Markets took a hard hit yesterday, and the damage is not random. The selling is circling the same group that led last year’s rally—software and the big names tied to artificial intelligence. The pain then bled into banks, where credit exposure to software firms is now under the microscope. The main point is simple: in a world reordering around AI, the line between winners and future casualties is thin.
“There’s a murder history playing out as the stock market gets smoked today. We aren’t searching for the killer. We’re searching for the victims.”
I am Taylor Sohns, CEO of LifeGoal Wealth Advisors, a CIMA and CFP. I spend my days inside the market’s moving parts. Today’s selloff is a reminder that themes can work until they don’t, and concentration risk shows up fast. Many investors believed AI would lift every boat tied to software and mega-cap tech. That story just hit a rough chapter.
The Selloff in Plain English
Software stocks, last year’s darlings, have dropped about 30% from recent highs. Even Microsoft, widely seen as a prime AI play, is down around 30%. On the day, major banks slid more than 4%. The link is not mysterious. When a sector that relies on borrowing stumbles, its lenders feel it next. That is old market math.
“Software… down 30%. The darling Microsoft, down 30%. Wait. I thought Microsoft was AI.”
Some investors now ask a hard question: if even the perceived winners can shed that much value in a hurry, how confident are we in our stock picking under an AI reset?
View this post on Instagram
Why Software Is Hurting
Software enjoyed years of easy growth. Then came AI hype, which pushed expectations forward. Prices ran first. Profits had to catch up. They have not kept pace. That gap is now closing the rough way—through price.
There are real pressures behind the drop:
- Spending mix shifts: Companies are still buying software, but they are cutting unused seats and pushing vendors for better pricing.
- AI is expensive: building and running AI features require significant computing resources. That raises costs before revenues scale.
- Monetization lag: Many vendors shipped AI add-ons fast. Customers are still testing them. Revenue ramps slower than slide decks promised.
- Cannibalization risk: Some AI tools replace existing modules. That can compress the old revenue base while the new one is not ready.
- Higher hurdle rates: Investors now demand stronger cash flows, not just growth. Valuations reset when growth is less certain.
Microsoft’s decline is the headline, but the theme is broader. When leaders pull back, it shakes confidence across the group. It also forces investors to rethink what “AI exposure” really means. Does a label equal durable earnings? Not always. The market is sifting stories from cash flows.
How The Stress Reached Banks
Today’s slide in financials, including JPMorgan, Bank of America, and Citi, is a reaction to perceived credit risk. Big banks and specialty lenders finance software companies through lines of credit, venture loans, term debt, and bridge facilities. When software weakens, lenders re-rate that risk.
Here is what investors are pricing in:
- Slower growth at borrowers: If software firms cut hiring or spending, their leverage metrics look worse.
- Funding needs rise: Companies with weak cash flow may lean harder on credit lines, raising default risk.
- Down-rounds and M&A delays: If equity markets cool, borrowers lose a key lifeline. That boosts pressure on lenders.
- Portfolio marks: Banks and private lenders revisit valuations. That can hit earnings and tighten new lending.
None of this says the banking system is in danger. It does say the market is recalibrating. When the software cycle cools, credit costs rise. Equity investors do not wait for charge-offs to show up; they move first.
Winners, Losers, and AI Reality Checks
AI is real. It is changing work, speeding coding, and reshaping content creation. But the winners’ list was never going to be as wide as the marketing suggested. Some firms will save more money with AI than they make from it. Others will enjoy usage gains but struggle to bill for it at scale. A few will turn AI into durable pricing power. Sorting that out takes time.
I look for three traits when judging an AI story:
- Clear monetization path: Can the company charge more for AI, or sell to new users who would not have bought before?
- Margin discipline: Are compute costs contained, and is the business lifting long-term margins, not masking pressure?
- Customer hold: Do clients lock in due to unique data, workflows, or compliance needs that AI alone cannot copy?
Companies that check these boxes can survive a reset and thrive. Those that rely on buzz or price hikes without value risk churn and revenue slip.
What This Means for Everyday Portfolios
The message is not to run from AI. It is to respect the concentration risk and the speed of narrative shifts. The market overpays for simple stories. It underprices messy transitions. AI sits squarely in the messy phase.
