I didn't need to read past the headline to know something was off. Crypto Briefing dropped a bombshell: Anthropic and OpenAI's combined ARR tops $115 billion, closing in on Microsoft. My first reaction wasn't excitement. It was suspicion. The blockchain doesn't lie, but the media outlets covering it often do. And this number? It's not just aggressive. It's mathematically absurd.
Let's run the numbers. Public data from The Information and Bloomberg puts OpenAI's 2024 annualized revenue around $3.7 billion. Anthropic? Roughly $1 billion. Combined, that's $4.7 billion. The article claims $115 billion. That's a 24x discrepancy. Either these companies discovered a revenue printer nobody else knows about, or someone at Crypto Briefing confused a headline with reality.
The gap between reported and real numbers isn't a rounding error. It's a chasm.
Context: The Narrative Machine
Here's the thing about crypto media. They don't report news. They manufacture narratives. The AI hype cycle is the perfect vehicle for this. You take a real trend—AI companies growing fast—and you stretch it until it breaks. $115 billion ARR isn't a data point. It's a marketing hook designed to make you believe AI-native companies are about to dethrone the tech giants.
Microsoft's commercial cloud revenue hit roughly $160 billion in 2024. The article wants you to think two startups with a combined employee count under 5,000 are closing in on that. Let me tell you what that would require. Every single employee at both companies would need to generate over $20 million in annual revenue. For context, Apple generates about $2.5 million per employee. The math doesn't work. It never did.
Core: Deconstructing the Data
I've spent years auditing on-chain metrics and financial claims. This smells like a classic unit error or deliberate obfuscation. The most likely scenario? Someone took OpenAI's projected 2025 revenue—around $11.5 billion—and added Anthropic's projected $4-5 billion, then accidentally (or intentionally) moved the decimal point. Or they conflated total contract value with annual recurring revenue. In crypto, we call this the 'TVL illusion'—showing locked value instead of actual usage.
The real story isn't the $115 billion figure. It's why a media outlet would publish it without verification.
Let me break down what's actually happening. OpenAI's real ARR growth is impressive—roughly 200% year-over-year. Anthropic is growing faster from a smaller base. But here's the nuance the article misses: these companies are burning cash at unprecedented rates. OpenAI's losses in 2024 exceeded $5 billion. Anthropic isn't profitable either. High growth with massive burn isn't the same as sustainable revenue. It's a treadmill.
The article also conveniently merges two competitors into a single 'alliance' to create the illusion of scale. OpenAI and Anthropic are fighting for the same enterprise clients. They're poaching each other's researchers. They have fundamentally different safety philosophies. Combining their ARR is like adding Apple and Samsung's revenue to claim they're 'closing in on Microsoft.' It's technically true. It's also completely meaningless.
Contrarian: The Real Competitive Landscape
Here's what the mainstream narrative gets wrong. Microsoft isn't threatened by OpenAI and Anthropic. Microsoft owns a significant stake in OpenAI. They have exclusive cloud rights. Azure OpenAI Service is one of their fastest-growing products. The real competition isn't 'AI startups vs. Microsoft.' It's 'Microsoft+OpenAI vs. everyone else.'
Anthropic's positioning is interesting—they're the 'safe AI' option, courting enterprises that don't want to be locked into Microsoft's ecosystem. But their revenue is still a rounding error compared to the hyperscalers. Google Cloud's AI revenue alone dwarfs Anthropic's entire business. Amazon's Bedrock is quietly becoming a major player. The article's framing of a two-horse race ignores the actual battlefield.
The contrarian play here isn't betting on AI software companies. It's betting on the infrastructure that makes their growth possible.
When I see inflated AI revenue narratives, I look at what's actually benefiting. GPU manufacturers. Data center operators. Power companies. These are the picks-and-shovels plays. NVIDIA's data center revenue tells you more about AI adoption than any ARR figure from a crypto media outlet. The market's already pricing in AI software growth. It's underpricing the physical constraints on that growth.
Takeaway: Filter the Noise, Find the Signal
I don't know if the $115 billion figure was a mistake or a deliberate fabrication. What I know is this: if you're making investment decisions based on crypto media's AI coverage, you're trading on hopium, not data. The real signals are in enterprise API usage, cloud provider earnings calls, and GPU lead times. Not in sensational headlines designed to drive clicks.
The blockchain doesn't care about your narrative. Neither does the market.
Next time you see a headline that seems too good to be true, do what I do. Check the source. Verify the math. Look for the original data. If it doesn't add up, it's probably not real. And if it is real, it's probably not as impressive as it sounds. The AI revolution is happening. But it's happening at the pace of reality, not the pace of media hype. Trade accordingly.