By ByteBot
OpenAI is in talks to raise $100 billion at an $830 billion valuation. That’s bigger than Walmart. And according to financial documents reviewed by Fortune, the company behind ChatGPT expects to lose $74 billion in 2028 alone. Not revenue. Losses. OpenAI projects massive annual deficits through 2028 before magically turning profitable by 2030. This isn’t sustainable innovation. It’s bubble economics. And for developers depending on OpenAI’s APIs, the bill is coming due.
The Brutal Financial Reality
The first half of 2025 tells the story. OpenAI generated $4.3 billion in revenue but burned through $2.5 billion in cash. That’s a 70% burn rate – for every dollar earned, they lost seventy cents. Full-year projections look worse: $13 billion revenue, $9 billion cash burn.
Where’s the money going? R&D consumed $6.7 billion in six months. Stock compensation for top AI talent hit $2.5 billion. Compute and talent costs are eating 75% of revenue. The long-term outlook gets wilder: OpenAI expects $200 billion annual revenue by 2030 – a 16× increase – and cash flow positive by 2029, right after losing $74 billion in 2028.
If you believe that timeline, I have dotcom stocks from 1999 to sell you.
Is This a Bubble?
According to Bank of America’s October 2025 survey, 54% of fund managers think AI equities are in a bubble. AI is now the top “tail risk” for investors – bigger than inflation or geopolitical threats. The share of the economy devoted to AI investment exceeds what we saw at the dotcom bubble peak.
But ChatGPT has 700 million weekly active users. That’s a real product, not speculation. Ten percent of businesses now use AI, up from 3.7% in September 2023. Current AI leaders trade at 26× forward earnings – high, but nowhere near the 70× at the dotcom peak.
The truth? Both sides are right. As MIT Technology Review analyzes, AI will transform the economy, but that doesn’t validate current valuations. The internet DID change everything – the dotcom bulls were correct. They were also spectacularly wrong about which companies would survive. Amazon emerged. Pets.com didn’t.
The question isn’t whether AI is real. It’s whether OpenAI at $830 billion is Amazon or Webvan.
The Infrastructure Money Pit
AI is genuinely expensive. OpenAI spent $3 billion on training compute and $1.8 billion on inference. GPT-4’s inference bill in 2024 hit $2.3 billion – 15× its training cost. Inference typically accounts for 80-90% of total compute costs over a model’s lifecycle.
The infrastructure arms race makes it worse. According to McKinsey, Meta is spending $72 billion on AI infrastructure in 2025. Microsoft: $80 billion. Amazon: $100 billion. Alphabet: $75 billion. Combined: $327 billion in a single year.
But here’s the flaw: DeepSeek proved in January 2025 that a frontier-class model can be trained for $5.6 million. Not billion. Million. Inference costs are dropping 10× annually. If competitors deliver similar quality at 1% of the cost, what’s OpenAI’s moat?
Real spending doesn’t guarantee reasonable returns.
What This Means for Developers
This isn’t just VC drama. It impacts every developer using AI tools. GitHub Copilot loses $20 per user monthly at its $10 price point because compute costs $30 per user. That math doesn’t work. The average organization now spends $85,000 monthly on AI tools, up from $63,000 in 2024. A chatbot with 1,000 daily users costs $13,000 to $40,000 monthly. And 71% of companies struggle to monetize AI effectively.
OpenAI will need to raise API prices significantly or cut services. If that $100 billion funding round fails, service disruption becomes real.
Smart developers are hedging. DeepSeek R1 matches OpenAI’s o1 model, runs locally, and carries an MIT license. Llama 3.3 70B delivers GPT-4-level performance. Tools like Ollama and LM Studio make local deployment trivial. Cost? Zero ongoing expenses.
When proprietary AI providers burn billions with no profitability path, betting your stack on their stability looks reckless.
The Verdict: Both Innovation and Bubble
AI will transform software development. Productivity gains are real and permanent. Current spending is unsustainable. OpenAI’s $830 billion valuation is disconnected from fundamentals. Both are true.
Some companies will emerge as winners. Many will fail. OpenAI might be Amazon or Webvan. We don’t know yet.
The smart play? Use AI tools – the gains are real. But don’t build on a single provider. Watch for warning signs: pricing changes, funding struggles, executive departures. Expect consolidation between 2026 and 2028.
We’re watching: Will that $100 billion round close at $830 billion in Q1 2026? Will API prices jump mid-2026? Will open-source models reach GPT-4 parity in 2027? Will those $74 billion losses materialize in 2028?
The AI revolution is real. The valuations are not. Embrace the technology. Hedge the risk. And make sure someone else is holding the bill when it comes due.
