The Fed and a Near-Certain Hike
The CME FedWatch tool shows about a 95% chance of a rate hike. At this point the Fed has no real choice - the market has priced it in, so a hike is close to forced. Every word from Kevin Warsh will be parsed for how dovish or hawkish he sounds. Attention then shifts to the October meeting and whether another hike comes. More telling: the December meeting now carries roughly a 30% chance (30.6%) of a third hike.
Do not fight the Fed. Monetary tightening is coming, and that is one of the large-scale economic factors flagged in the early-September market outlook. Even though the market looks undervalued, there is heavy volatility and more downside risk for the rest of this month and into early October, because the big-picture pressures have kept getting worse since they were first written about.
Warsh does not want to forecast or reveal anything. Expect little real information tomorrow. The useful question is not what the Fed says word for word, but what it means for valuations. Parsing every Fed statement is a job better left to economists.
Market Valuation: Cheap, but Only Because of Seven Stocks
Morningstar's method is bottom-up: cover more than 700 stocks trading on US exchanges, compare each price to its intrinsic value, and roll that up. That gives a market about 10% undervalued. Many other strategists use a top-down approach instead.
The catch: almost all of that discount sits in seven mega-cap names - Nvidia (NVDA), Broadcom (AVGO), Tesla (TSLA), Amazon (AMZN), Meta (META), Microsoft (MSFT), and Alphabet (GOOGL). Strip those out and the rest of the market is close to fairly valued. Individual cheap stocks exist here and there, but the undervaluation is concentrated in those seven. So even with the market cheap - usually the point to consider moving overweight - it is too early to go there.
Look at a name like Broadcom (AVGO): the market gives full credit for the 2027 outlook, guidance, and earnings growth, but the multiple steps down sharply from 2027 to 2028. The market gives no credit for growth past 2027. That gap between the internal outlook and what is priced in is one of the biggest differences in the market today.
Sticking With AI Leaders
The upside is real in those mega-cap AI names, but the market may not get comfortable with the AI buildout story into 2028 until well into the first quarter or first half of next year. So those stocks could stay flat in the near term.
The AI buildout has a long lead time. From deciding to build a data center to securing land, placing orders, and constructing it, you are locked in well ahead. That is why 2027 output is already baked in - the orders are placed and Taiwan Semi (TSM) already knows what is coming next year from tools it already has. Only when people start thinking about 2028, early next year, should those stocks start working to the upside.
The reported weekend news about a possible slowdown or "pacing" of AI models has not changed the strategy or outlook. It hasn't hit yet, and near-term buildout is already committed.
Avoiding the AI "Followers"
Stay with the technology leaders in AI, not the followers - the second-derivative beneficiaries that rose too far. The commodity-oriented tech hardware names have been a warning for a while. They jumped a couple hundred percent early in the year: SanDisk, Seagate (STX), Western Digital (WDC), Ciena (CIEN). Many have since fallen 30%, 40%, even 50%, and there is still more downside expected.
The reason these companies did so well is a supply-demand imbalance. If you need memory to open a data center, you pay whatever it costs - you cannot tell your boss the opening is late for lack of memory. So these firms charge what they want and earn huge margins. But they are now revamping product lines, redeveloping production, and building new facilities. Much of that new supply comes online around 2028. Once supply and demand rebalance in 2028, earnings at those companies start falling off.
An Emerging Non-AI Opportunity
Outside AI, more opportunities are appearing. Procter & Gamble (PG) is an example. It looked significantly overvalued a couple of years ago, peaked in 2024, and has been falling since - now it looks attractive again. It is a core-type holding: a wide economic moat, low uncertainty, and long-term durable competitive advantages. You can start building a position, collect a solid dividend while you wait, and it should hold up better on the downside. If the broader market sells off further, it becomes a strong candidate to dollar-cost average into.
The Contrarian Call: SpaceX Is Overvalued
SpaceX (private, tracked as SPCX) carries a fair value of $62 a share, while it trades at more than half that - roughly 40% off its August lows. The approach here breaks the company into parts. The space business is fairly forecastable: number of rocket launches and revenue per launch. Other lines can be modeled too - an estimate of what xAI is worth (using comparable AI companies) and Starlink.
The problem is the future business lines that do not exist yet. The market gives too much credit for the eventual value of those. Take data centers in space - that could happen eventually, but the base case applies probability weights to the upside and downside of all these unproven businesses. That is where the overvaluation comes from.
This is not a bear case on the business itself. Revenue was $19 billion last year, $37 billion this year, and is projected to reach $73 billion by 2030 - nearly a fourfold rise over five years. Earnings were negative last year, likely break-even this year, and only about 90 cents a share by 2030. At today's price, the stock trades at 160 times a five-year forward multiple, which is far too high.
The core debate on SpaceX is how to value a company with little available information, where much of the worth rests on the future and on pricing that future. The overvaluation conclusion comes down to how much of that future you are willing to believe and pay for today.
