S&P 500: The Winners and Losers of September 2026
The S&P 500 slipped about 0.5% in September 2026, a mild decline that badly understated how weak the month actually was for the typical stock. Roughly 78% of index constituents finished lower, and nine of the eleven sectors declined. Technology and communication services were the only exceptions, which is why the index held up while the median stock did not. The backdrop made that split sharper. The Federal Reserve raised rates for the first time in roughly three years, the 10-year Treasury yield ended the quarter near 5.3%, and 30-year mortgage rates moved back above 7%. Those conditions should have been difficult for long-duration growth stocks, but AI infrastructure enthusiasm was strong enough to let semiconductors decouple from everything else. Meanwhile, a regulatory campaign aimed at mortgage credit costs produced the single worst monthly decline of the year. In this article, we break down the top winners and losers in the S&P 500 for September 2026 and examine the key drivers behind their stock movements.
Top S&P 500 Performers in September
Among the top performers were Moderna (MRNA), Intel (INTC), and Advanced Micro Devices (AMD). Two of the three were driven by a widening view of what AI demand actually covers. Here’s a closer look at why each of these stocks stood out:
Moderna ($MRNA) +37.22%
Moderna, a biotechnology company built around messenger RNA technology, rose roughly 37% in September, making it the index’s best performer for a second consecutive month. The unusual feature of the move is that the catalyst behind it did not occur in September at all.
The fundamental event was the August 19 announcement by Moderna and Merck that the Phase 3 INTerpath-001 trial of intismeran autogene combined with Keytruda met its recurrence-free survival and distant metastasis-free survival endpoints in resected high-risk melanoma. What made that result so consequential is that the development program extends well beyond melanoma into multiple tumor types. Investors were not capitalizing a single indication. They were raising the probability assigned to Moderna’s entire oncology platform, and that repricing took more than one month to work through.
Moderna entered September around $140 and reached roughly $203 by September 29, at which point the monthly gain exceeded 40%. No comparably transformative new clinical result appeared during the month, which makes the September advance a mixture of genuine fundamental revaluation and momentum.
The final session provided the first real pushback. On September 30, Citi downgraded the shares from Hold to Sell while raising its price target from $60 to $80, arguing that a roughly $80 billion valuation already assumed intismeran would succeed commercially across several difficult cancers rather than melanoma alone. The stock fell about 5.4% to close at $192.57. That downgrade is useful as a marker, because it shows the debate had shifted from whether the melanoma program works to how much broader success was already priced in.
Intel ($INTC) +34.32%
Intel, the semiconductor company that designs and manufactures CPUs and is attempting to build a contract chipmaking business, rose roughly 34% in September on a sequence of reinforcing catalysts rather than any single announcement.
The month started from $89.51 and produced three distinct legs, each confirmed by unusually heavy volume. On September 8, Intel rose about 9.1% to $104.47 on roughly 140 million shares, against 65 million on the last day of August, amid an analyst upgrade and reports that Intel could raise CPU prices by as much as 10% in a tight supply environment. The same day, Intel Foundry announced a collaboration with ASML to accelerate industry readiness for High-NA EUV lithography, the manufacturing technology underpinning its most advanced process nodes. On September 17, the stock gained 7.7% to $108.80 on about 150 million shares, following reports that SK Hynix was in exploratory discussions about producing memory chips in the United States, potentially using or leasing capacity at Intel’s delayed Ohio complex. Such an arrangement would improve utilization and financing economics at one of Intel’s most capital-intensive projects.
The largest leg came on September 21, when Intel jumped roughly 12.1% to $121.78 on approximately 192 million shares, nearly three times the August 31 volume. That session was driven by a broader realization that inference workloads increase demand for conventional server CPUs, not only for GPUs, which reframed Intel’s core product line as an AI beneficiary rather than an AI casualty.
Intel closed September at $120.23. The caveat worth holding onto is that several of these catalysts were analyst commentary, reported pricing actions, and exploratory discussions rather than signed contracts or reported earnings.
Advanced Micro Devices ($AMD) +29.96%
Advanced Micro Devices, a semiconductor company specializing in CPUs and GPUs for data centers, PCs, and gaming, rose roughly 30% in September, crossing a $1 trillion market capitalization along the way and then using that valuation strategically.
AMD’s advance rested on a stronger fundamental base than Intel’s. Its prior quarter had already shown data center revenue more than doubling year over year, with EPYC demand accelerating and Instinct deployments scaling. The September enthusiasm was therefore building on a measurable operating trend rather than sentiment alone. Around September 21, AMD gained close to 10% in the same session that lifted Intel, as investors recognized that agentic and inference-heavy workloads require general-purpose CPU compute, networking, and memory alongside accelerators, which broadens the addressable opportunity for AMD’s full portfolio.
