When Revenue Acceleration Overwhelms Quality

Key Takeaways

• AI-driven scarcity is allowing companies perceived as lower quality to post the revenue growth, margin expansion and rising returns that investors associate with quality, lifting estimates, multiples and stock prices.
• This apparent quality can last for years, but history shows it is not necessarily structural. When capital spending slows or expectations catch up, smaller earnings beats and a slower pace of upward earnings revisions can expose how much of the improvement depends on scarcity related pricing power.
• Active managers need to recognize durable regime shifts without confusing cyclical improvement with structural quality. The opportunity is to separate companies whose economics have genuinely improved from those borrowing quality from the cycle.

Over the past two years, perceived lower-quality companies have not just outperformed higher-quality peers; many have begun to look like them.

That can happen in a generational investment cycle. When vigorous AI-related demand meets limited industry supply, companies gain pricing power. Revenue accelerates, fixed costs are absorbed and margins, returns and cash flow improve. Those are the same metrics investors normally associate with quality, so estimates rise, multiples expand and the stocks perform.

For a time, lower-quality companies can masquerade as high quality. The improvement in the numbers is real, but it may be superficial. Intense competition, limited differentiation, customer concentration and a lack of durable pricing power have not necessarily disappeared; they simply matter less while revenue and earnings surprises keep coming.

High-quality companies—profitable, growing, financially sound and disciplined in their use of capital—have historically earned a return premium. However, the past two years have been an exception with quality’s outperformance eroding from its most recent peak, leading to underperformance for the rolling periods ending in 2026 (Exhibit 1). The market often treats improving quality as durable quality until the earnings-revision cycle says otherwise.

How Tailwinds Create the Appearance of Quality

Consider the sequence. Years of underinvestment leave an industry with too little supply just as AI-related spending steps higher. Scarcity temporarily softens competitive pressure and creates pricing power, lifting revenue and incremental margins. Higher margins improve returns, cash generation and balance sheets. The market then rewards those improving quality metrics with higher earnings estimates and, often, a higher multiple.

That is why a company with a non-durable edge, heavy competition or uneven execution could outperform in this environment. Its structural weaknesses are still there, but they are temporarily hidden by stronger reported economics, which may be supported by unsustainably high demand. The business may not be higher quality, but it is reporting numbers that resemble it.

The size and persistence of the surprise become the key variables. A company can keep working as an investment as long as revenue, margins and earnings are improving faster than the market expects. Each better than expected result, or “beat,” lifts the next forecast and raises the hurdle for the following quarter. That combination has been especially powerful across selected beneficiaries in memory, optical networking and engineering and construction (Exhibit 2).

Eventually, Expectations Catch Up

Strong operating results do not have to reverse for a stock to stop working, they only have to become less surprising. After several quarters of earnings beats, consensus estimates rise, investors pay a higher multiple and the bar moves closer to the company’s actual earnings power.

This is where the second derivative matters: revenue may still be growing, margins may still be expanding and earnings may still be topping estimates – but by less. A slower pace of upward estimate revisions, smaller upside surprises or flattening incremental margins can signal that investor expectations are catching up to reality.

At that point, the market will start to ask what is structural. Can the company retain pricing power as supply improves? Is the margin gain tied to a durable competitive advantage or simply to scarcity? Can returns stay elevated when customers have more alternatives and input costs are harder to pass through? Importantly, a lower forward P/E does not necessarily mean expectations are low; it may mean that substantial earnings growth is already embedded in the forecast (Exhibit 3). A stock can disappoint even while earnings grow if it no longer clears the higher bar.

The optical networking boom of 1999–2001 offers a useful, if imperfect, comparison with today’s AI infrastructure buildout. Internet traffic was surging, telecom companies had easy access to capital and new equipment allowed them to send much more data over the fiber already in the ground, creating a rush of spending and unusually positive results for suppliers such as JDS Uniphase (JDSU). When carriers later cut spending after building more network than demand required, orders fell, customers worked down inventories and JDSU’s adjusted operating margin—after removing inventory write-downs, restructuring charges and other one-time items in each year—fell from 32% in fiscal 2000 to 25% in fiscal 2001 and -29% in fiscal 2002; its split-adjusted stock price fell from a quarterly high of US$146.53 to a quarterly low of US$2.28 (Exhibit 4). While JDSU’s decline was unusually severe, the broader telecom equipment sector also experienced significant revenue pressure, inventory corrections, restructuring actions and margin compression, supporting the broader cyclical comparison.