Every earnings season, I see the same headlines:
- “Missed analyst expectations.”
- “Lackluster guidance.”
- “Stock crashes after earnings.”
And every time, I ask myself one question:
Since when did analysts’ assumptions become more important than actual business performance?
recently reported another solid . Revenue grew. AI-related business continued to expand. Cash flow remained strong. Margins were healthy. The company once again demonstrated why it is one of the highest-quality businesses in the semiconductor industry.
Yet the headlines focused on one thing:
“It didn’t beat expectations by enough.”
Think about that for a moment.
A company can grow revenues by billions of dollars, generate enormous cash flow, increase shareholder value, and still be labelled a disappointment because someone sitting behind an Excel spreadsheet expected a slightly different number.
As someone who has spent years building valuation models and earnings forecasts, I can confidently say this:
Most forecasts are assumptions, not facts.
Analysts drag cells in spreadsheets, adjust growth rates, tweak margins, and arrive at a target. The market then starts treating those assumptions as reality.
But reality rarely follows an Excel model.
If forecasting was that easy, analysts would consistently identify the next winners long before the market does. Yet history repeatedly shows that many of the biggest opportunities emerged when consensus was wrong.
The bigger question investors should be asking today is this:
Why are investors obsessing over a few cents of earnings variance while ignoring the massive amount of capital still flowing into AI infrastructure?
Every major cloud provider, enterprise software company, hyperscaler, and AI startup continues to spend aggressively. The AI race is far from over.
Broadcom remains one of the critical enablers of that ecosystem through networking, custom silicon, connectivity solutions, and infrastructure software.
The market may debate quarterly expectations. The business continues to execute.
I remember seeing similar reactions in other quality companies. sold off after earnings. A few days later it was making new highs. The business hadn’t changed. Only sentiment had.
- This is why successful investing requires separating noise from signal.
- Noise is what analysts expected.
- Signal is what the company actually delivered.
Long-term wealth is rarely created by following financial television headlines or reacting to analyst revisions. It is created by identifying exceptional businesses, understanding their competitive advantages, and remaining patient when short-term sentiment becomes disconnected from long-term fundamentals.
Broadcom’s earnings did not change my view of the company. If anything, the emotional sell-off reminded investors of an important lesson: Markets often punish quality companies for failing to meet expectations. Great investors use those moments to evaluate opportunity, not panic. Don’t let a spreadsheet tell you what a business is worth.
- Study the business.
- Understand the industry.
- Follow the cash flows.
- And remember: the stock market often transfers wealth from those who react to headlines to those who understand fundamentals.
The Question Nobody Wants To Ask: What surprises me is not that analysts are scrutinizing Broadcom’s earnings. What surprises me is what they are not scrutinizing.
Every quarter, investors hear endless discussions about whether a company beat consensus estimates by a few cents. Analysts build complex models, adjust assumptions, revise targets, and debate minor earnings variances.
But where is the same level of scrutiny when it comes to the hundreds of billions of dollars flowing into AI and semiconductor infrastructure?
The entire market knows that capital expenditure across AI, data centers, networking, semiconductors, power infrastructure, and cloud computing has exploded to unprecedented levels.
The real question is: Are the revenues being generated today sufficient to justify the scale of investment being made? History teaches us that every major technological revolution attracts excessive capital at some point.
Railroads. Telecommunications. The Internet. Housing. And now Artificial Intelligence. The technology itself may be transformational. That does not automatically mean every investment made during the boom will generate an acceptable return on capital. Even legendary investors like Ray Dalio have recently warned that AI exhibits characteristics often seen during speculative periods. The technology may change the world, but that does not mean valuations cannot become detached from reality.
When I look at the sector, I see enormous investments being made across the entire ecosystem. What I don’t see yet is proportional revenue generation across many parts of that ecosystem. Perhaps those revenues will come. Perhaps they won’t. But surely that deserves more attention than whether Broadcom exceeded an analyst’s spreadsheet by a few percentage points. There is another issue investors should consider. Many companies are aggressively pursuing productivity gains through AI while simultaneously reducing headcount.
In the short term, markets reward lower costs and higher margins. In the long term, however, economies still depend on consumers. Consumers buy homes. Consumers take loans. Consumers buy cars. Consumers spend on travel, entertainment, and services.
If technology eventually displaces employment faster than new opportunities are created, the economic consequences could be far more important than the next quarterly earnings estimate.
These are the questions investors should be debating. Not whether a world-class company missed a spreadsheet model by a fraction.
Because bubbles never announce themselves while they are forming. People usually recognize them only after they burst.
