S&P 500 vs Nasdaq 100

SPY is the SPDR S&P 500 ETF Trust, the oldest US-listed ETF (launched 1993), tracking the S&P 500 Index of 500 large-cap US companies selected by a committee based on size, liquidity, and earnings. QQQ is the Invesco QQQ Trust, launched 1999, tracking the Nasdaq-100 Index of the 100 largest non-financial companies listed on the Nasdaq exchange.

Both are market-capitalization weighted. The critical difference in selection: the S&P 500 can include companies listed on any US exchange and has a financial-services sector. The Nasdaq-100 is restricted to Nasdaq-listed stocks and excludes financials entirely, which mechanically tilts it toward technology, consumer discretionary, and communication services. The overlap between the two indices is substantial at the top end. Most of the Magnificent Seven (Apple, Microsoft, Nvidia, Amazon, Alphabet, Meta, Tesla) are in both, but they carry meaningfully higher weights in QQQ.

SPY sector weights as of early 2026: technology 34.2%, financials 12.7%, healthcare 9.7%, consumer discretionary 10%, industrials 8%, communication services 8.8%, the rest spread across consumer staples, energy, utilities, real estate, and materials. QQQ sector weights: technology 51-58%, communication services 16.2%, consumer discretionary 13.1%, with healthcare, industrials, and utilities making up the remainder.

The practical meaning of those weights is that QQQ is a concentrated bet on the "digital economy" plus adjacent consumer names, while SPY is a broader claim on the US large-cap economy. When technology leads the market, QQQ outperforms by definition. When financials, healthcare, industrials, or energy lead, SPY outperforms. This relationship is mechanical and predictable at the sector level; the interesting questions are about timing and the drivers of sector leadership.

QQQ is substantially more concentrated than SPY at the top. As of early 2026, QQQ's top 5 holdings (Nvidia ~9%, Apple ~7.6%, Microsoft ~5.7%, Amazon ~5.5%, Broadcom ~4.5%) total approximately 33% of the fund. Adding Alphabet ~6.7%, Meta ~3.7%, and Tesla ~3.5% pushes the top-8 concentration above 47%. SPY has the same companies near the top but at lower individual weights, typically capped more tightly by the S&P 500 methodology, giving the top-10 roughly 30-35% of the fund.

This concentration has two implications for the pair comparison. First, an idiosyncratic event at a single top holding affects QQQ more than SPY (an Nvidia earnings miss moves QQQ roughly 40% more than SPY on equal-weight terms). Second, when a single theme dominates (AI, in 2023-2025), the concentration works for QQQ; when the theme disperses, it works against QQQ.

QQQ has outperformed SPY on a total-return basis over 10 and 15-year windows, though not uniformly. Over the 10 years ending early 2026, QQQ returned approximately 20.45% annualized versus SPY's 14.85%. Over 5 years, the gap is wider because the 2020-2024 period included both the COVID-era growth boom and the 2023-2025 AI rally.

Over a full 25-year window that includes the 2000-2002 dot-com crash, QQQ's advantage shrinks dramatically. An investor who bought QQQ at the March 2000 peak waited until August 4, 2016 to see nominal breakeven, a 16-year recovery. The same investor in SPY was nominally whole by 2007, then took another five years to recover from the 2008 crash. Recency effects exaggerate QQQ's apparent outperformance; longer windows that include tech cycles beginning-to-end show the volatility trade-off more clearly.

The deepest historical divergence between the two benchmarks was the 2000-2002 dot-com crash. QQQ peaked on March 24, 2000 at a closing price of $117 and bottomed on October 7, 2002 at $20, a drawdown of 83%. SPY peaked in August 2000 near $150 and bottomed at $78 on October 2, 2002, a drawdown of about 48%.

The 35-percentage-point gap between the two drawdowns is the historical upper bound for the sector-concentration risk QQQ carries. It reflected the fact that internet-stock valuations unwound while the broader economy had already absorbed much of the Y2K-related spending and moved into a milder recession. Any analysis of "is QQQ better than SPY" that ignores this episode is selling on a short window and understating the downside distribution.

The 2022 Federal Reserve tightening cycle was the second-worst modern divergence between QQQ and SPY. QQQ finished 2022 down 32.58%, while SPY finished down approximately 18%. The gap reflected QQQ's greater sensitivity to rising real yields: long-duration growth stocks, which dominate QQQ, lose more value when discount rates rise because more of their intrinsic value sits in future cash flows.

This rate-sensitivity is the main structural wedge between the two ETFs beyond sector composition. When real yields rise meaningfully (as in 2022, when 10-year TIPS went from about negative 1% to above positive 1.5%), QQQ underperforms even without specific tech-sector bad news. When real yields fall, QQQ outperforms.

QQQ's 2023 return of 54.76% (versus Nasdaq-100 Index total return of 55.13%) was among its strongest calendar years since inception. The drive came from AI-related demand and the concentration of AI infrastructure names (Nvidia, Microsoft, Broadcom, Meta, Alphabet, Amazon) inside the Nasdaq-100. QQQ returned 25.60% in 2024 and continued strong performance through 2025 as AI capital expenditure and model deployment expanded.

The trailing one-year performance as of April 2026 reflects the cumulative effect: QQQ up roughly 42% versus SPY up 31%. The question increasingly asked by allocators is whether AI-driven tech concentration resembles the late-1990s setup (where valuations ran ahead of realistic earnings) or the genuine 1996-1998 productivity-boom setup (where earnings eventually caught up). The answer is not knowable in real time.

QQQ runs meaningfully higher volatility than SPY. Trailing 12-month realized volatility for QQQ is approximately 15.6% versus 11.0% for SPY, a ratio that has been stable for decades. The volatility translates directly to drawdown distributions: QQQ's maximum drawdown since inception was 83% (dot-com); SPY's was 55.19% (2008 financial crisis).

The practical consequence is that a 100% QQQ allocation is risk-equivalent to roughly 140-150% of an SPY allocation on the same volatility target. Allocators who compare the two as substitutes often under-size QQQ relative to their actual risk tolerance, which is why QQQ drawdowns frequently produce capitulation selling at exactly the wrong moment. Reducing to risk-weighted sizing (about 70% QQQ for every 100% SPY on a matched-risk basis) flattens the comparison.

SPY's expense ratio is 0.0945% (0.09 bps annually). QQQ's was 0.20% until December 22, 2025, when Invesco completed a long-planned conversion of QQQ from a Unit Investment Trust to an open-end ETF structure following shareholder approval on December 19, 2025. The expense ratio dropped to 0.18%, and the restructure removed several operational drags specific to the old UIT format, including inability to loan out securities for income, limited flexibility on cash holdings, and inability to use derivatives for brief rebalancing periods.

For ordinary investors, the restructure is a small tailwind worth a few basis points per year of tracking error improvement. For tax optimization, QQQ (like SPY) is a highly tax-efficient ETF that rarely distributes capital gains thanks to in-kind creation and redemption. Post-2025, QQQ's structure now matches SPY's, and the historical advantage SPY had on operational flexibility is largely closed.

The honest answer is that most investors should hold both in some combination. A common framework: SPY as the core US equity allocation, QQQ as a growth-tilt overlay sized to reflect real-yield views. In regimes of falling real yields (early recovery, Fed easing), tilt toward QQQ. In regimes of rising real yields (mid-cycle tightening), tilt toward SPY or neutralize the tilt.

For single-holding choices, SPY is the textbook default because of its lower volatility, broader sector exposure, lower cost, and smaller historical drawdowns. QQQ makes sense as a single holding only for investors who explicitly want concentrated tech/growth exposure, accept the 80%+ drawdown distribution, and have a long enough horizon that a 10-15 year recovery period from a bubble top does not force capitulation. Neither is universally "better"; they are different risk and return distributions appropriate to different allocation decisions.