Listen to the audio version of this article (generated by AI).

Fed holds, but two officials want a hike… AI’s surprising share of inflation… the price problem that outlasts inflation… Luke Lango’s 2028 warning… Microsoft and Alphabet just reported

Earlier today, the Federal Reserve held its benchmark rate at a target range of 3.75% to 4%.

The decision itself was no surprise. Heading in, futures traders had priced in roughly a 10% chance of a hike.

As for the official statement, it offered little substantive detail. Beyond saying that “the accumulation of more information” will guide its next move, it offered no hint about December.

What’s more interesting is today’s vote…

It was 10-2, with Fed Governor Lisa Cook and Cleveland Fed President Beth Hammack both dissenting in favor of a quarter-point hike. That’s a notable shift from September, when the committee voted unanimously to raise rates. The Fed is divided again – and the dissenters want to move faster, not slower.

Cook’s dissent is especially relevant to today’s Digest because she pointed at the AI buildout as her top concern heading into next year. As she recently put it at the New York Fed, “It’s already causing inflationary pressure.” Those pressures, she added, “may not resolve very quickly.” We’ll return to this idea in a moment…

Then came Warsh’s press conference, which was vintage Warsh: loads of words with little message. He called inflation “too high,” insisted there’s “no soft inflation target,” and once again refused to signal the Fed’s next move.

The one real piece of news was structural. Starting next year, the Fed will vote on rates at only seven of its eight meetings, using the other for a broader strategy discussion. In 2028, it drops to six rate decisions. Fewer decision points could mean bigger market moves around each one.

Now, returning to AI and the Fed, the most interesting part of the press conference was when Warsh was asked whether rate hikes could affect the AI investment surge. His answer:

…with respect to the new technology and this investment surge – do interest rates matter?

I believe they do. And I don’t believe that they’re somehow divorced from the rest of the economy.

Fair enough – rates do matter, and we’ll look at where they bite below. But “matter” isn’t the same as “decide.”

The risk the Fed can’t fix

Wall Street will spend tonight speculating about hikes in December and beyond. And tomorrow’s PCE inflation report could influence opinions on where rates are going and push stocks around over the coming days.

But for AI investors, the Fed can only slow or speed up this trade. It can’t stop it.

The bigger risk is one the Fed can’t fix – and its rate hikes may make things worse.

To unpack this, let’s start with how Warsh himself described the AI spending boom two weeks ago – a “shock.” The problem is that the consequences of this shock will linger long after the Fed stops hiking or even starts cutting rates.

Warsh’s new word for AI

Before becoming Fed Chair, Warsh was one of AI’s biggest optimists on inflation. Last November, he wrote that AI would be “a significant disinflationary force.”

But two weeks ago in Bangkok, at the IMF’s annual meetings, he described the AI spending boom very differently:

If you’re a central banker… you’ve got to recognize this as a shock.

His point was about timing: AI-related spending is hitting the economy now – shocking it – while the productivity payoff arrives later.

In other words, AI comes with a bill before it delivers a benefit. The long-term promise is AI making everything cheaper and more efficient. But first comes the cost of building it – and right now, that cost is rippling through the economy.

How big is the bill?

Bigger than most investors realize.

Researchers at the Minneapolis Fed found that AI-driven demand for memory chips and computer hardware pushed prices in that category up 12.2% over the year through July. That alone added roughly 0.4 percentage points to core PCE inflation – about as much as President Trump’s entire tariff program. Goldman Sachs estimates AI’s contribution could reach 0.5 points by year-end.

For context, core PCE – the Fed’s preferred inflation gauge – ran at 3.0% in August against the Fed’s 2% target. That makes AI responsible for about 40% to 50% of the overshoot.

You can see it in your own budget. A 32GB memory kit for a home PC that cost $100 to $200 a year ago now runs about $450.

Meanwhile, data center demand has driven electricity costs to climb where those centers cluster. As of June, residential power rates were up 10.4% in Pennsylvania and 9.7% in Ohio from a year earlier – roughly double the national average.

Why rate hikes can’t fix this

The hyperscalers aren’t going to stop building AI because the fed funds rate climbs a quarter point. They’re funding the buildout mostly from mountains of cash (and now, some new equity issuance as we saw with Meta Platforms Inc. (META) last week).

Hammack – one of today’s two dissenters – acknowledged as much, noting that higher rates could slow activity outside the AI sector without necessarily interrupting tech investment.

Now, as we noted above, Warsh pushed back on this idea today. And at the margins, he’s right – rates “get in the cracks,” as he put it in today’s press conference. You can see those cracks in companies building AI with borrowed money.

As we highlighted in the Digest last week, bonds of AI cloud provider CoreWeave Inc. (CRWV) touched a 15% yield earlier this month as lenders demanded higher yields to fund its debt-heavy buildout. Its stock fell, and so did peer Nebius Group N.V. (NBIS).

As The Wall Street Journal put it today:

Borrowing costs have risen for some lower-rated companies tied to the AI build-out.

But those are the edges of the buildout, not its core. Higher rates may thin the herd of debt-funded players. They won’t stop the hyperscalers’ spending, which is driving most of AI’s price pressure.

So, for the hyperscalers driving most of this inflation, a hike barely registers. Instead, it lands on the family with a mortgage, a car loan, and a rising power bill. The Fed’s main tool doesn’t touch AI-driven inflation – it just adds to Main Street’s bill. And as you’ll see, that bill matters more for your AI stocks than any rate decision.

