A Better Bargain, Not an Off Switch: Where the AI Trade Goes From Here

The fight ahead is about who pays the bill, not whether AI matters. Here's what to watch and how to keep a retirement plan steady through it. (Part 3 of 3)

Over the last two days, we covered what happened to the AI trade and who is paying for the AI buildout. Today, the last part: where this goes next, and what it means for a retirement plan.

This morning offered a perfect example of why the answer isn't simple. CoreWeave, one of the companies building data centers for AI, reported results that beat expectations. Its quarterly revenue jumped 168% to $3.66 billion, it raised its forecast for the year, and its shares jumped about 20% [CNBC, 2026-11-11; Quartz, 2026-11-11]. The same company also lost $709 million in the quarter [Quartz, 2026-11-11]. And just weeks ago, the cost of insuring its debt against default had climbed to levels usually associated with financial distress [CNBC, 2026-11-11].

Booming demand and serious financial strain, in the same company, on the same morning. That is the AI story right now.

What Happens Next

The Fight Is About the Bill

Last week's election didn't turn AI into a political target overnight. It gave louder voices to people who were already asking hard questions about data centers. According to the Brookings Institution, those questions are practical: who pays for the extra electricity, where the water comes from, how the land gets used, and who funds the upgrades to the power grid. Brookings also notes that many of the promises developers have made so far are voluntary, and the new political landscape could make some of them mandatory [Brookings, 2026-11-04].

There's a local side, too. A data center brings a big wave of construction jobs, but far fewer permanent ones. One review of more than 1,200 data centers found that even the largest generally employ fewer than 150 permanent workers [Business Insider via Yahoo Finance, 2026]. For a town weighing higher power bills against a handful of long-term jobs, the math can be a hard sell.

No Single Switch in Washington

It's easy to picture Congress flipping AI off. That isn't how this works. The new Congress takes office in January, and the President still holds a veto. Just as important, Brookings points out that many decisions about energy and land use have historically been made at the state and local level [Brookings, 2026-11-04].

So the likely path is a patchwork: local zoning fights, state utility rules, federal negotiations, and concessions from the companies themselves. Messy, slower, and more expensive, but not a shutdown.

A Better Bargain

My view is that the strongest case for AI isn't "stop complaining and build." It's "build, but strike a deal that voters will accept, so construction can continue." That means developers paying more of the costs they create, clearer permitting rules, and real benefits for host communities. That bargain could keep much of the investment going. It cannot guarantee every project, every job, or every stock price.

What Does It Mean for Me?

Two Very Different Recoveries

History offers two lessons, and both matter right now.

In April 2025, tariff fears knocked the market down hard. Then it came back. The S&P 500 set a new record by late June, roughly eleven weeks after the bottom [CNN, 2025-06-27].

In 2000, the dot-com bubble burst. The Nasdaq set a record on March 10, 2000. It did not close above that level again until April 2015, more than fifteen years later [Reuters, 2015-04-23].

What separated the two was not how bad the first day felt. In 2025, the problem was a policy shock that could be reversed. In 2000, companies were priced for profits that never showed up.

Timing Shock or Profit Shock?

That leads to the question for the next year. A timing shock means AI projects get delayed: revenue arrives later, but it still arrives. A profit shock means the spending never earns enough to justify itself. The first can be painful. The second can last a long time.

Right now, the evidence points more toward timing. CoreWeave's results this morning suggest demand is real. Some Wall Street firms expect certain projects to slip a year rather than disappear [Reuters, 2026-11-09]. But it's too early to rule out the second kind, especially for companies that borrowed heavily against future customers.

Three Paths to Watch

These are scenarios, not predictions:

My base case is the first path: a contested, more selective buildout, with winners and losers sorted company by company. That is a judgment, not a promise, and markets can move a long way while the answer becomes clear.

What That Means for Your Plan

For someone ten years from retirement, a slow, messy AI story is survivable. For someone drawing income now, the risk is different. A big chunk of a typical index fund now rides on this one theme, and a drop that lines up with withdrawals can do lasting damage. The goal isn't to guess which path we get. It's to make sure your plan can live with any of them.

What We're Doing About It

Planning for All Three Paths

When a story this big is still unfolding, our work with clients starts with their plan, not the headline. How much income will the portfolio need to provide over the next few years? Where would that money come from if stocks are down when it's needed? How much of the portfolio rides on a single theme, and does that fit the client's timeline? No approach avoids every decline, and losses are always possible. But clear answers to those questions make a bad market far easier to live through.

One Roof, One Picture

A shift like this touches more than the portfolio. It can change the best order to draw from accounts, how gains get harvested for taxes, and how an estate plan holds up. We handle investments, tax planning, and estate planning under one roof, so those pieces get looked at together instead of in three different offices.

The Bottom Line of This Series

AI may well change the economy. That was never the question. The questions are how long it takes, who pays for it, and what price is fair for the companies involved. Last week, the market started asking those questions all at once. Expect it to keep asking for a while. A plan built for that is worth more than any prediction.

Take Action

If this series made you wonder how much of your own retirement rides on AI, or whether your income plan could handle a slow, messy few years, let's talk. Reply to this email, call the office, or schedule a review here. For direct contact: [email protected].

───This post originally appeared on The Lighthouse on Substack — weekly financial analysis written specifically for people within ten years of retirement. Subscribe free at The Lighthouse on Substack or upgrade for weekly Trades of the Week picks.───

Investing involves risk, including the possible loss of principal. This article is for informational purposes only and does not constitute individualized investment advice. Past performance is not indicative of future results. Fixed income investments are subject to interest rate, credit, inflation, reinvestment, and liquidity risks. Equity investments are subject to market risk and may decline in value, sometimes significantly, over short or extended periods. Yields and market levels referenced are as of the dates indicated and change continuously.

Asset Guidance Group, LLC ("AGG") and its personnel may hold positions in securities or assets discussed, subject to AGG's personal-trading policy. AGG and its personnel do not undertake to disclose specific positions or to notify readers of position changes.

PulseSignal Weekly is AGG's impersonal research publication; subscribing to it does not establish an advisory relationship with AGG or provide individualized advice.

AGG is an investment adviser registered with the State of Georgia (CRD #291448). Registration does not imply a certain level of skill or training.

Wallace R. Nichols, MBA, JD, CFP® Chief Investment Officer Asset Guidance Group, LLC