Megalomania? Hypergrowth on Credit: OpenAI's (ChatGPT) $100 Billion Bet Against Economic History

The dissonance between technological promise and economic reality

OpenAI has set out to redefine the boundaries of artificial intelligence. But while the company is making ambitious predictions about the performance of its models, it is also planning revenue growth that exceeds all historical benchmarks. Epoch AI's current analysis paints a remarkable picture: OpenAI aims to increase its revenue from $13 billion in 2025 to $100 billion by 2028. This corresponds to a required annual growth rate of 97 percent over three years. By comparison, even the fastest-growing companies in recent technology history, such as Tesla and Meta, needed seven years to make the leap from $10 billion to $100 billion in annual revenue, and Google even needed a full decade. OpenAI aims to reach this milestone in just three years, a pace that, according to Epoch AI, has no historical precedent.

These ambitions raise fundamental questions. Is this the legitimate extrapolation of a technological revolution whose transformative potential is rewriting the rules of the market economy? Or are we witnessing a repetition of historical patterns in which exaggerated growth expectations and massive infrastructure investments inevitably lead to overcapacity and economic disruption? The answer probably lies somewhere in between and requires a nuanced consideration of the technological, economic, and structural factors that determine OpenAI's growth trajectory.

This article analyzes OpenAI's growth strategy in the context of economic history, examines the underlying market mechanisms, and assesses the likelihood of the company achieving its goals. It highlights both the innovative strengths and the structural risks associated with such an aggressive expansion strategy. The analysis is divided into eight sections: a historical overview, an identification of the core factors driving the current AI boom, a review of the current situation, comparative case studies, a critical assessment of risks, an outlook on potential development paths, and concluding strategic implications.

Suitable for:

• Profit over principles? The sex revolution – ChatGPT gets dirty and why OpenAI is now focusing on eroticism

From research laboratory to the most valuable startup in the world

OpenAI's history is inextricably linked to the rise of large-scale language models and the broader commercialization of artificial intelligence. Founded in 2015 as a non-profit research institution, the company initially positioned itself as a counterweight to major technology corporations, pursuing the goal of developing artificial general intelligence for the benefit of all humanity. The founders, including Sam Altman and Elon Musk, recognized early on that developing advanced AI systems would require enormous capital resources.

The decisive turning point came in 2019 with the transformation to a hybrid structure combining for-profit and non-profit elements. This realignment enabled OpenAI to secure an initial investment of one billion dollars from Microsoft. The partnership with the software giant proved strategically valuable: OpenAI gained access to Microsoft's Azure cloud infrastructure and the necessary computing resources, while Microsoft, in return, received exclusive access to OpenAI's technology.

The company's revenue initially grew moderately in the following years. In 2020, OpenAI generated only $3.5 million in revenue, but a year later, it reached $28 million. The breakthrough came in November 2022 with the release of ChatGPT, a chatbot based on GPT-3.5, which reached one million users within five days and surpassed the 100 million user threshold within two months. This viral success instantly transformed OpenAI from a research lab into a commercial powerhouse.

Revenue growth accelerated dramatically. In 2023, OpenAI surpassed the $1 billion mark in annual revenue for the first time, reaching $1.6 billion. In 2024, revenue more than doubled to $3.7 billion. For 2025, the company forecasts annualized revenue of $13 billion, representing a 251 percent increase over the previous year. This momentum is supported by a growth rate of approximately 3.2x per year since the end of 2023.

Parallel to revenue growth, the company's valuation soared to dizzying heights. A financing round in March 2025 valued OpenAI at $300 billion. Just a few months later, in October 2025, the valuation reached the $500 billion mark through a secondary share sale to investors such as SoftBank, Thrive Capital, and T. Rowe Price. This made OpenAI the most valuable startup in the world, even surpassing Elon Musk's SpaceX.

This historic development highlights the extraordinary speed with which OpenAI has evolved from a research project to one of the dominant players in the global AI industry. At the same time, it raises the question of whether these valuations are based on realistic assumptions about future growth and profitability or whether they represent an overvaluation reminiscent of previous technology bubbles.

Drivers, players and the mechanics of the AI ​​market

The current AI boom is driven by a complex interplay of various factors. At the heart of this is technological innovation itself: Large-scale language models have made remarkable progress in natural language processing, logical reasoning, and solving complex tasks in recent years. These capabilities open up application possibilities in virtually all economic sectors, from customer service automation to software development and scientific research.

The key players can be divided into several categories. First and foremost are the developers of large-scale language models such as OpenAI, Google with Gemini, and Anthropic with Claude. These companies compete for technological leadership and market share, with OpenAI currently holding a dominant position with ChatGPT. ChatGPT's market share in the AI ​​assistant space is estimated at 62.5 percent.

A second key group is the infrastructure providers. Nvidia dominates the AI ​​accelerator market with a market share of approximately 95 percent. The company's graphics processors, especially the H100 and A100 series, have become indispensable for training and running large language models. Nvidia is profiting massively from the AI ​​boom and has multiplied its valuation in recent years. Recently, however, other players, such as AMD and Broadcom, have entered the market, attempting to challenge Nvidia's dominance.

Cloud providers such as Microsoft Azure, Amazon Web Services, and Oracle constitute a third important category of players. They provide the computing capacity required to train and run AI models. OpenAI's close partnership with Microsoft and Oracle is particularly important in this regard.

The economic incentive structures driving these players are complex. For OpenAI and its competitors, it's about establishing a dominant market position in a technology segment that has the potential to transform large parts of knowledge work. McKinsey estimates that generative AI could contribute between $2.6 and $4.4 trillion annually to global economic output. Given such forecasts, even investments in the hundreds of billions seem justified.

For infrastructure providers like Nvidia, this creates direct demand for their products. Market mechanics follow a self-reinforcing logic: The more capital flows into the development of larger and more powerful models, the greater the demand for computing power and thus for chips. This dynamic has led to a veritable arms race, with companies like OpenAI signing long-term supply contracts worth hundreds of billions of dollars.

Another key driver is the availability of capital. The low interest rates of recent years and the general euphoria surrounding artificial intelligence have led investors to be willing to invest enormous sums in AI startups. OpenAI closed a $40 billion financing round in the first half of 2025 alone and also secured a $4 billion revolving credit facility. This capital infusion enables the company to pursue its ambitious expansion plans despite massive operating losses.

The regulatory framework also plays a role, albeit an ambivalent one. On the one hand, there are efforts in important markets such as the European Union to regulate AI systems more strictly, which could increase development costs. On the other hand, governments, particularly in the US, actively support AI development. The Stargate project, which, with a total volume of $500 billion over four years, represents the largest AI infrastructure initiative in history, was launched with strong support from the Trump administration.

The underlying market mechanisms exhibit characteristics typical of technology markets. It is a market with high fixed costs and low marginal costs: Developing a large language model costs hundreds of millions to several billion dollars, while the cost of answering a single user query is comparatively low. This leads to strong economies of scale and favors the emergence of oligopolies or even monopolies.

At the same time, it's a market with network effects: The more users use a platform like ChatGPT, the more valuable it becomes through the data generated and user feedback, which can contribute to improving the models. However, these network effects are less pronounced in the case of large language models than, for example, in social networks, because users can switch relatively easily between different providers if a competitor offers a better model.

Indicators of unprecedented expansion and its limits

OpenAI's current situation is characterized by a discrepancy between impressive growth and massive financial losses. In the first half of 2025, the company generated revenue of $4.3 billion, which was already 16 percent higher than the entire previous year. At the same time, however, OpenAI recorded an operating loss of $7.8 billion. The loss margin is thus 181 percent of revenue, which clearly shows that the company is spending almost two dollars more for every dollar it earns.

The main cost drivers are clearly identifiable. Research and development consumed approximately $6.7 billion in the first half of 2025 alone. A significant portion of this is attributable to the computational costs for training new models and operating ChatGPT. Estimates for the training costs of the next model generation vary considerably: While GPT-4 cost an estimated $100 to $200 million, the training costs for GPT-5 could range between $500 million and $2 billion, depending on the source. These exponentially rising development costs represent a key challenge.

Added to this are personnel costs, which are also rising rapidly. OpenAI granted its employees stock options worth $2.5 billion in the first half of 2025, almost double the amount in the entire previous year. Intense competition for AI talent is driving up salaries and forcing companies to offer generous compensation packages.

ChatGPT's user base continues to grow dynamically. In October 2025, the platform recorded between 700 and 800 million weekly active users. This represents double the number from February 2025, when the number was 400 million. The platform processes 2.5 billion requests daily and ranks fifth among the most visited websites worldwide.

The central problem, however, lies in the conversion rate. Only five percent of users pay for a subscription, be it ChatGPT Plus for $20 a month or ChatGPT Pro for $200 a month. This equates to approximately 40 million paying users. Even this comparatively low conversion rate is above the average in the generative AI industry, where only three percent of users are willing to pay. Nevertheless, the fact remains that 95 percent of the user base currently generates no direct revenue.

Approximately 75 percent of total revenue comes from consumer products, primarily ChatGPT subscriptions. The enterprise business, while growing, remains comparatively small. In June 2025, OpenAI reported three million paying business customers for its ChatGPT Enterprise, ChatGPT Team, and ChatGPT Edu products. By September, this number had risen to five million. While this represents healthy growth, the B2B segment lags significantly behind the consumer business.

The $500 billion valuation implies a price-to-sales ratio of approximately 38.5 times the projected revenue of $13 billion for 2025. By comparison, software companies are typically valued at two to four times their annual revenue. Even high-quality SaaS companies with strong growth rarely achieve multiples above ten. OpenAI's valuation is thus several times higher than historical averages and reflects investors' extreme growth expectations.

These expectations are based on the assumption that OpenAI can achieve its $100 billion revenue target by 2028. To achieve this, the company would have to overcome several challenges: The number of paying users would have to increase dramatically, possibly to 200 to 300 million. At the same time, it would have to develop new revenue streams, such as advertising, e-commerce integrations, or high-priced productivity tools for businesses.

The infrastructure commitments OpenAI has entered into exacerbate the pressure to succeed. The contracts with Nvidia, AMD, and Broadcom total approximately $1.3 trillion over a decade. The Stargate project envisions investments of $500 billion over four years. These commitments far exceed current and even projected revenues and require continued capital injections from investors or a significantly faster improvement in profitability.

A new dimension of digital transformation with 'Managed AI' (Artificial Intelligence) – Platform & B2B Solution | Xpert Consulting - Image: Xpert.Digital

Here you will learn how your company can implement customized AI solutions quickly, securely, and without high entry barriers.

A Managed AI Platform is your all-round, worry-free package for artificial intelligence. Instead of dealing with complex technology, expensive infrastructure, and lengthy development processes, you receive a turnkey solution tailored to your needs from a specialized partner – often within a few days.

The key benefits at a glance:

⚡ Fast implementation: From idea to operational application in days, not months. We deliver practical solutions that create immediate value.

🔒 Maximum data security: Your sensitive data remains with you. We guarantee secure and compliant processing without sharing data with third parties.

💸 No financial risk: You only pay for results. High upfront investments in hardware, software, or personnel are completely eliminated.

🎯 Focus on your core business: Concentrate on what you do best. We handle the entire technical implementation, operation, and maintenance of your AI solution.

📈 Future-proof & Scalable: Your AI grows with you. We ensure ongoing optimization and scalability, and flexibly adapt the models to new requirements.

More about it here:

• The Managed AI Solution - Industrial AI Services: The key to competitiveness in the services, industrial and mechanical engineering sectors

Lessons from the rise of digital giants and their limitations

A look at comparable companies and their growth paths offers valuable insights into the feasibility of OpenAI's ambitions. Google, now Alphabet, reached the $100 billion mark in annual revenue within ten years of its IPO in 2004. The company benefited from near-monopolistic access to the lucrative search engine market and was able to generate high margins from advertising revenue. Google's business model was based on low marginal costs and strong network effects, enabling continued profitability.

Meta, formerly Facebook, also took seven years to grow from $10 billion to $100 billion. Meta also benefited from strong network effects and a high-margin, advertising-based business model. Key to Meta's success was its ability to monetize a massive user base, initially on desktop and later on mobile devices. The acquisition of Instagram and WhatsApp further expanded its user portfolio.

