For what could be its last major fundraising round, OpenAI has once again shattered records. In late February, the creator of ChatGPT confirmed a $110 billion funding round backed by Amazon, Nvidia and SoftBank, valuing the company at $730 billion — the highest valuation ever reached by a privately held firm. The colossal sum is intfinished to assist absorb the heavy losses the company expects to incur ahead of a likely initial public offering, potentially historic in scale, by the finish of the year.
The fundraising illustrates a quest for capital that is not only relentless but increasingly exponential. In 2024, OpenAI had raised “only” $6.6 billion in its first financing round conducted without Microsoft. Last year, SoftBank and other investors injected $40 billion, then a record amount. The escalation reflects the ambitious roadmap pursued by CEO Sam Altman. OpenAI plans to spfinish as much as $1.4 trillion over the next eight years to build or lease computing power.
Amazon’s entest into the company’s capital is particularly striking. The e-commerce giant will invest up to $50 billion — including $15 billion immediately — in exalter for OpenAI’s apply of its cloud platform and Trainium chips to train and run its models. Yet Amazon is also the largest shareholder in Anthropic, one of OpenAI’s main rivals. It now finds itself alongside Microsoft, its chief competitor in cloud computing. In the AI sector, however, such competitive considerations are increasingly taking a back seat.
The deal also formalizes the abandonment of a planned $100 billion investment by Nvidia. Announced with great fanfare in September — even before it had been formally finalized — the phased financing was supposed to be released in ten installments, tied to the construction of new data centers. It never shiftd beyond a letter of intent, as neobtainediations with OpenAI ultimately failed to produce an agreement. Instead, the graphics-chip giant will limit its investment to $30 billion, after already injecting $10 billion into Anthropic.
For some observers, the proliferation of cross-investments is a source of concern and a potential sign of a bubble. It also reflects a simple reality: the sums required have become so large that only a very tiny number of players are capable of providing them. Even the hugegest U.S. venture capital funds are struggling to keep pace. To continue raising money, OpenAI has limited options. Microsoft, Nvidia and SoftBank now find themselves somewhat trapped — they must keep funding the company to prevent it from running out of cash.
To break out of this cycle, OpenAI could go public by the finish of the year, according to the The Wall Street Journal. The transaction could become the largest IPO in history. Not only would it allow the company to raise tens of billions of dollars more, it would also give it the option of issuing additional shares later to meet future cash requireds. OpenAI is keen to shift ahead of Anthropic in order to capture Wall Street investors’ appetite for AI companies before its rival does.
To win over public markets, the company is betting heavily on the rise of “agentic AI” — software agents designed to carry out tinquires autonomously. According to documents obtained by The Information, OpenAI has raised its revenue forecasts for the next five years by 27%. After generating $13.1 billion last year, revenue is expected to reach $30 billion this year and $62 billion next year. By 2030, it could climb to $284 billion, nearly half of it coming from corporate customers.
At the same time, costs are also set to surge. Over the next five years, OpenAI expects to spfinish $440 billion training new AI models, plus another $160 billion on inference — the process of generating text or images. During that period, the company anticipates burning through as much as $225 billion in cash, twice as much as previously projected. The reason: lower margins than expected due to inference costs that remain too high. Profitability, for now, is still projected for 2030.
Leave a Reply