Anthropic, it is said, turned its operating profit positive for the second consecutive quarter after adjustments, with a gross margin exceeding 80%.
On Polymarket, this message is immediately turned into a yes-or-no question: Can AI really make money, or does it only look good after stripping out the most expensive costs?
I’ve been in this industry—this trick is something I’ve seen too many times.
“Adjusted operating profit” excludes the cut shared with partners like Amazon, as well as the training costs themselves—training costs are precisely the most cash-intensive part of this industry.
Yes, the product-level margin is high, but whether R&D costs for frontier models will rise alongside their capabilities is something nobody has explained clearly.
This post has spread aggressively: 455 quote-tweets to 219 reposts, which suggests people are arguing rather than simply agreeing. 407,000 views may indicate attention, but it’s not evidence of money flowing.
What this message truly boosts is the narrative of “AI + crypto”—DePIN, GPU compute networks, AI infrastructure tokens. The logic is: “AI can make money → real demand for compute → these tokens can capture the upside.”
The logic itself isn’t necessarily wrong, but there’s a missing link in the middle: whether the demand driven by model growth truly transmits to token holders or node operators—not something that can be automatically cashed out just because “an AI company is making money.”
Utilization rates, real revenue, and contract terms—these three are the standards for whether tokens can actually capture this wave of upside. Social hype is not.
Whether a tweet can change the market depends on whether money truly flows in that direction, not on the number of quote-tweets.
$RENDER #DePIN #AI
On Polymarket, this message is immediately turned into a yes-or-no question: Can AI really make money, or does it only look good after stripping out the most expensive costs?
I’ve been in this industry—this trick is something I’ve seen too many times.
“Adjusted operating profit” excludes the cut shared with partners like Amazon, as well as the training costs themselves—training costs are precisely the most cash-intensive part of this industry.
Yes, the product-level margin is high, but whether R&D costs for frontier models will rise alongside their capabilities is something nobody has explained clearly.
This post has spread aggressively: 455 quote-tweets to 219 reposts, which suggests people are arguing rather than simply agreeing. 407,000 views may indicate attention, but it’s not evidence of money flowing.
What this message truly boosts is the narrative of “AI + crypto”—DePIN, GPU compute networks, AI infrastructure tokens. The logic is: “AI can make money → real demand for compute → these tokens can capture the upside.”
The logic itself isn’t necessarily wrong, but there’s a missing link in the middle: whether the demand driven by model growth truly transmits to token holders or node operators—not something that can be automatically cashed out just because “an AI company is making money.”
Utilization rates, real revenue, and contract terms—these three are the standards for whether tokens can actually capture this wave of upside. Social hype is not.
Whether a tweet can change the market depends on whether money truly flows in that direction, not on the number of quote-tweets.
$RENDER #DePIN #AI