The Anthropic xAI Deal Is An Admission Of Failure
The chart shows US business paid AI subscription growth. Musk's AI company is in last place. He shoved xAI into SpaceX in order to siphon off SpaceX cash flow and IPO proceeds to finance his struggling venture. On its own, no one would lend to xAI given its third tier position in the AI service and its $1 billion per month cash burn.
Now
Anthropic is leasing a huge chunk of xAI data center capacity to
estimate its models. Supposedly worth about $1.25 billion per month. I
say supposedly because contracts can vary from rock solid to as soft as
jello. Indeed, some industry contracts are really just letters of
intent, non-binding expressions of customer interest. In this case both sides can cancel the agreement with only 90 days notice. Again, the Musk
cheerleaders are prematurely celebrating. What this deal implies is that
there's very little demand for xAI's models. So just as X pivoted from
social media platform to AI, it is now pivoting to the AI data center
business (also called neo cloud). This explains in retrospect Musk's
failed OpenAI lawsuit. He's desperate.
Let's
suppose xAI does get $1.25 billion a month. Does this reduce its huge
losses? Perhaps, but probably not as much as one might naively expect. A
lot depends on the cost of electricity, which is the single largest
data center operating expense and often consumes 33% to 50% of revenues.
The other question is whether xAI must buy GPUs to get the money? A
good guess is that the two xAI data centers are not fully populated. In
that case we would see a big uptick in depreciation and other costs.
Are
Neo Clouds doing well in general? Prominent examples include CoreWeave
and Oracle. CoreWeave is one of the most promising of the AI data center
players. It has a $100 billion revenue backlog. But given grid power
constraints, its unlikely to to realize most of that revenue. In 1Q2026,
its operating loss was surprisingly small, but it paid a half billion
in interest. Moreover, its total liabilities were $50.8 billion, mostly
debt and long term leases. A NEO cloud is a highly leveraged business
complicated by the fact that there are only a few buyers. Hence buyers
have strong negotiating power. In CoreWeave's case two customers
comprise 65% of revenue. Lose them and it's kaput. Oracle is struggling
much more than Coreweave. Its stock price has fallen from $328 a share
in August to $200 today. The story is the same: a Neo cloud involves
massive capex and debt, but also big operating expenses in the form of
power and only a few, very large customers that have great negotiating
power.

Comments
Post a Comment