SF Compute argues credit, not chips, is AI's bottleneck
Evan Conrad of San Francisco Compute argues that independent, physically settled forward contracts can unlock GPU financing, a model that depends on refusing the profits of becoming a cloud.
Loading articles…
The complete subject coverage could not load. Published articles remain below.
Crusoe's CEO models AI infrastructure on oil majors, not winner-take-all
Chase Lochmiller, CEO of Crusoe, explains how a vertically integrated model, from electrons to tokens, shapes his view of competition, leadership scaling, and the public backlash against data centers.
Modal's CEO: inference tokens commoditize, platform breadth endures
Eric, founder and CEO of Modal, argues that inference serving is heading toward commodity pricing while durable advantage lies in multi-product infrastructure, even as compute stays chaotic and capital-hungry.
Baker: forward compute prices exceed spot while enterprise AI stays early
Investor Gavin Baker, reacting to Anthropic's S-1, argues demand is outrunning compute supply, while counterpoints on customer concentration and training-heavy compute test how far that optimism can run.
Anthropic's $1.48 trillion valuation rests entirely on sustained growth
An investor valuing Anthropic's newly filed prospectus finds the same data supports both bull and bear readings, while a founder-controlled governance structure complicates any price he would pay.
Google will forgo cloud revenue to stay at AI's frontier
Alphabet CEO Sundar Pichai argues frontier models are non-negotiable even as capacity shortages, imperfect alignment, and overbuilding risk test how Google allocates scarce compute.
Intelligence hits physical limits, so own what it can't replace
Cliff, an investor with a concentrated Carvana portfolio, argues intelligence is bounded by chaotic, physical reality, so AI demand may undershoot the buildout while data, attention, and scale gain value.