When Revenue Acceleration Overwhelms Quality

• 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.

"The market often treats improving quality as durable quality until the earnings revision cycle says otherwise." 

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 post recent peak, leading to underperformance for most of the rolling periods ending in 2026 (Exhibit 1). The market often treats improving quality as durable quality until the earnings-revision cycle says otherwise.

Exhibit 1: Quality Has Been Challenged Recently

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 can outperform in this environment. Its structural weaknesses are still there, but they are temporarily hidden by stronger reported economics, which are 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).

Exhibit 2: Scarcity Beneficiaries Delivered Extraordinary Two-Year Returns

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.

Exhibit 3: Consensus Already Assumes a Substantial Earnings Step-up

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 strong results for suppliers such as 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 $146.53 to a quarterly low of $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.

Exhibit 4: Optical Networking Profits Reversed When Carrier Spending Fell

Quality-oriented managers should not dismiss this episode as an irrational low-quality rally. Generational investment cycles can persist for years, and the compounding cost of missing a durable trend can be substantial.

Quality is an attribute, not a catalyst, and it cannot become a crutch. Durable competitive advantages usually give a company the best opportunity to reinvest, grow and raise returns in a defensible way. However, when those advantages are not producing growth, quality alone is unlikely to drive a stock higher. Active managers must adapt as the evidence changes, recognize generational regime shifts and consider new sources of growth and innovation as they emerge.

In some cases, the improvement is more than cosmetic. Sustained pricing power can strengthen balance sheets, fund reinvestment and raise through-cycle returns. ClearBridge’s work on AI’s physical constraints points to emerging earnings power across semiconductors, memory, networking, optical connectivity and power infrastructure.

"The eventual narrowing of earnings surprises is not only a risk; it is also a stock selection opportunity.  "

The task is to separate durable compounders whose existing advantages are being amplified by AI-driven investment and the buildout of its physical infrastructure; cyclicals whose through-cycle economics are genuinely improving; and fragile beneficiaries whose higher margins and returns still depend on scarcity, pricing and repeated upside surprises.

The eventual narrowing of earnings surprises is not only a risk; it is also a stock selection opportunity. When expectations catch up, the market becomes less willing to treat every beneficiary as high quality. Active managers can then separate companies whose competitive position, pricing power, balance sheet and returns have structurally improved from those whose metrics were temporarily flattered by the cycle.

Static quality scores alone will not do it. Deep fundamental research into the durability of pricing power; incremental margins; cash conversion; customer concentration; recurring revenue; switching costs; reinvestment discipline; and how much of the improvement is already embedded in consensus estimates and valuation will matter. The key question is simple: what remains when revenue growth normalizes?

Suppliers across memory, optical connectivity and power equipment have now announced meaningful capacity expansions. Although those additions will take time, they matter because new supply can eventually ease scarcity and pricing power — the forces currently lifting revenue, margins and returns. At the same time, AI infrastructure spending is consuming most or all of several hyperscalers’ operating cash flow. Capital markets can fund spending ahead of cash generation for a time, but not without limit—or without growing evidence that the investment will earn an adequate return.

We do not profess to know when this generational investment cycle will slow. But as supply responses broaden, expectations rise and the funding burden grows, we believe the next upside surprise is more likely to be smaller than the initial shock. The next phase of the cycle should therefore reward discernment over exposure—not simply which companies benefited from scarcity, but which remain high quality after scarcity and surprise begin to fade.