HOLT TMT The Durability Debate - US Tech at Midyear, Inflection Opportunities & META

Research evidence excerpt

HOLT TMT The Durability Debate - US Tech at Midyear, Inflection Opportunities & META

May 2003 – June 2004 (14 months). Driver: the recovery off a deeply depressed

base. Ending: benign – easy comparisons rolled off and the group chopped

sideways through 2004, with semis correcting harder.

May 2009 – June 2010 (14 months). Driver: the rebound from the Great

Recession drawdown. Ending: benign – faded into the spring of 2010 correction

while estimates kept improving into 2011.

May 2017– September 2018 (17 months). Driver: tax reform, broad global

growth, memory cycle strength, and cloud CapEx. Ending: the most abrupt

deterioration – breadth broke weeks before Q4 2018’s -20% drawdown, as the

market began pricing a cloud CapEx deceleration for 2019, DRAM pricing rolled

over, and the Fed kept tightening.

August 2020 – February 2022 (19 months). Driver: COVID/WFH demand pull-

forward. Ending: price led again – the Nasdaq peaked in November 2021, three

months before breadth broke; the 2022 bear market followed.

May 2025 – Present (15 months and counting). Driver: the AI CapEx build-out,

rebounding PMIs and industrial activity, and surprising software estimate strength.

Ending: open – but note the benign endings were rebounds off depressed bases.

Today’s setup is a boom, and in booms the tell was price, not estimates.

Two lessons stand out from this history. First, every prior streak ended within 14 to 19

months – at 15 months, the current run is entering the zone where every predecessor

rolled over. Second, in the two episodes that ended badly (1999-2000 and 2020-2022),

price peaked months before estimates did – strong estimates tell you a rally is supported;

they will not tell you when it ends.

Can the streak last? The estimate surge is real and has rightly powered Tech’s