Most conversations I hear is whether ChatGPT (genAI) is conscious. No. Whether we’ll lose our jobs to AI. Not really. Or whether we’ll reach Artificial General Intelligence (AGI) in the next couple years. We won’t. What I don’t hear much about is the financial viability of genAI. Can genAI be profitable? Because it’s certainly not right now. Everything in genAI startups are subsidized by venture capitalist (VC) money. And it’s in the best interest of CEOs like Sam Altman from OpenAI to exaggerate genAI’s future profitability. Sound decisions and realistic expectations be damned. Many of OpenAI’s decisions this year, most clearly Sora 2’s release, makes the most sense through a financial lens.

To be clear, all genAI startups are losing money right now. I’ll focus on OpenAI because they’re the most well known for introducing ChatGPT. Between January and July 2025, OpenAI reported total cash outflow (all company expenses) at $13.5 billion. In that same period, they earned $4.3 billion in revenue. That’s a $9.2 billion loss.

Software companies throwing money into a blackhole is nothing new. Uber lost $25 billion during its first 5 years before becoming profitable. Facebook (Meta) lost $60 billion on Mark Zuckerberg’s failed bet on the ‘Metaverse’. So while OpenAI throwing money into a blackhole is nothing new, they’re doing so with a key difference: genAI fails at economies of scale.

Places like Instagram or Uber rely on millions of users for economies of scale. In contrast, this scale does not happen with genAI. Each prompt to genAI is computationally expensive. As an analogy, Instagram gaining new customers is like adding an additional rail-cart to a train. While a train is initially expensive, it’s fairly cheap to add a new rail-cart once established. OpenAI, however, is basically giving each customer a Tesla truck.

All levels of subscription lose money. From the $20 dollar entry tier to the $200 premium tier loses money. While software giants like Instagram profit from users languishing on their platform, OpenAI loses money for every. single. prompt. Improved efficiency is not destined to drive costs down either. To date, cheaper genAI mostly redirects savings into running models that “think” (aka cost) more to create better responses for users.

Those issues aside, OpenAI is signing massive new contracts seemingly every week. In September 2025, the company signed a $300 billion 5-year contract with Oracle. With payments beginning in 2027, that’s a $60 billion per-year commitment. At the bare minimum, OpenAI is looking at $90 billion in operating costs beginning in 2027. The math is simple. With only 4.3 billion in revenue in the first half of 2025, I fail to imagine them servicing $90+ billion debt within two years. And as of November 2025, OpenAI has $1.4 trillion in financial commitments across the next 8 years. How do they expect to survive financially?

So far I’ve focused solely on OpenAI. But they’re not the only players in the game. In fact, the four largest software companies: Amazon, Microsoft, Google, and Facebook (Meta) are spending $364 billion this year combined. 2025 isn’t even the crescendo of their spending. 2026 and 2027 are looking to be even bigger. To compare, the Apollo Space program cost $313 billion (adjusted for inflation) across 13 years. And yes, these companies are also losing money on genAI. In what public financial information is shown, we know between the four they’ve made $30 billion in revenue. A $334 billion loss in 2025.

Why does it matter if software companies throw money into a furnace?
Because your life is more intertwined with the success and failures these companies experience than you’d like to think. Take 401k retirement funds as an example. The majority of these invest in the S&P 500 Stock Index Fund. The way index funds work is that your money is distributed unevenly across the top 500 US companies. Meaning if I put $100 into the S&P 500 index fund and Microsoft is worth 10% of the total value of the 500 companies, then $10 goes into Microsoft stocks.

And so, Amazon, Microsoft, Meta, Google, and Nvidia combined hold ⅓ of the S&P 500’s total value. I haven’t even mentioned Nvidia yet. They alone hold 8.5% of the S&P 500. Who is Nvidia? Simply put, they’re selling picks and shovels in a gold rush where nobody's even found gold yet. 88% of Nvidia’s revenue in 2024 and 2025 came from GPUs, which are expensive specialized hardware required for genAI. (Seriously, a single commercial GPU costs $70,000 and Microsoft bought half a million of them this year) Should genAI funding dry up, then demand for GPUs will drop and Nvidia’s valuation will plummet. Should Nvidia drop, so will the S&P 500.

Through all this, what is OpenAI trying to achieve? They know their finances better than anyone else. And if my understanding is accurate, there is no profitability in sight. Yet, genAI is here to stay. Does OpenAI believe they will reach profitability, or are they expecting the government to bail them out? What happens if OpenAI fails? Will it be the spark in a massive chain reaction? Get replaced by another ambitious genAI company? Or get bailed out by the government a la 2008?