Tesla presents an interesting comparison case because it operates in a capital-intensive industry with lower margins. Tesla also reached its $100 billion revenue target in about seven years, but benefited from a period of exceptionally high valuations for electric vehicle manufacturers and a charismatic CEO who embodied the brand. Tesla struggled with profitability issues and negative cash flow for years before crossing the profitability threshold.

Comparing these companies reveals both parallels and important differences to OpenAI. All three companies benefited from technological innovations that transformed existing markets. All three had strong brands and charismatic leadership figures. However, Google and Meta achieved profitability significantly earlier in their development than OpenAI. Tesla, on the other hand, recorded losses over long periods but was able to bridge these through continuous capital raising.

A critical difference lies in the nature of economies of scale. At Google and Meta, the cost per user decreases significantly as the user base grows, as infrastructure costs remain relatively constant. At OpenAI, however, computational costs increase almost proportionally with usage, as every request to ChatGPT consumes computing resources. CEO Sam Altman admitted that OpenAI is losing money on the $200 ChatGPT Pro subscription because users are using the service more intensively than expected. This points to a fundamental problem: Without dramatic cost reductions, growth doesn't automatically lead to improved profitability.

Another relevant comparison concerns companies that failed in their attempt to sustain extremely rapid growth. During the dot-com bubble in the late 1990s, hundreds of internet companies emerged with similarly ambitious growth forecasts. The majority failed because revenues didn't keep pace with expectations, and investors eventually lost patience. The telecommunications sector also saw massive misinvestments as companies built fiber optic networks with capacities that far exceeded actual demand.

Chinese AI development offers another interesting point of comparison. DeepSeek, a relatively unknown Chinese startup, caused a stir in early 2025 when it released a language model that could compete with leading Western models, but reportedly cost a fraction of the development costs. DeepSeek's R1 model reportedly cost a mere $5.6 million to develop, compared to over $100 million for GPT-4. If it is confirmed that comparable performance can be achieved with significantly fewer resources, this will challenge the assumption that massive investments in computational power are the only path to advanced AI systems.

Suitable for:

• Is ChatGPT from OpenAI and Google Gemini AIaaS – Artificial Intelligence as a Service?

Dislocations, uncertainties and the anatomy of possible undesirable developments

The risks associated with OpenAI's growth strategy can be divided into several categories. First, there are significant technological uncertainties. The so-called scaling laws, according to which larger models with more training data and computing power automatically improve, may reach their limits. There are indications that newer models no longer show the same performance leaps as previous generations. For example, OpenAI's GPT-5 reportedly consumed less computing power during training than GPT-4.5 without producing significantly better results. This could indicate that the simple scaling laws are no longer valid and new approaches are required.

The competitive situation is becoming increasingly fierce. Google with Gemini, Anthropic with Claude, and Meta with its Llama models are all investing heavily in the development of competing systems. Each of these players has significant resources and established distribution channels. Google can integrate Gemini into its search and productivity tools, while Meta can integrate its models into Facebook, Instagram, and WhatsApp. OpenAI lacks a comparable ecosystem, increasing its dependence on ChatGPT as its primary distribution channel.

The cost structure represents a structural problem. The computational costs for running large language models are enormous and increase with usage. OpenAI spends an estimated 60 to 80 percent of its revenue on computational costs alone. This leaves little room for profitability, especially given the additional costs for personnel, research, and operations. A significant reduction in inference costs would be necessary, but whether and when this will be achieved remains uncertain.

Dependence on a few infrastructure providers poses additional risks. Nvidia virtually completely controls the market for AI accelerators, giving the company considerable pricing power. While OpenAI is trying to reduce this dependence through contracts with AMD and Broadcom, these alternatives require time to build up production capacity. Should chip supply bottlenecks or drastic price increases occur, this could significantly impact OpenAI's expansion plans.

Regulatory risks are increasing. Questions regarding copyright on training data, data protection, and liability for AI-generated content remain largely unresolved. Should courts or legislators decide that AI companies must pay for the use of copyrighted training data, this could dramatically change the cost structure. Stricter data protection regulations or restrictions on certain use cases could also slow growth.

