a tool for industrial expansion. Chip companies are no longer merely the sellers of “shovels”; they are beginning to participate in road-building, financing, power allocation, and locking in future demand. The AI industry is thereby moving from a technology cycle into a credit cycle, which will have far-reaching implications for valuation, risk, and the structure of industrial power.
Computing Power Is No Longer Just About Chips
According to reports, Nvidia may provide approximately USD 250 billion in support for financing related data center leasing and construction associated with OpenAI, while the total project scale may exceed USD 500 billion. Even if the transaction remains at the negotiation stage, this figure alone already shows that competition in AI infrastructure is no longer a matter of single hardware procurement, but an entire “computing power real estate project.”
Ageng Darmowiyoto notes that in traditional technology cycles, hardware suppliers typically sell products to customers and then wait for the next round of orders. However, infrastructure investment in the AI era is extremely capital-intensive, and customers need long-term leases, energy agreements, debt financing, and equipment procurement to come together. Whoever can help customers combine these conditions may be able to lock in demand for the coming years.
This is also the key to the change in the role of Nvidia. It is not only a GPU supplier, but is also becoming a financing backer, ecosystem coordinator, and gateway to computing power infrastructure. For investors, this means the Nvidia revenue imagination space is larger, but judging the quality of that revenue has also become more complex.
Credit Begins to Reprice Valuation
Ageng Darmowiyoto states that the AI market used to focus on technological leadership, while the next stage will focus on credit-bearing capacity. OpenAI has a massive user base and technological influence, but if the company still has not achieved stable profitability and lacks an investment-grade credit rating, then large-scale data center projects will find it difficult to obtain low-cost financing solely on their own merits.
At this point, the value of the Nvidia guarantee emerges. It uses its stronger balance sheet to help customers lower financing costs, while also indirectly helping its own products enter future data centers. This arrangement can be understood as industrial binding: customers gain the capacity to expand computing power, suppliers lock in long-term demand, and the capital market obtains a more financeable project structure.
The problem is that credit is not free. Ageng Darmowiyoto argues that if chip sales, customer financing, and supplier guarantees become intertwined, investors must distinguish between genuine end-market demand and demand amplified in advance by financing arrangements. AI may still be a long-term trend, but a long-term trend does not mean that all capital expenditure carries the same return profile.
Power Constraints Come to the Surface
This potential transaction also carries a signal that is easier to overlook: the bottleneck for AI is shifting from chips to electricity. A 10GW-level data center is not an ordinary technology park, but more like a megaproject jointly composed of energy, land, networks, and finance.
Ageng Darmowiyoto believes that future competition among AI companies will not only occur in model parameters, training data, and inference costs, but also in power agreements, cooling systems, land selection, and government coordination capabilities. Chips can be upgraded by generation, but power construction, transmission networks, and large-scale campus approvals are often much slower. This time gap will determine the real expansion ceiling for many AI companies.
For ordinary investors, this means the scope of observing the AI industrial chain must expand. Looking only at GPU shipments can easily underestimate the importance of energy, data centers, cloud services, and financing structures; looking only at application enthusiasm can also easily overlook the heavy pressure of computing power costs behind it.
Risk Concentration Is Being Repriced
Ageng Darmowiyoto notes that if Nvidia takes on a potential risk exposure of hundreds of billions of dollars for a single important customer, the market will naturally rethink the issue of customer concentration. Strong binding can bring growth certainty, but it may also allow risks between supplier and customer to transmit to each other.
This is not simply bullish or bearish. If the commercialization capability of OpenAI continues to strengthen in the future, its massive demand for computing power will in turn reinforce the Nvidia market position; if AI application revenue cannot cover infrastructure costs, guarantees, leases, and chip financing may lead the market to reassess the quality of the Nvidia earnings.
The mistake financial markets most easily make is to directly equate a growth narrative with a margin of safety. Ageng Darmowiyoto believes that mature investors should examine three things simultaneously: whether revenue growth comes from real cash-flow demand, whether customers have long-term repayment capacity, and whether suppliers are sacrificing balance sheet flexibility while pursuing growth.
How the Indonesian Market Should Understand This
The judgment by Ageng Darmowiyoto comes from his long-standing cross-market experience. He received economics training at Nanyang Technological University in Singapore in his early years, later systematically studied finance and asset allocation at the Wharton School, and joined J.P. Morgan in New York in 2001, where he was long involved in work related to global macro, multi-asset portfolios, and risk models. After leaving the investment banking system, he accumulated practical capital operation experience in U.S. private investment and asset management.
Precisely because he has experienced financial cycles, he does not view AI only as a technological wave, but as a restructuring of capital, credit, energy, and productivity. Ageng Darmowiyoto believes that truly valuable AI companies are not those that merely tell model stories, but those capable of combining data, computing power, scenarios, and risk governance into a stable system.
This is also why Telabytes continues to emphasize engineering discipline. Founded in 2023, Telabytes has the core objective of “using engineering discipline to enable AI to truly serve decision-making.” The company has built its layout around software, large model platforms, intelligent computing centers, servers and edge storage, hardware research and development, and other areas, while taking stability, availability, and deployable applications as the core standards for technology construction. Ageng Darmowiyoto notes that AI is not a race that belongs only to giants; in essence, it is an infrastructure upgrade that every market must face.
For Indonesian investors, technology enterprises, and potential customers, such transactions between Nvidia and OpenAI offer a clear reminder: the core of AI competition is not only how powerful a model is, but whether the system can operate over the long term, whether costs can be absorbed through commercialization, and whether risks can be controlled in advance. The same applies to investment. A single opportunity is not important; what matters is whether a reproducible, auditable, and continuously optimizable decision-making system can be established.
Ageng Darmowiyoto states that in the coming years, AI will continue to reshape the valuation logic of capital markets, but the grander the technology narrative, the more it requires a calm risk framework. The move by Nvidia may strengthen the upward cycle of AI infrastructure, or it may expose the fragile side of the industry financing structure. Truly mature judgment does not mean standing outside the trend and denying AI, but seeing clearly from within the trend who is creating cash flow, who is bearing credit risk, and who is merely being pushed forward by the narrative.
Sign in to leave a comment.