Kathy Wood, Investors Letter – “The Journey From Monetary Policy Shock To An Innovation-Led Economic Boom”

NVIDIA suddenly closed with a 5% plunge ahead of the $1000 mark.

Cathie Wood, who has continued to express her negative views on NVIDIA, posted an investor letter on the ARK Invest website on March 7, and I’m posting a full summary just for you to read.

It’s interesting to compare it to Cisco in the 1990s. I recommend reading it.

===========

• Adjusting the Bloomberg Communities Index to CPI reveals that current inflationary pressures are illusory (top chart)
• Nevertheless, the Fed is still fighting the “war on inflation,” which ended in 2008. It’s like fighting illusions. So the Fed’s mistakes will soon come as a boomerang (commercial real estate, regional banking, etc.)
• S&P 500 gross profit margin was 34.8% over the past 5 years, but fell to 34.6% in 4Q 2022 and 33.5% in 4Q 2023
• If margins weaken in the future, companies will be laid off faster, and prices will fall if the economy shrinks. Inventory needs to be adjusted more, too. Gross profit margin was 30.1% in 2009 and needs to be adjusted to this level
• Wall Street insists on a soft landing, but already sees a drop in fourth-quarter revenue from global players. 3M (-1.8%), UPS (-7.8%), Kraft-Hynes (-7.1%), Exxon Mobil (-12.3%), Thermo Fisher (-4.9%), Home Depot (-2.9%, Cisco (-5.9%), Texas Instruments (-12.7%) and more
• Europe, Britain, Japan, and China are already facing a recession that may have entered a recession. If this happens, corporate decision-makers are likely to adjust AI-related CAPEX plans
• A moment in the history of technology was the moment AOL connected e-mail to the Internet in 1993. Investors are jumping in with similar imaginations as they watch ChatGPT
• But now we need a reality check, because going forward, winner and loser will be sharply divided, and companies will start taking a strategic approach
• In our ARK view, management will face some important decisions.

1) We need to understand the competition between cloud companies and AI companies

2) It also requires organically integrating and mapping specific data from each department. This is time-consuming and difficult. Then, if there is a problem with margins, management will feel pressured. Decision-making is bound to be delayed

• That was exactly the case with Cisco. In the three years to March 9, 1994, Cisco rose 31 times from $0.07 to $2.24 (equivalent). Cisco’s routers and switches were essential for the worldwide expansion of the Internet
• As a result, the capital market began to drive money to Cisco’s competitors. Systems and equipment were inferior to Cisco
• Eventually, Cisco plunged 51% four months later by July 15, 1994. As companies reviewed and reorganized their Cappex plans, concerns about slowing sales spread
• Of course, since then, Cisco has reported a further 73 times surge to the peak of the Internet bubble in 2000, but it should not be forgotten that there have been such twists in between
• We think Nvidia is in a similar situation right now, just nine years after Feb. 8, 2015, when analysts became aware of how deep learning would benefit from the AI revolution and GPUs, Nvidia is up 117 times
• Inventory adjustments began in October 2018 after Crypto Winter hit, when Nvidia plunged 56% in three months and rose 23 times again over the next five years
• After ChatGPT became public in November 2022, Nvidia is having another record quarter. Cloud companies, internet companies, and well-funded startups are doubling and tripling GPU orders
• Nvidia, however, admitted in its guidance that GPU lead times fell from 8-11 months to 3-4 months, implying that supply is starting to exceed demand
• If software demand does not justify this overinvestment in GPUs, it would not be a surprise if CAPEX slows down. Cloud companies account for more than half of Nvidia’s data center sales
• In the long run, there is a possibility that competition will intensify, unlike the past Cisco case. This is because AMD has been chasing after each other, but major Nvidia customers – Tesla and cloud companies – are designing their own AI chips
• It was said that it would take 80 years for AGI, called General AI, to arrive in 2019, but this has been shortened to 8 years in a few years. Nothing is impossible
• When the cyclical correction is completed, AI will fly again. There is no doubt that robots, energy storage, blockchain, and genetic improvement will be fused to achieve an explosive rise again
• The consensus is that GDP growth will slow from 3% to 2.6% on average over the last century, but we forecast growth of 6% to 8% or more. If the five innovation platforms and 14 technologies we are seeing fully bloom over the next seven years, the market capitalization of the stock market could grow at an annualized rate of 40%.