The risk of an infrastructure bubble is real. The historical parallels to the telecommunications bubble of the late 1990s are striking. Back then, massive capital inflows led to the buildout of network capacity that far exceeded actual demand. When the bubble burst, 85 to 95 percent of the fiber optic cables laid remained unused, and dozens of companies went bankrupt. Today, observers see similar patterns in the data center boom: massive capacities are being built, but their full utilization is uncertain. If demand for AI services falls short of expectations, many of these investments could become worthless.

The $500 billion valuation implies extremely optimistic assumptions. Investors who buy at this valuation apparently expect an IPO at a valuation of over $1 trillion within two to three years. This would make OpenAI one of the ten most valuable publicly traded companies worldwide. By comparison, Apple took decades to reach such a valuation and has massive cash flows and an established product range. OpenAI, on the other hand, is incurring heavy losses and is dependent on a single product.

The social and environmental costs of AI expansion are increasingly being discussed. The energy consumption of large language models is considerable. The Stargate project plans to require 10 gigawatts of electricity, equivalent to the energy needs of approximately 7.5 million households. In times of climate crisis, this raises questions about the sustainability of such investments. Furthermore, negative social impacts, such as those resulting from the automation of jobs, could lead to political backlash.

Scenarios between breakthrough, stagnation and correction

The future development of OpenAI and the broader AI industry can be outlined along several scenarios. In the optimistic scenario, OpenAI succeeds in achieving its ambitious growth targets. This would require several conditions to be met: Technological development continues and new model generations offer substantial improvements. The conversion rate of paying users increases significantly, possibly to 15 to 20 percent, which would result in 120 to 160 million paying subscribers. New revenue streams such as advertising, e-commerce, and high-priced enterprise products are successfully developed and contribute significantly to overall revenue. Inference costs decrease significantly due to technological advances and increased competition in the chip market. In this scenario, OpenAI would become profitable and could go public at a valuation exceeding one trillion dollars.

In the moderate scenario, OpenAI continues to grow but falls short of its most ambitious targets. Revenue might reach $40 billion to $60 billion by 2028 instead of $100 billion, which would still represent exceptional growth. However, profitability remains elusive as costs keep pace with growth. OpenAI would have to rethink its infrastructure plans and possibly renegotiate some contracts. Its valuation would be adjusted, perhaps to $200 billion to $300 billion. An IPO would still be possible, but at more modest valuations. In this scenario, the AI ​​market establishes itself as an oligopoly with several large players competing for market share.

In the pessimistic scenario, OpenAI faces significant growth obstacles. Technological development slows, and new models don't offer sufficient added value over existing solutions. Competitors like Google and Anthropic gain market share. The conversion rate stagnates at low single-digit percentages. At the same time, costs remain high or even continue to rise. In this scenario, OpenAI could struggle to close further capital rounds at attractive valuations. The company would have to drastically reduce its spending and potentially sell assets. The extensive infrastructure commitments would become an existential burden. This scenario could lead to a broader correction across the entire AI sector, similar to the bursting of the dot-com bubble.

A disruptive scenario would be the commercialization of fundamentally more efficient AI architectures. If approaches like those demonstrated by DeepSeek were to find wider adoption, this could fundamentally change the industry's cost structure. In this case, the massive investments in traditional scaling would lose value. OpenAI would have to adapt its strategy and could lose its lead in the process. At the same time, this would accelerate the democratization of AI and allow more competitors to enter the market.

Another key element is the development of AI agents capable of autonomously performing complex tasks. If reliable agents can be developed that act as virtual employees and enable companies to achieve significant productivity gains, this could usher in a new phase of growth. OpenAI is positioning itself for this market, but the technological challenges are considerable. Current AI systems are prone to hallucinations and errors, which limits their reliability for critical business processes.

Regulatory developments will also play a key role. Governments in the US, Europe, and China are developing different approaches to AI regulation. Stricter regulations could slow innovation but also foster greater trust and broader acceptance. Conversely, a regulatory vacuum could lead to abuse and societal disruption, ultimately leading to more stringent interventions.

The geopolitical dimension is gaining importance. The AI ​​competition between the US and China is increasingly perceived as a strategic conflict. Export controls, investment restrictions, and government support programs could significantly influence the competitive dynamics. The Stargate project is explicitly designed as a contribution to American technological leadership.