The good news that isn’t

Now, here’s a twist – AI’s contribution to inflation may be near its peak.

Memory prices are still rising, but far more slowly. Contract prices for conventional memory chips jumped 90% to 95% in the first quarter of this year, then rose just 13% to 18% in the third quarter. Goldman expects its memory-linked price index to peak around November, and Oxford Economics projects AI’s drag on core inflation diminishing through 2027. Power costs are leveling off, too.

So, by next year, AI’s price pressure may stop getting worse. But that’s not the same thing as getting better.

As we’ve highlighted in the Digest many times, inflation measures how fast prices are rising – not how high they already are. And certainly not the fallout of high prices.

We have a great example of this with the housing market.

The latest Case-Shiller national home price index, released yesterday, rose 2.1% over the year through August. That trails the 3.4% inflation rate, meaning home values have fallen in real terms for 15 straight months. By that measure, housing has no significant inflation problem at all.

Yet the median existing-home price was still higher in September than a year earlier – the 39th straight month of annual gains – and sales are crawling along near their slowest pace in roughly three decades.

You see, after the price explosion a few years ago, prices never came down even after inflation faded. The result is a price level that makes buying completely unrealistic for most people, inflation levels be damned.

And that’s the trap. The Fed may win today’s inflation fight, with rate hikes cooling the rest of the economy and AI’s price spike running its course. But AI-driven prices may not come back down meaningfully. So, Main Street can keep losing the affordability war long after the Fed declares victory.

So, where will Main Street’s frustration over high prices go?

To the ballot box.

When Gallup asked Americans whether they want an AI data center in their community, 70% said no. And the backlash is bipartisan, with critics ranging from Bernie Sanders on the left to Josh Hawley on the right.

Just last week, NBC’s Meet the Press devoted a segment to whether AI anger could help decide control of Congress in next Tuesday’s midterms.

What about AI’s long-term payoff? Any chance it arrives soon enough to ease the pressure?

Well, in one way, it may be starting to show up. According to the BLS, U.S. worker productivity rose 2.2% over the year through the second quarter, and some economists see early signs of AI at work.

But so far, those gains are going to profits, not paychecks. Over that same quarter, real hourly compensation fell 3.1%, and the share of business output paid to workers as wages and benefits fell to 52.9%, the lowest on record.

So, the payoff may be arriving – just not for Main Street. And that only adds fuel to this simmering anger.

Regular Digest readers know our technology expert Luke Lango, editor of Innovation Investor, has been warning about exactly this.

Here’s Luke:

The force that will derail the AI Boom is not a technological failure, demand collapse, or even a recession.

It is politics – specifically, a populist backlash against AI that is already building momentum.

Luke has tied that backlash mostly to the 2028 election cycle – “not this earnings season or even this year.”

And this is where we come back to the Fed. The backlash comes from high prices, not the inflation rate. So, it won’t fade when AI inflation peaks – or when the Fed stops hiking.

So, what do you do with this?

Let’s start with a simple metaphor – the AI trade as a race car…

Interest rates are the brake. Rate expectations tap it or ease off it from week to week, often driven more by emotion than substance.

The hyperscalers’ spending is the gas pedal – the longer-term force that actually drives this trade forward.

Finally, politics is the emergency brake. Ultimately, it will decide whether the AI trade keeps running – or skids out when that brake gets yanked.

Too many AI investors are focusing on the brake, paying less attention to the gas pedal, and hardly any attention to the emergency brake.

Luke’s advice for handling this is simple – use the window while it’s open:

Make your money now. The window for transformational wealth creation in this AI cycle is the next two to three years. This trade will not last forever.

In practice, that means three things:

• Know where the political heat lands. The parts of the AI trade that show up on a voter’s monthly bill – especially power-hungry data centers where rates have spiked – are the likeliest targets for new fees, rules, and siting fights.
• Favor builders that pay their own way. Last week, the market punished Meta after its report showed negative free cash flow alongside $825 billion in spending commitments. Companies funding the buildout from cash rather than debt hold up better if rates stay high or the politics turn.
• Write down your exit plan now. Decide in advance which AI names you’ll hold through a political storm and which you’ll sell – and when.

To follow along with Luke’s analysis of the AI boom, his Peak Watch List that we covered in yesterday’s Digest, and which specific AI stocks he’s positioning in before 2028, click here to learn about joining him in Innovation Investor.

Before you go – a check on the gas pedal

Microsoft Corp. (MSFT) and Alphabet Inc. (GOOGL) were scheduled to report after the closing bell today, so their results are likely out by the time you’re reading this.

The key number to watch is spending.

Alphabet alone has guided to roughly $180 billion to $190 billion in capital expenditures this year. Whether these two keep raising their budgets – and whether Wall Street rewards or punishes them for it, as it did Meta last week – will set the tone for tomorrow. Amazon.com Inc. (AMZN) reports after the close tomorrow.

So, to sum up the week: Today, the Fed kept its foot off the brake – for now. Tonight’s reports tell us how hard the hyperscalers are still pressing the gas. And next Tuesday’s midterms will give us our first real read on the emergency brake.

We’ll report back on all of it.

Have a good evening,

Jeff Remsburg

(Disclosure: I own MSFT, GOOGL, AMZN)