김상욱 retweetledi
Based on the news, $NVDA should have rallied on stellar earnings last wk with software hit on Citrini outlining the worst case for AI driving unemployment followed by $XYZ (+19.7% last wk) cutting 40% of its workforce stating AI improvements as the reason.
Instead, Nvidia closed down 6.7% and $IGV (software ETF) closed up 1.0% while the S&P was down just 0.4% for the wk. But to be fair, there was mixed but generally positive performance from some of the important software names that reported earnings during the week. $WDAY was -2.9% last week, $CRM +5.2%, $SNOW -2.4%, $ADSK +8.3% and $INTU +7.5%.
This bounce also occurred despite further financial “cockroaches” to use the CEO of $JPM (down 3.4% last wk) Jamie Dimon’s phrase from October of last year following the failures of Tricolor and First Brands. MFS, which collapsed on Friday, had backing from firms including $BCS (-6.6% last wk), $APO (-12.6%), $JEF (-16.9%) and $TPG (-2.4%.) The dividend cuts by private credit companies $FSK (-16.7% last wk) and $MFIC (-13.1%) impacted the space and their associated parent companies $KKR (-13.3% last wk) and Apollo. Given I view credit as the lifeblood of the economy, this bothers me.
I do not believe we have seen the ultimate lows in software (and potentially disintermediated sectors including financials) for one simple reason, not enough time has passed. Marc Andreessen wrote his “software is eating the world” op-ed for the WSJ in August of 2011. This optimism has generally been justified ever since. But given software stocks peaked in late October of last year as agentic AI became a threat, I doubt 4 months is enough time to truly figure out what companies are in trouble and which are not after at least 14 years of optimism.
Having said that, I believe that the 1% bounce in the $IGV last week despite the above big picture reasons that could have driven it down further, makes a continuation of that bounce potentially likely.
Not all software names will be roadkill as corporations adopt AI agents. Data is the lifeblood of any company and needs to be protected so ultimately the database and security companies should fare better. In addition software for highly regulated industries and those where deterministic outcomes are needed (eg the plane must not crash) are more insulated and could benefit (AI testing the plane design.) But remember, AI agents can communicate with the underlying data and do not need the elaborate GUI interfaces that humans need from many of the software companies in the application layer.
As for $NVDA, despite the market clearly disagreeing last week, I continue to believe that at a 23x CY26 PE versus the S&P at 23x, this continue to be a great risk versus reward for those that still have optimism on AI. Revenue growth accelerated from 56% y/y to 73% over the past two quarters with a forecast for it to accelerate again in the April quarter.
While the memory of massively missed industry forecasts should sit in the back of everyone’s mind from the dotcom bust, I believe we are still in the early stages of the capex necessary to deal with the new workloads generated by recent agentic models. Token generation since the end of last year according to OpenRouter, have more than doubled year-to-date. While others such as $AMD or Cerebras will gain share, with use of custom ASICs by the big hyperscalers, Nvidia remains the gold standard.
$GOOGL is my preferred play on this trend in terms of the hyperscalers with the complete vertical stack as I have outlined in prior posts. $CSCO should benefit from networking the datacenters and the associated corporate network upgrades regardless of who wins. $AAPL should finally have AI capabilities for their 2.5B installed base of devices by year-end and is benefitting from free riding on the enormous capex investments of others.
But I also believe shorts continue to be important as well. Regardless of public commentary, in terms of actions, Nvidia invested only $30B in the latest OpenAI funding round versus the $100B Letter of Intent announced in September of last year. The larger issue for me is OpenAI increased their revenue forecasts by 27% by 2030 but doubled their cash burn to nearly $220B through 2029. This will obviously impact financial services companies who made poor lending decisions to datacenter providers. $CRWV (-10.9% last week) reported a greater Q1 loss than expected while forecasting capex of $30-35B in 2026 vs sales expectations of $12-13B.
As we approach the appointment of a new Fed chairman in mid-May, the market will start to look forward to rate cuts, whether they should happen or not. Remember the Fed did not raise rates despite the highest inflation in roughly 40 years in 2021 and the S&P finished up 27%.
I believe U should also continue to get MESI in terms of positioning. (And yes I believe Messi’s eighth Ballon d'Or in 2023 arguably gives him the edge in GOAT status over Pele my sentimental favorite.) The S&P sectors of Utilities, Materials, Energy, Staples and Industrials are all up between 11-24% year-to-date. These sectors are asset heavy with low obsolescence risk from AI with potential efficiency improvements instead. These sectors, however, are up just 42% on average since the end of 2022 since the launch of ChatGPT and the AI trade versus the 79% advance in the S&P, 117% rise in the Nasdaq and blistering 335% return in the Magnificent 7.
Clearly the reaction to events which will continue to evolve in Iran will color the trading in the market to start the week. But I see violent trading ranges as here to stay within the sectors exposed to AI for the rest of the year. Selectivity will continue to be important as wide dispersions continue.
Best of luck in the week ahead.
English