Between visionary ambition and economic disillusionment

OpenAI's plan to increase revenue from $13 billion to $100 billion within three years represents one of the most ambitious growth plans in the history of the technology industry. The analysis shows that while this plan is not impossible, it would require a number of favorable conditions, the simultaneous occurrence of which must be considered unlikely.

OpenAI's strengths are undeniable. The company boasts technological leadership in large-scale language models, a strong brand, and a massive user base. ChatGPT has established itself as a synonym for generative AI, much like Google is for internet search. Partnerships with Microsoft and Oracle ensure access to essential infrastructure resources. The company's capital base has been strengthened through several rounds of financing.

At the same time, the challenges are immense. The low conversion rate of paying users, the high and ever-increasing development costs, intensified competition, and structural profitability problems pose significant obstacles. The infrastructure commitments entered into far exceed foreseeable revenues, creating enormous pressure to succeed.

There are several implications for policymakers. First, the massive government support for AI infrastructure should be critically examined. The Stargate project may have symbolic value, but its economic viability is questionable when private investors risk hundreds of billions without a robust business case. Second, regulatory frameworks should be developed that enable innovation while simultaneously addressing risks. Third, the energy question must be resolved: The massive power demand of AI data centers conflicts with climate goals and requires coordinated responses.

For business leaders, this development means that AI investments should be approached strategically, but without excessive expectations. The productivity gains from AI are real, but they will occur gradually and require significant organizational adjustments. Companies should experiment, but not rely on immature technologies to build their business models.

For investors, the question of appropriate valuation arises. The current valuation of $500 billion only seems justified if OpenAI not only achieves but exceeds its growth targets and simultaneously finds a path to profitability. The risk-return ratio is extremely unfavorable for late investors. Early investors, however, who entered at significantly lower valuations, can realize significant gains even with moderate success.

The long-term importance of OpenAI and broader AI development for the global economy should not be underestimated, regardless of whether the company achieves its specific revenue targets. Large language models will transform parts of knowledge work and enable significant productivity gains. The question is not whether this transformation will happen, but how quickly and which companies will benefit from it.

History teaches us that technological revolutions are often accompanied by financial excesses. The railroad, electricity, automobile, and internet revolutions all saw phases of massive overinvestment followed by painful corrections. Yet these technologies ultimately proved transformative. The investors who profited the most were often not those who built the infrastructure, but those who used that infrastructure to develop innovative business models.

OpenAI is at a turning point. The company must prove that it can not only develop impressive technology but also translate it into a profitable business model. The next two to three years will be crucial. If OpenAI fails to achieve its goals, the repercussions will reach far beyond the company and shake up the entire AI sector. If it succeeds, however, it would rewrite the rules of corporate growth and potentially mark the beginning of a new era in business history.

The key finding of this analysis is that OpenAI needs new scaling laws, not only for the performance of its AI models, but above all for its own business model. The laws of physics and mathematics that govern the training of neural networks are one challenge. The laws of economics and the market, which determine how a company can grow sustainably and become profitable, are at least as significant. OpenAI must master both to realize its vision.

☑️ Our business language is English or German

☑️ NEW: Correspondence in your national language!

Konrad Wolfenstein

I would be happy to serve you and my team as a personal advisor.

You can contact me by filling out the contact form or simply call me on +49 89 89 674 804 (Munich) . My email address is: wolfenstein ∂ xpert.digital

I'm looking forward to our joint project.

☑️ SME support in strategy, consulting, planning and implementation

☑️ Creation or realignment of the digital strategy and digitalization

☑️ Expansion and optimization of international sales processes

☑️ Global & Digital B2B trading platforms

☑️ Pioneer Business Development / Marketing / PR / Trade Fairs

Benefit from Xpert.Digital's extensive, fivefold expertise in a comprehensive service package | R&D, XR, PR & Digital Visibility Optimization - Image: Xpert.Digital

Xpert.Digital has in-depth knowledge of various industries. This allows us to develop tailor-made strategies that are tailored precisely to the requirements and challenges of your specific market segment. By continually analyzing market trends and following industry developments, we can act with foresight and offer innovative solutions. Through the combination of experience and knowledge, we generate added value and give our customers a decisive competitive advantage.

More about it here:

• Use the 5x expertise of Xpert.Digital in one package - starting at just €500/month