OpenAI is a high-stakes bet on ever-bullish markets
This week, The Wall Street Journal (WSJ) lifted the lid on the internal financial projections of OpenAI and Anthropic – the companies at the centre of the artificial intelligence (AI) boom.
By way of background, Dario Amodei, once a senior leader at OpenAI, parted ways with Sam Altman in 2021 amid strategic and perhaps personal differences and launched Anthropic that same year, seeded it with a $124 million Series A. The debut of ChatGPT in late 2022 blindsided the Amodei and handed OpenAI 100 million users overnight, along with an 18-month revenue lead. Anthropic pivoted sharply to enterprise-grade deployments of its Claude models, a bet that has now propelled its private-market valuation to $183 billion – still trailing OpenAI’s $500 billion mark, but closing the gap through disciplined business to business (B2B) focus.
Indeed, Anthropic aims to make its flagship Claude models the preferred AI for stickier business clients and most recently signed professional services firm Cognizant Technology Solutions as a customer in one of its biggest deals yet.
Anthropic is guiding toward breakeven in 2028, while OpenAI is staring down a projected $74 billion operating loss that year, equal to roughly 75 per cent of its anticipated revenue. In other words, OpenAI expects to generate $100 billion in revenue in 2028 – up from $20 billion this year – but will lose or spend all of that, plus another $74 billion.
The conclusion I reach is that OpenAI’s success and maybe its survival hinge on successfully continuing to raise a lot of capital. And that strategy is dependent, in no small part, on public-market sentiment remaining euphoric.
The numbers
The Wall Street Journal (WSJ) got its hands on OpenAI’s latest investor deck, which stated:
“OpenAI forecasts its operating losses that year [2028] to swell to about $74 billion – or roughly three-fourths of revenue – thanks to ballooning spending on computing costs.”
The WSJ also reports “OpenAI expects thinner margins than Anthropic from its sales for the next five years. Yet it is investing far more in the chips and data centres needed to build its AI technology, and doling out more stock-based compensation to attract top researchers.”
Compared to Anthropic’s financials, OpenAI expects to consume 14 times more cash before achieving profitability in 2030. This year, OpenAI says it will burn $9 billion against $13 billion in sales (although the $13 billion figure conflicts with Sam Altman’s recent interview on X, where he claimed OpenAI’s revenue would be substantially more than $13 billion this year).
Anthropic, by contrast, burns $3 billion on $4.2 billion of revenue this year. The numbers then diverge materially.
Table 1. Comparison of OpenAI and Anthropic cash burn/revenue
|
Year |
OpenAI Cash Burn / Revenue |
Anthropic Cash Burn / Revenue |
|
2026 |
57% |
33% |
|
2027 |
57% |
9% |
|
2028 |
75% |
Breakeven |
Source: Company projections via WSJ
The driver of the difference is the computation, a.k.a. ‘compute’. OpenAI is pre-committing to $1.4 trillion in cloud and chip contracts over eight years – agreements signed after these forecasts were drafted, meaning the actual 2028 loss could be materially higher than $74 billion.
As Sam Altman posted on X: “We believe the risk to OpenAI of not having enough computing power is more significant and more likely than the risk of having too much.”
In other words, we should expect over-investment by OpenAI. It’s a feature. OpenAI is setting aside significantly more computing capacity for future AI research than Anthropic is. In fact, the company is stockpiling $100 billion for moonshots in video (Sora), browsers (Atlas), consumer hardware, e-commerce, and humanoid robotics.
So here it is in black and white; OpenAI has no path to self-funding until and possibly after 2030. Therefore, every dollar of capital expenditure (capex) beyond current cash reserves requires external capital – equity rounds, structured debt, or prepaid cloud commitments from Microsoft and others.
And here’s the WSJ’s take: “The strategy requires near-constant fundraising to keep the startup alive and could backfire if markets cool on the technology or its near-term profitability.”
In mid-October, the Nasdaq wobbled, falling 3.5 per cent in a single day amid AI capex fears. If shifting sentiment drives the S&P 500’s forward price-to-earnings (P/Es) to compress, or if 10-year yields spike, causing the cost of OpenAI’s next $20–$30 billion raise to jump overnight, OpenAI’s valuation – and those of its listed peers – could fall punitively.
Anthropic
Anthropic’s playbook is different. It’s growing costs in step with revenue, prioritising enterprise, and avoiding compute-intensive modalities. Roughly 80 per cent of the company’s revenue already comes from corporate customers who value Claude’s coding competency. Compute-intensive tasks, such as image and video generation, have been de-emphasised. The result is a cash burn rate that falls to nine per cent of revenue by 2027 and a 2028 breakeven.
The good and the…not so good.
- OpenAI
- Good: First-mover scale, perceived multitrillion-dollar total addressable market.
- Not so Good: Requires perpetual positive market sentiment and excellent execution on $1.4 trillion of commitments. A failed product launch or a macro shock could trigger a funding ‘down round’, with implications for the entire AI theme.
- Anthropic
- Good: Capital-light path to profitability; lower execution risk than OpenAI thanks to stickier business customers and less reliance on compute-heavy consumer models.
- Not so Good: Cedes consumer mindshare and multimodal leadership; caps long-term valuation ceiling unless enterprise pricing power proves durable.
My thoughts
OpenAI is not so much a software company as it is a compute arbitrage bet pitched as a consumer app. Its Profit and Loss (P&L) only works if it’s graphic processing unit (GPU) utilisation stays very high, inference margins keep expanding, and investors remain willing to fund massive negative operating margins for another five years.
There’s also the explicit assumption that market sentiment stays red-hot through to 2030.
If you believe the artificial intelligence (AI) hype cycle has another half-decade of runway, OpenAI’s aggression is rational. If you suspect there could be speed humps along the way, the revelation of a $74 billion 2028 loss may be just the first.
Roger Montgomery is the Founder and Chairman of Montgomery Investment Management. Roger has over three decades of experience in funds management and related activities, including equities analysis, equity and derivatives strategy, trading and stockbroking. Prior to establishing Montgomery, Roger held positions at Ord Minnett Jardine Fleming, BT (Australia) Limited and Merrill Lynch.
He is also author of best-selling investment guide-book for the stock market, Value.able – how to value the best stocks and buy them for less than they are worth.
Roger appears regularly on television and radio, and in the press, including ABC radio and TV, The Australian and Ausbiz. View upcoming media appearances.
This post was contributed by a representative of Montgomery Investment Management Pty Limited (AFSL No. 354564). The principal purpose of this post is to provide factual information and not provide financial product advice. Additionally, the information provided is not intended to provide any recommendation or opinion about any financial product. Any commentary and statements of opinion however may contain general advice only that is prepared without taking into account your personal objectives, financial circumstances or needs. Because of this, before acting on any of the information provided, you should always consider its appropriateness in light of your personal objectives, financial circumstances and needs and should consider seeking independent advice from a financial advisor if necessary before making any decisions. This post specifically excludes personal advice.
