My understanding is that the S-1 showed a Q1 loss of xAI of $2.47 billion in Q1. But with the Anthropic Colossus-I rental agreement at $1.25B/month or $3.75B/quarter, xAI should now be net-neutral to cash-flow positive.
If Colossus-II rents networked GB-200s, that could be up to +$47B/year at $9/hour/GB-200 for 555,000 GB-200s. For reference, current rental rates are $10-$27/hour for the same hardware. With Anthropic at a 55% month-over-month growth rate (implying a $150B/year run rate by August, or, more likely, sometime in late 2026), it seems very possible that xAI could be highly profitable as the only available compute resource.
I'm not saying +$47B Colossus-II deal will happen, but even a small fraction of that remains highly material to xAI economics. xAI is likely already cashflow neutral. (Where am I wrong?)
the xAi revenue comes from renting their infrastructure to their biggest competitor. Anthropic is going to IPO for aprox 1/2 the valuation, is profitable, and can cancel the contract with 90 days notice.
My understanding is that the S-1 showed a Q1 loss of xAI of $2.47 billion in Q1. But with the Anthropic Colossus-I rental agreement at $1.25B/month or $3.75B/quarter, xAI should now be net-neutral to cash-flow positive.
If Colossus-II rents networked GB-200s, that could be up to +$47B/year at $9/hour/GB-200 for 555,000 GB-200s. For reference, current rental rates are $10-$27/hour for the same hardware. With Anthropic at a 55% month-over-month growth rate (implying a $150B/year run rate by August, or, more likely, sometime in late 2026), it seems very possible that xAI could be highly profitable as the only available compute resource.
I'm not saying +$47B Colossus-II deal will happen, but even a small fraction of that remains highly material to xAI economics. xAI is likely already cashflow neutral. (Where am I wrong?)