Here is how I am thinking about positioning after a day like this:
- Cut single-theme concentration: Cap exposure to any one theme, even a popular one, so a single break does not drive results.
- Favor quality balance sheets: Cash, low net debt, and recurring revenue help when growth cools.
- Blend factors: Mix growth, value, and dividends to reduce reliance on a single style at a time.
- Use broad ETFs as core: Keep single-stock picks as satellites, not the whole plan.
- Rebalance with rules: Set thresholds to trim winners and add to laggards. Do not guess at peaks.
- Keep dry powder: A short-term reserve helps you buy when volatility spikes.
- Harvest losses: Use tax rules to improve after-tax returns without changing overall exposure.
- Mind credit spillovers: For bond sleeves, stress test exposure to cyclical credit and extend duration only where paid to do so.
These steps do not require a perfect forecast. They reduce the cost of being wrong, which is the real job in choppy markets.
Spotting The Next “Victim”
The hunt for the next victim is already on. That search can lead to hasty moves. A better approach is to test business models under pressure. Ask simple, tough questions.
For tech and software:
- Is the company’s AI feature set hard to copy, or could a platform vendor replicate it in a quarter?
- Do customers sign multi-year deals with clear upsell milestones, or is usage optional and easy to trim?
- Are compute costs rising faster than revenue per user?
For banks and lenders:
- How heavy is the exposure to venture-backed or pre-profit borrowers?
- What share of loans reprice soon, and at what spreads?
- How conservative are loss reserves versus prior cycles?
For the broader market:
- Where did valuations get stretched most during the AI rush?
- Which sectors benefit from AI productivity without paying heavy compute bills?
- Which industries gain pricing power from AI-driven efficiency and data advantages?
These checks help separate fair drawdowns from structural cracks. They also surface bargains where fear outruns fundamentals.
What I’m Watching Next
Over the coming weeks, earnings reports will do the real talking. I’ll track AI monetization metrics, not just user counts. I want to see attach rates, price uplift, and churn. On the cost side, I care about gross margin trends tied to AI features. If margins slip while usage rises, that is a red flag.
On the banking front, I will study credit provisions, commentary on software borrower health, and any tightening in underwriting standards. If banks raise reserves and hint at rising delinquencies in venture or software-linked books, the equity market will keep discounting that risk.
I also watch second-order effects. If software firms pause hiring or cut vendors, ad budgets, consulting, and cloud usage may slow. That ripple does not stay inside one ticker.
The Emotional Trap to Avoid
Days like this tempt investors to chase whatever held up and dump what fell. That often reverses right after the panic. The better move is to revisit your thesis. If it holds, use weakness to improve entry points. If it broke, size it down and move on. No heroics, just discipline.
Anchoring to last year’s winners is dangerous. So is declaring AI “over.” We are moving from a simple story to a harder one. Harder stories can still pay—if the math works.
Key Takeaways
- Software has dropped about 30% from highs; even Microsoft is down a similar amount.
- Banks fell more than 4% today as investors priced in credit risk to software borrowers.
- AI is real, but monetization is uneven and costs are heavy. Earnings, not labels, will decide winners.
- Reduce single-theme exposure, focus on cash flow quality, and stick to a rules-based rebalance.
- Watch margins, attach rates, and bank credit provisions in the next earnings cycle.
I started by asking who the victims are. For now, software and its lenders are on the hot seat. That does not make them permanent losers. It does mean the easy phase of the AI trade is over. From here, results—not stories—earn the premium.
Frequently Asked Questions
Q: How can I tell if an AI-featured stock is worth holding?
Check whether the company can charge more for AI in a way that customers accept. Look for rising attach rates, stable or improving gross margins, and low churn.
Q: Why did bank stocks fall on the same day as software?
Many banks and lenders finance software companies. If software weakens, lenders face higher credit risk, so investors price in larger loan losses and push bank shares lower.
Q: What portfolio moves make sense after a sharp tech selloff?
Trim concentrated positions, add to diversified core holdings, and rebalance. Favor companies with cash, low debt, and pricing power. Use weakness to upgrade quality, not to chase.
