Cathie Wood on the Market's Election Reaction, AI Spending Gap, and Early-Cycle Signals

Wood begins by acknowledging that markets have had a difficult few weeks with considerable drama, and says she will address the election's ramifications, particularly for technologically enabled innovation and AI. She says the discussion will then move through fiscal policy, monetary policy, and economic and market indicators.

Before turning to the data, she says she wants to put everything in perspective by first paying tribute to John Nosta.

Wood describes John Nosta as a dear friend and an ARK supporter who joined Brainstorm almost every week since the firm started in 2014. She says he died this week at a young age, calling him an incredible intellect and a devoted husband and father who leaves behind a wife and three children.

Wood notes that Nosta had just published a book, The Borrowed Mind: Reclaiming Human Thought in the Age of AI, and had planned to launch it with conferences alongside Brian Romel, who also joins Brainstorm. She says the book draws on his background in medicine and psychology to examine the AI revolution and its meaning for individuals. Although she has not finished it, she says she was fascinated by it and encourages viewers to support his family by buying the book, calling it the least she can do for a man who believed in ARK from the beginning.

Wood presents a chart of midterm elections from 1958 to 2022, excluding a forecast for the coming year. She says that whether the administration is Republican or Democratic, the market is generally up in the year after a midterm election. She attributes this partly to the pre-election anxiety that usually builds ahead of a midterm and the large turnover of seats held by the incumbent party, which she says happened again this time, although historically similar turnovers have occurred before, as in the Obama administration.

She also compares the current period to the 1980s and 1990s, which she describes as a long strong market that she believes could be seen as lasting 20 to 25 years if her view of the technology revolution is correct, interrupted occasionally by corrections. Looking at those decades and Trump's last term, she says most years were followed by gains, the exception being 1986 in the chart she showed.

Wood acknowledges the 1987 case. After a year in which investors took comfort in portfolio insurance, the market fell 22% in one day, but she says that even in the year after that midterm, the market rose modestly. On average, she says the market was up roughly 13 to 14% in the following year, and she hopes that past patterns will serve as a guide.

Wood says the results may be somewhat more balanced than they first appeared on the surface. She attributes voter anger to the Iran war and the inflation that accompanied it, especially oil inflation and higher gas prices, which she says angered people across Democratic, Republican, and independent groups.

She acknowledges a mistake in her team's thinking: they had assumed the Iran war would end quickly. Once it did not, she says the market grew skittish and began sniffing out a defeat for President Trump and the Republicans from around May and June. She notes that many AI stocks peaked around that time and that uncertainty was gradually priced in. Defense stocks, she says, were hit hard because the expectation is that increased health care spending would have to be funded from elsewhere, given the size of the deficit and debt. She adds that additional taxes on banks and energy companies, which have benefited from the environment, could also be considered.

Wood says she was struck that election advertising in Florida, where she lives, focused heavily on data centers, even though the state has no large data center. She says this made her see the issue from a very different angle.

She admits that companies racing into the AI age and building data centers without sufficient attention made a mistake. They were not sensitive to local politics or to how communities might receive them, assuming they would create well-paying jobs and bring tax revenue that would leave the community better off. She acknowledges that headlines about rising electricity prices were a real concern, and she says some of the criticism was political. She also addresses water use, saying that Meta and Amazon are using what is effectively recycled water with no real additional usage, and that concerns about pollution from generators and about blight or displacement of people, including in Texas, are being taken seriously.

Wood says that Anthropic and its CEO, Dario, in particular have caused many of the negative headlines. She cites a 10% probability of human extinction within the next decade and massive job losses as claims that she says set the stage for the AI revolution. She says she cannot understand why he has been saying these things.

She states that ARK does not believe either outcome is true and that the firm expects a labor shortage instead. Wood says the one positive thing emerging from the drama is that shortages of power, chips, and possibly more data centers than otherwise would have existed are forcing efficiencies into the new AI world. She argues that when large amounts of money are thrown at a new technology, much of it is typically wasted, and that this cycle has a chance to introduce efficiencies early. She also suggests that this should be a moment to encourage people, including students, to understand and embrace the change rather than be disrupted by it.

Wood contrasts the current period with the tech and telecom bubble, when she says large sums were typically wasted. By introducing efficiencies very early in the technology revolution, she argues, the cycle should be longer-lasting and healthier, with much less fallout than the internet bubble and bust. She says she has made this point many times, adding that this is the real revolution and that its seeds were planted 20 to 25 years ago and are now flourishing.

She also turns to credit markets, showing a chart of credit default swap spreads for high-yield, lower-quality companies going back to about 2012. She notes the sharp spike during COVID, which she describes as a heart attack for those companies. Current levels, she says, have not yet reached a new high, although the metric is a little elevated and she is somewhat concerned about what the stock market suggests about the future, without being in shock.

Wood cites Ramp data showing that the top 1% of companies spend nearly $10,000 per employee per month on frontier AI models, while the top 10% spend a little over $800. The rest, she says, spend about $15 per month per employee. For her, that gap is evidence of how early this cycle is.

Wood says some fast adopters are already seeing dramatic productivity benefits and expects them to become more competitive over time, pulling their rivals into the AI race. She says ARC is harnessing AI aggressively, and that companies that want to compete will need to do the same.

Wood says there are big losers in software, referencing the so-called SaaS-pocalypse. She notes that Atlassian had a difficult week and that ServiceNow dropped about 27% in one day, even though its CEO, Bill McDermott, has described the ServiceNow platform as the execution layer of AI. The market, she says, does not believe that. She sees technology shifting toward frontier labs, and expects OpenAI and Anthropic, which she says have stolen the show, to continue succeeding. She also thinks SpaceX AI will have more success than most people understand because of orbital computing capacity, which she expects to become more important given the data center backlash on Earth.

Wood says Gemini is part of the pack but has not performed as well as the others, and that there is some leapfrogging under way. She expects software to keep moving up but not nearly as fast as hardware. She adds that embodied AI and robotics have barely begun. Citing work by Tasha Keiny and Daniel, she says the humanoid robot opportunity could be worth $26 trillion over the next 10 to 15 years, similar to 20% of today's global GDP. Robotaxis, she says, are an even nearer-term opportunity worth around $10 trillion.

Wood says natural gas has until recently been priced locally, based on local supply and demand, but the US is now exporting more. She notes that the difference between US and European natural gas prices is large. Over the past year, gasoline prices rose roughly 30%, while natural gas prices, the largest feedstock for the power grid, fell by a little over 35%.

She says this gives the US a competitive advantage relative to Europe and that she does not think people understand that the US grid is enjoying lower natural gas prices. She also notes that US natural gas storage is at normal levels ahead of winter, while Europe's is well below, suggesting possible demand for US gas in Europe, with the question being whether enough infrastructure exists to deliver it.

Wood says that one of the good numbers released this week was productivity, which grew at an annualized quarterly rate of 3.6%, while unit labor costs fell 0.7%. She says the Fed will pay a lot of attention to this. She explains that the gap between productivity and compensation per hour was a little below 3% at an annual rate.

If productivity keeps accelerating, Wood says, compensation per hour should turn around as well, which she calls a win-win for businesses and consumers. She believes the election is prompting soul-searching on this score and suggests technology could help address affordability problems. She lists other ways productivity gains could be shared, including through company margins, more R&D, investment in the technology revolution, and lower prices, and she expects the mix to differ across countries.

Wood describes free AI personal assistants offered by several developers, naming Anthropic's Duet, OpenAI's ChatGPT, Meta's Muse, and Instinct, which she says is a private company, though she is unsure of the exact names. She says that anyone who has experimented with these tools gets a personal assistant and a personal system at no cost that removes tasks forcing users to open apps and search.

She says that any app relying on transactions is probably going to face challenges. She adds that these assistants will delight both businesses and consumers and that productivity numbers are already coming out of enterprises.

Wood says the yield curve is still flattening despite all the drama. She recalls that during the Industrial Revolution the curve went negative because short rates rose, mirroring nominal GDP growth, while long rates were held back by 100 to 500 basis points because long rates focused on the deflationary effects of new technologies.

On the deficit, she expects the deficit to reach 2.8% of GDP by the end of 2028, citing likely gridlock and vetoes in the new Congress. She says President Trump could bring back the art of the deal and negotiate with Democrats, and she notes he has said Democrats would like working with him. She also points to bipartisan support around space, framed as national security and competition with China, saying the US is about 10 years ahead of China because of Elon Musk's work on reusable rockets.

Wood says non-farm payrolls came in stronger than expected, while the household survey was roughly in line with somewhat weak expectations. Average hourly earnings rose 0.4% for the month, a little higher than expected and a 5% annualized rate, though she does not expect that pace to continue right away. She suggests that companies may be starting to share productivity gains with workers, especially as the unemployment rate has fallen toward the 3s, a level seen only a few times in history.

She attributes the labor shortage to retiring baby boomers and immigrants who have left. The unemployment rate for the 16 to 24 age cohort ticked down from 9.9% to 9.3%, which she says may signal entry-level jobs freeing up. Citing Ramp data, she says more AI-intensive companies are hiring faster than non-AI-intensive or early-adopter companies.

Wood says the trade deficit will likely not please President Trump, but the US is growing much faster than the rest of the world, and most of the recent import growth is related to data centers. She notes that the effective corporate tax rate is 8%, the lowest in the world, which she says is drawing more foreign direct investment because returns will be there. Accelerated depreciation helps, and profit margins are near their highs, which productivity could push higher. She suggests that if companies want to keep labor happy, they will have to share some productivity gains with employees.

Wood also points to relationships among assets. She says metals falling relative to gold suggests China is still in a difficult position, while the S&P relative to oil and to gold is holding up, unlike past periods when inflation eroded market returns. She expects Bitcoin to keep rising relative to gold, but notes that the correlation between Bitcoin and gold has been around 0.1 since 2019.

Summing up, Wood says the economy is on a growth trajectory much stronger than most people recognize, and that third-quarter GDP growth of 3.4% proves it. She expects the acceleration to continue and says a housing pickup could make the economy even stronger, given housing's large multiplier effect. She expects inflation to be much lower than expected, citing productivity, oil prices, and AI costs seeping into the cost structure of industries.

She says that if companies do this well, the combination could push nominal GDP growth up, which would lift short and long rates, with long rates rising less than short rates. She does not expect a huge increase from current levels, though she acknowledges that investors overweighted in bonds will find it difficult. She says it will not be a crisis, except possibly for baby boomers heavily weighted in bonds who live hand-to-mouth and cannot absorb that capital depreciation.

Wood closes by saying the team misses John Nosta, that they held their first Brainstorm without him, and that they are trying to reach his family. She says she will see viewers after Thanksgiving, and adds that there is much to be thankful for.