Now Playing
Valuation Reset, Data Center Gridlock & the Capital Rotation | Ep 1
Audio is available on Spreaker — see link below.
220 unicorns have lost up to 68% of their peak value as AI makes legacy SaaS models obsolete — while billions flow into power, compute, and physical infrastructure. Today's briefing maps the capital rotation reshaping the startup and VC landscape.
Audio is available on Spreaker — see link below.
Over two hundred and twenty companies that were once valued at a billion dollars or more have lost between fifty and sixty-eight percent of their value. Not because markets crashed.
Companies last funded in twenty twenty-one have lost an average of sixty-eight percent of their peak value. Those funded in twenty twenty-two are down fifty-two percent.
While those valuations compress, capital is moving decisively in the opposite direction. Not toward software.
The infrastructure bet comes with a constraint that's now moved from analyst concern to operational reality. Thirty to fifty percent of the one hundred and forty planned U.S. data centers targeting sixteen gigawatts of capacity are facing delays or outright cancellations in twenty twenty-six.
SoftBank's commitment to build Europe's largest AI data center hub in Dunkirk, France, reads differently in that context. A five gigawatt project, with an initial three point one gigawatt phase targeted for twenty thirty-one.
Back to those two hundred and twenty companies. The practical test for most of them is a twelve to twenty-four month adaptation window.
The two things worth tracking closely from here. First, whether the data center delay rate stabilizes or worsens through twenty twenty-six, and whether hyperscalers can offset that with on-site power generation fast enough to keep model training timelines intact.
Chapter summary auto-generated from the verified script. Listen to the full episode for the complete content.