• Lykeion
  • Posts
  • Markets Update - April '24

Markets Update - April '24


  1. Q1 Key Takeaways

  2. Innovations that Make Money

  3. Data Centers and Energy Demand

1. Q1 Key Takeaways

It’s been a risk on market, with positive returns across the board – credits due to the liquidity cycle and continued enthusiasm around AI. Best performing market of the major developed ones is Japan (mostly driven by a continued devaluation of the Japanese Yen), followed by Italy and Germany/US.

As we mentioned last month, rate cut expectations have been significantly reset since the beginning of the year, leading yields to once again move higher (Roger opined back in December that the Fed may not be cutting at all this year). Financial markets are now mostly in agreement with the Fed’s rate cut expectations.

Real yields are also in positive territory for the first time in a couple of years.

There is also a reasonable case to be made that the peak hiking cycle, globally, is behind us now, with the only major Central Bank still in hiking mode being the BoJ – (read our Does a BoJ Rate Hike Matter? piece for more on this).

The best-performing sector year-to-date is Energy – credits due to a 20% move higher in oil prices. Energy continues to be the only sector in the S&P 500 that trades at a discount to its 20-year PE average (more on this below).

Overall, global equity markets continue to be significantly in favour of US equities (which now comprise 64% of global financial markets, a 40-year high). About US equities:

  • US tech market valuations are back at 2021 peaks.

  • PE discount of international markets vs the US is now at 35%, the largest ever. That is partially justified by the difference in growth of US companies vs international ones.

We don’t find any particular reason for this to change any time soon.

 2. Innovations That Make Money

Fresh all-time highs are, counterintuitively, not a bad time to invest. Historically, investing at new all-time highs has led to better returns than investing on any regular day for the S&P 500 since 1988 – as long as investors commit to remain invested for more than six months.

More importantly, whilst equity markets are soaring, there are few signs of exuberance around the lower quality part of the market currently, which is usually a good indicator that risk appetite is still under control. Investors have not been chasing the young and unprofitable companies like they did back in 2021 and the IPO market, still scarred by the post-pandemic SPAC mania, is still very unenthusiastic.

Remember, in the long term, the reason why tech has done so well is because it tends to enable investors with great investment opportunities not only because it brings to market innovative products and services, but, more importantly, because those innovations sometimes will make a lot of money – which you see below.

In fact, making money (i.e. generating free cash flow) is actually a better long-term indicator of the performance of your investment than the level of innovation that a product promises to bring to market. And unless you’re trading or at least actively trying to time the ins and outs of your investments, we’d recommend you pick your investments based on the visibility and growth outlook of free cash flow rather than based on how disruptive a certain product promises to be.

All of this brings us to the key question: is AI a true innovation with a clear path toward helping companies make more money? Or does it promise more than it can deliver?

  • So far, the market has favored AI suppliers to a significantly larger extent than AI beneficiaries (the companies that will actually use the AI).

  • What this means is that financial analysts have had an easier time modelling increases in margin and growth in earnings that come from increased GPU demand (to suppliers of GPU) rather than modelling the increase in margins or grow earnings of companies that one would expect to be able to integrate AI within their product offering.

  • Again, this is (1) a sign that risk appetite is not exuberant yet and (2) that we’re in a transition phase in which we’re still trying to understand if AI is indeed an innovation that will help big tech make more money, or just another hype cycle.

  • If that’s the case, then there’s likely still plenty of money on the table to be made. So keep that question front and center going forward.

 3. Data Centers and Energy Demand

Loads of headlines this month about Amazon’s purchase of a nuclear energy-backed data center for $650 million in Pennsylvania.

We think this matters and is a story that we’ll see many more times going forward, albeit in different formats, with the key narrative being: big tech understands that AI requires a shit ton of electricity, and if they believe that AI is truly an important technology for their business model, then they need to make sure they have priority access to electricity in order to ensure power reliability. This is likely the key reason behind Amazon’s deal and also the reason why Meta has recently added former Enron executive and energy trader John Arnold to its board. Tech companies will increasingly need people within their organization that understand the energy market.

In fact, AI’s strain on data centers is already making some noise. As per Michael Cembalest at JPM:

“Last December, Google announced that its Gemini Ultra LLM finally outperformed GPT-4.

But look closely at the labelling: 32x means that Gemini ran through its paces 32 times, picking the best answer from the set. In other words, this took 32x more compute time and energy than a typical prompt.

Energy is a hot topic in AI circles. As shown on the right, Dominion Resources, which serves 6 million customers in 15 states has sharply increased its projected power needs for the coming decade. The sole reason for the change: increased demand from data centers.

Dominion may not be representative of all utilities, but with the cost of transformer equipment already rising sharply, electrification of transport/heating and increased power demand from AI may turn electricity into a much scarcer resource. McKinsey, BCG and S&P Global all expect US power demand to grow by 13%-15% per year until 2030.”

It’s funny to see these reality checks on one side, whilst on the other we still have the IEA forecasting peak oil demand scenarios, credits due to the increased expected proliferation of renewable energy.

We still live in a world, and a financial market, that believes that there will not be a significant strain on energy supplies in the coming decade, despite:

  •  More than a decade of underinvestment in oil and gas exploration.

  • Significant reduction in above and below-ground inventories, especially after Saudi Arabia decided that they would not increase spare capacity to 13 million barrels of oil output a day (from a theoretical 12 million, although no one besides the Saudis knows for sure).

  • 7 billion people living in developing countries, where economies are growing the most, consuming around 3 barrels of oil per capita per year, compared to 13 for developed markets (and 22 for the US).

  • History tells us that when a country’s economy grows, energy consumption per capita follows.

  • The performance of the Energy sector against the broader market has likely bottomed in 2020. Not only are demographics and geopolitics a tailwind for the asset class, but even tech innovation and adoption are starting to become one as well. Hard to see how the sector won’t continue to deliver strong results to its shareholders in the coming future.

Enjoyed this one? We’d be thrilled if you helped with our growth story and shared Lykeion with a few friends… rewards await!