OpenAI’s horizons are expanding. The latest funding round of the startup behind ChatGPT, unveiled on Monday, raised up to $40 billion from investors including SoftBank Group at a $300 billion valuation – nearly double what it fetched only six months ago. Yet the company’s path to profitability remains opaque.

Boss Sam Altman admittedly has some encouraging-sounding figures. OpenAI is hoping to generate nearly $13 billion of revenue this year, Bloomberg reported citing sources familiar with the matter, about quadruple last year. While cash flow remained negative in 2024, Altman expects this to turn positive in 2029 – when OpenAI expects to make $125 billion in revenue. Although positive operating cash flow doesn’t necessarily translate into net profit, the pathway will reassure SoftBank boss Masayoshi Son.

Yet OpenAI looks some way off generating the revenue to cover its costs. Over 90% of its 500 million users worldwide pay nothing to use the service. In 2025, just under 5% could pay the $20-a-month charge to access the group’s more advanced AI models with fewer limitations on usage and performance, according to people familiar with the situation – generating about $5.5 billion of sales. Another 0.3% could for ChatGPT Pro, a yet-more advanced service, contributing another $3.6 billion. The trouble with this 70% chunk of OpenAI revenue is that expenses may keep rising faster than the top line.

In 2024, the computing costs of training large language models and then running them amounted to $5 billion, the Information reported in October citing an OpenAI financial document, against OpenAI revenue of only $4 billion. The cost of compute could admittedly fall given the rise of cheaper players like China’s DeepSeek. But OpenAI’s aim to train and develop a smarter AI model means it will by the end of the decade still probably spend 60% to 80% of its annual revenue to either train or run models, the Information said.

Altman’s best workaround is to increase the $4 billion of 2025 revenue that should come from businesses. A significant chunk of this “enterprise” revenue comes from the so-called Application Programming Interface (API) arm, where the company sells direct access to their models. Since the API business charges by usage, it should be easier to cover the cost of running models. But intense pricing competition from other providers such as Mistral AI and Anthropic mean this isn’t a slam dunk, either. OpenAI’s market share of enterprise LLMs fell to 34% in 2024 from 50% a year ago, according to Menlo Ventures data.

In other words, to hit his $125 billion revenue target in 2029, Altman needs to not only grow OpenAI’s top line by 10-fold but increase his enterprise sales by even more to replace loss-making consumer business. The conveyer belt of cash from funding rounds may yet mean he can do so. But at the very least, it’s a stretch.

Context News

OpenAI said on March 31 it would raise up to $40 billion in a new funding round led by SoftBank Group at a $300 billion valuation. The Japanese tech investment group said in a statement that it has agreed to fund OpenAI with $10 billion in mid-April and an additional $30 billion in December, contingent on the artificial intelligence firm transitioning to for-profit by the end of the year. SoftBank plans to syndicate out $10 billion of its total investment to other co-investors it did not name. A person familiar with the matter told Reuters that the remainder of the funding would come from Microsoft, Coatue Management, Altimeter Capital and Thrive Capital. If OpenAI’s restructuring fails, SoftBank said its total investment amount in the round would drop to $20 billion. OpenAI had closed a $6.6 billion funding round in October 2024, which valued the company at $157 billion.

Breakingviews

Reuters Breakingviews is the world’s leading source of agenda-setting financial insight. As the Reuters brand for financial commentary, we dissect the big business and economic stories as they break around the world every day. A global team of about 30 correspondents in New York, London, Hong Kong and other major cities provides expert analysis in real time.

More By This Author: