$10,000 Invested in XLK on New Year’s Eve Is Worth $13,434 Today

Ten thousand dollars dropped into Technology Select Sector SPDR Fund (NYSEARCA:XLK) at the close on December 31, 2025 was worth about $13,434 at the close on June 4, 2026. The same ten grand in SPY was worth about $11,102. That is the entire "three to one" story, 34% for the tech SPDR against 11% for the S&P 500, and it is why every tech-watching group chat has been screenshotting the same chart for weeks. The math is total return through the most recent close, dividends in. XLK pays a yield of about 0.62%, so for a five-month window the distributions barely move the needle. What is doing the moving is price. Even QQQ, the tech-heavy Nasdaq-100 proxy that usually serves as the "but tech is winning" benchmark, is up 21% year to date, which means XLK is beating QQQ by roughly fourteen points too. A sliver of the tech complex is pulling everything. Two stocks doing most of the work Open the hood and the engine is obvious. The top three holdings, NVIDIA (NASDAQ:NVDA | NVDA Price Prediction), Apple (NASDAQ:AAPL), and Microsoft (NASDAQ:MSFT), account for 40% of XLK’s net assets. NVIDIA alone is 14.93% and Apple is 13.23%. Combined, that is 28.16% of the fund concentrated in two names. The semiconductors and semiconductor-equipment bucket as a whole runs 38.64% of the portfolio, with software at another 32.7%. When you buy XLK, you are buying chips and code with a heavy lean on two companies that happen to be the two most valuable on the planet. NVIDIA is up 17% year to date and Apple is up 15%. Those numbers look ordinary next to XLK’s 34%, until you remember that the fund is rebalanced to systematic concentration rules and that the other 60% of the basket (Broadcom, Palantir, AMD, Oracle, Micron, and so on) is also riding the same AI infrastructure wave. The State Street fact sheet still shows an expense ratio of 8 basis points, so virtually none of the return is being clipped on the way out. The mechanism is AI capex, and the receipts are quarterly NVIDIA reported Q1 FY27 on May 20, 2026. Revenue came in at $81.6 billion, up 85% year over year, with Data Center revenue at $75.2 billion growing 92%. Networking inside that segment, which is the InfiniBand and NVLink fabric that ties Blackwell systems together, tripled to $14.8 billion, up 199%. Non-GAAP gross margin held at 75.0% at this scale, which is the part that has analysts checking their models twice. Jensen Huang, on the call, described it the way Jensen Huang describes everything. "The buildout of AI factories, the largest infrastructure expansion in human history, is accelerating at extraordinary speed." The line that matters more for forward modeling is the balance-sheet line. NVIDIA has $119.0 billion in total supply-related commitments, the company guided Q2 to $91.0 billion in revenue, and that guide explicitly excludes any China data-center compute. So the run is happening without the China contribution that used to show up in the year-ago quarter, when $4.6 billion of H20 product was still flowing. Strip China out and the comp gets harder. They are doing it anyway. Apple, the other engine, reported Q2 FY26 on April 30, 2026 with revenue of $111.2 billion, up 16.6%, and iPhone revenue of $56.99 billion, a March-quarter record powered by the iPhone 17 lineup. Services hit a fresh all-time high at $30.98 billion. That is the eighth consecutive EPS beat. Tim Cook’s framing, that it was Apple’s "best March quarter ever" with double-digit growth across every geographic segment, is exactly the boring, dependable income statement an index fund wants from its second-largest holding while its largest holding goes vertical. The capital-return side tells you what management thinks of valuation here. NVIDIA authorized an additional $80 billion in buybacks and raised the quarterly dividend from a penny to $0.25. Apple authorized a fresh $100 billion buyback and bumped the dividend 4%. Two of the largest balance sheets in the world are choosing to retire their own shares rather than do anything else with the cash, and that vote keeps lifting per-share metrics across the entire XLK basket. What has to keep being true The forward-look question is whether the conditions that produced a 3:1 spread over the S&P in five months can hold for the next twelve. Three things have to keep going. First, hyperscaler capex. NVIDIA has publicly framed AI infrastructure spending at $3 to $4 trillion annually by the end of the decade, with the largest hyperscalers contributing about $600 billion this year. Hyperscale customers are roughly 50% of NVIDIA’s Data Center revenue, so the leading indicator a reader can actually watch is the next round of capex guides from Microsoft, Amazon, Google, and Meta. If those numbers flatten or get walked back, NVIDIA’s $119 billion of supply commitments turns from a confidence signal into an inventory problem. Second, the China line stays out of the model. NVIDIA is currently delivering 85%+ growth with zero H20 shipments to China and a Q2 guide that assumes the same. A reopening would be a tailwind. A further restriction would be a tax on a single line item that the rest of XLK cannot easily replace. Third, Apple has to keep printing the boring quarters. Greater China at $20.5 billion in Q2 and an installed base above 2.5 billion active devices is what keeps Services compounding. If iPhone 17 demand normalizes the way iPhone cycles eventually do, Apple becomes a stabilizer rather than a contributor, and XLK’s outperformance over SPY narrows toward whatever NVIDIA does next. A 3:1 spread over the S&P is not a baseline to expect through year-end. It is the product of a specific concentration meeting a specific capex regime, on top of valuations that are no longer cheap. The setup is broadly intact, the leading indicators (hyperscaler capex guides, NVIDIA’s quarterly Data Center revenue report, Apple’s Services growth rate, and any move on China compute restrictions) are all observable on a release calendar, and XLK at 8 basis points is the cleanest way to own the trade. The thing to remember is that you are effectively buying NVIDIA and Apple with a software hedge. As long as that sentence keeps describing the world, XLK keeps lapping the index. The day it stops describing the world, you will know from the capex guides before you know from the chart.