Can Policy Solve AI’s Chipflation?

Welcome to Thoughts on the Market. I’m Mayank Maheshwari, Morgan Stanley’s Asia Energy analyst.

Today: how AI’s rapid growth is forcing Asia into a massive energy buildout across power grids, fuels, storage and dependable energy and power generation.

It’s Tuesday, June 9th at 8am in Singapore.

Every time you ask AI to draft a note, summarize a file, plan a trip or generate an image, the response feels instant and easy. But behind it sits a very physical system: data centers, electricity, cooling, fuel, metals, power lines, storage tanks and ships.

There is no AI without energy. And in Asia, the power and energy needs could get much bigger.  And right now, we are at a critical inflection point where energy, AI, and security converge into [a] once-in-a-generation investment cycle.

We see a super cycle with $5 trillion plus in new investments in energy over next five years, almost double of what we have seen in the past decade. And this has global implications as Asia consumes almost half of the world's energy needs – but produces only about a third of it at home. Energy markets may be global, but energy insecurity is local. It shows up in electricity prices, fuel shortages, factory delays, food supply pressure and household budgets.

By 2030, Asia’s energy use could rise by about 38 exajoules. That increase is roughly equal to all the energy the Middle East consumes today. Power demand alone could reach about 19 trillion  units a year when expressed in kilowatt-hours. That is around four trillion more units of electricity usage than in 2025, driven by data centers, industry, and onshoring of businesses.

AI is now part of that demand story. By 2030, data centers could use roughly one-sixth of all new power units in Asia. That makes AI a major new load on the power system.

Meeting this demand requires a major investment cycle. Asia’s annual energy investment could rise to roughly US$1.1 trillion a year over the next five years. Much of that spending goes into the power system itself: generation, grids, storage and the equipment needed to connect everything.

Grids may be the biggest bottleneck. Think of [the] grid as the highway system for electricity. You can build more power plants, but if the roads clog up, the power does not reach homes, factories or data centers. Asia’s grid investment needs could reach close to about US$1 trillion by 2030. Transformer lead times have stretched to years in some cases, which shows how tight the equipment supply chain has become.

The hardest part is keeping the lights on every hour of the day. Baseload power means electricity that can run around the clock. Asia is adding a large amount of renewable power to its energy infrastructure. But that source depends on when the sun shines or the wind blows. That is why coal, gas and nuclear remain part of the conversation.

Storage also moves from useful to essential. Batteries help smooth out renewable power demand when supply rises and falls during the day. Global energy storage installations could rise from about 500 gigawatt hours in 2025 to around 3,000 gigawatt hours in 2030.

Powering AI also reaches beyond electricity. Data centers need power, but the system around them needs dependable fuels, grids, batteries, metals, refining, storage and shipping. Electricity has to be generated, moved, backed up and supplied through physical infrastructure. That is why this story pulls in copper and aluminum for grids, fuel refining for transport and petrochemical supply chains, and fertilizers because energy security also connects to food security.

The future may look digital, but it will be powered by something far more physical: the largest energy buildout Asia has seen in decades.

Thanks for listening. If you enjoy Thoughts on the Market, please leave us a review wherever you listen and share the podcast with a friend or colleague today.

Andrew Sheets: Welcome to Thoughts on the Market. I'm Andrew Sheets, Global Head of Fixed Income Research at Morgan Stanley.

Today, why is everything still so expensive?

It's Thursday, June 11th at 2pm in London.

The Federal Reserve has a so-called dual mandate, tasked with keeping the labor market healthy and prices stable. It is currently having much more success with the former than the latter.

Let's start with that good news.

Last Friday saw solid data from the U.S. jobs market, reducing some of the fears from earlier this year that artificial intelligence and other factors would lead companies to make do with fewer workers. The U.S. unemployment rate sits at just 4.3 percent, a historically low level. Measures like initial jobless claims indicate no large uptick in firings.

Yet the success within the U.S. labor market is mirrored by struggles with inflation. The Fed tries to keep inflation, the annual increase in a broad set of prices, to about 2 percent per year. Their preferred measure of these prices, so-called PCE inflation, well, it's been materially above this target over the last three months, six months, twelve months, and indeed, the last five years.

As for another key measure of inflation that was reported yesterday, CPI, overall prices increased more than 4 percent. While that was close to expectations, it still represents prices that are rising much faster than the Fed would prefer.

This leads to a dilemma. One diagnosis of what's going on is that elevated inflation is a sign that conditions are simply too loose and too accommodative at these levels of interest rates. Corporate capital expenditure and merger activity is surging, regulation is being eased, and the U.S. government is spending a lot more than it's taking in. All of these are consistent with a hot economic cycle, which in the past would've warranted higher interest rates to bring the economy back down to a more sustainable speed.

But it might not be that simple.

The surging spend that we're seeing on AI data centers feels pretty unique and almost insensitive to other dynamics. Indeed, we've seen a 700 percent increase in the price of memory over the last year. Yet it's done little to slow demand for this construction as the large, well-capitalized companies behind the AI buildout see it as so essential to their future success.

U.S. consumers are also still spending, boosted perhaps by record levels of household wealth. As just one example of this, my colleagues in Equity Research note that the price of airline tickets has gone up 25 percent over the last year, yet there's been no sign of people flying less.

Now, the positive story would be that while there are some high-profile categories like computer memory or airfare that are seeing these large price increases, the broader inflation picture is actually set to get better as the year goes on, and costs for things like housing and tariff-impacted goods moderate. That is our view at Morgan Stanley, where our economists think that inflation will ultimately be lower over the next twelve months – and lower than many in the market expect.

But there's definitely uncertainty.

This month, June, is one where central banks may appear to have a renewed commitment towards inflationary pressures; with the ECB hiking rates today and our expectation that the Bank of Japan will hike rates next week, while the Fed will remove their easing bias. And our more benign economic base case for inflation does assume that oil will start flowing through the Strait of Hormuz pretty soon. It may not, and that could also lead to more sustained inflationary pressure.

The big story on inflation has not gone away. Our assumption that pressures could ease in the second half of the year is a key and differentiated input to our forecast for lower bond yields and higher stock prices in 12 months' time. But it does rely on a change of the status quo.

As of now, inflation is still too high.

Thank you, as always, for your time. If you find Thoughts on the Market useful, let us know by leaving a review wherever you listen. And also, tell a friend or colleague about us today.