“Our LT Weekly Update is a comprehensive weekly analysis aimed at helping our key corporate and investor clients cut through the noise and stay in front of what we view as the most important themes and developments driving the digital economy.”
- Leslie Mallon, Head of LionTree Public Markets

Geopolitical concerns spiked even higher this week, sharply dragging down the major indices (the S&P 500 fell -1.9% and Nasdaq fell -2.1%). The war’s implication on oil dominated headlines…Saudi Arabia’s warning that crude could exceed $180/bbl if disruptions persist past April caused a bit of a stir. Regarding the Fed rate meeting, as expected, current rates were maintained, but the hawkish bent concerned the Street, plus PPI was hotter than expected.

Regarding key sector themes/developments/upgrades, we focused on the below:

Nvidia Materially Ups Its Demand Targets As The “OpenClaw Revolution” Begins

Nvida’s GDC and financial analyst events were highlights this week. Most incremental was that the Co substantially took up its Blackwell and Vera Rubin “high confidence” demand forecasts from $500bn in 2026 to $1 trillion+ through 2027. AND on top of that will be demand from new products like Groq, which is coming in Q3. At the events, the “OpenClaw revolution” was a big focus of the conversation (it essentially is the operating system for agentic computers) and Huang contends that every software company now needs to have an “OpenClaw Strategy”. He also tried to ease concerns about the future health of the software sector, arguing that these companies will dramatically create more value with AI.

In the future, AI agents will be producing tokens 24/7…and eventually with Physical AI (which is much larger than digital AI), “the world is going to produce tokens every single day, continuously. It will not stop.” The cost per token will go down while token “smartness” will keep going up while they will increase the throughput. He is not worried about future demand.

Overall, NVIDIA believes its architecture is the only platform capable of running every AI domain across all AI models (language, biology, computer graphics, vision, speech, proteins, chemicals, robotics) and environments (edge or cloud), positioning it as the lowest-cost, highest-confidence infrastructure for AI investments. Mgmt is confident in Nvidia’s lead as Moore’s Law has “run out of steam,” with NVIDIA’s accelerated computing, specifically Grace Blackwell and Vera Rubin, delivering 35x to 50x performance per watt improvement over Hopper.

While all of this sounded exciting, one thing that investors would like to see is proof points of a revenue growth inflection at its key customers as a result of the heavy levels of AI investment…