On September 28, AMD announced a definitive agreement to acquire World Labs, the spatial intelligence research lab led by Fei-Fei Li, in an all-stock transaction valued at approximately $8.2 billion and expected to close by the end of 2026. World Labs builds models that generate, reconstruct, and simulate interactive 3D environments, along with technology for robotic learning. Li will join AMD as executive vice president and chief scientist. Strategically, the deal pushes AMD further from being a pure silicon vendor toward the integrated hardware and software model that has been central to Nvidia’s competitive position. The all-stock structure also mattered, since AMD effectively spent its appreciated equity rather than $8.2 billion of cash.
Commercial validation followed on September 30, when Hewlett Packard Enterprise disclosed a roughly $1.2 billion Vultr server order based on AMD’s Helios rack-scale platform. AMD rose only about 0.7% that day, which indicates the order confirmed the thesis rather than creating it. AMD closed the month at $611.76.
Worst S&P 500 Performers in September
Among the biggest decliners in the S&P 500 for September 2026 were Fair Isaac (FICO), Gen Digital (GEN), and Equifax (EFX). Two of the three were hit by the same regulatory campaign against mortgage credit costs, though through different economic channels. The third fell for an entirely unrelated reason. Here’s a closer look at what drove their underperformance.
Fair Isaac ($FICO) -48.36%
Fair Isaac, the company behind the FICO credit score that has underpinned U.S. mortgage underwriting for decades, lost nearly half its value in September in what was the cleanest event study of the month. Investors abruptly reassessed the durability of what they had treated as a near-monopoly.
The first shock came on September 4, after the Federal Housing Finance Agency director said he had directed Fannie Mae and Freddie Mac to make the competing VantageScore model available to all lenders rather than a limited set. FICO fell roughly 20% in early trading. On September 9, Fannie Mae and Freddie Mac formally announced broad availability of VantageScore 4.0 for eligible loan deliveries, and the agency signaled it was studying further changes to mortgage credit reporting, including potentially relying on fewer bureau reports. The implication went beyond losing unit share. Competition threatened FICO’s ability to exercise pricing power on every mortgage score it sells.
The decisive turn came at month-end. On September 28, Rocket Mortgage announced it would become the first lender to make VantageScore 4.0 its preferred credit scoring model for all eligible loans beginning in the fourth quarter. After testing 1.4 million credit reports using both models, the company found VantageScore helped more clients qualify while reducing scoring costs, with borrowers who saved money saving an average of $1,600 at closing. That changed the question from whether regulators would permit competition to whether major lenders would actually switch.
On September 29, with the agency moving toward a unified Fannie and Freddie mortgage pricing grid incorporating VantageScore alongside FICO, the stock fell 26.5% in a single session to $617.87, its worst day since 1989. Another 4.1% decline followed on September 30. Worth noting is what the selloff did not mean. FICO has published research arguing its Score 10T model is more predictive than VantageScore 4.0 across all mortgage loan types. Investors were not concluding the product was inferior. They were concluding that regulation and competition could destroy a large share of the economic rent the franchise had historically collected, and mortgage rates above 7% amplified the damage by shrinking transaction volumes at the same time.
Gen Digital ($GEN) -28.97%
Gen Digital, the consumer cybersecurity company behind Norton, Avast, and LifeLock, fell roughly 29% in September for reasons that had almost nothing to do with its existing business. The company had entered the month with improving operating momentum, having reported an 11% rise in both bookings and revenue in August alongside raised full-year guidance, and traded as high as $31.65 on September 3, a 52-week high.
The break came on September 24, when reports emerged that Gen had made an early-stage unsolicited approach for GoDaddy at a valuation of roughly $12 billion. The shares fell about 12% to approximately $23.07. The concern was straightforward. GoDaddy is a materially different business from consumer cybersecurity and identity protection, and a transaction of that size would likely require substantial leverage, equity issuance, or both. Higher Treasury yields made debt-financed acquisition math look more expensive, which amplified the worry.
The decline then continued across multiple sessions rather than reversing, falling another 6.3% on September 25 and a further 3.6% to $20.85 on September 28. That cascading pattern suggests investors were progressively incorporating financing, integration, and strategic fit risk rather than reacting to a single headline. Gen’s own corporate communication during this stretch concerned an AI product initiative rather than any definitive transaction.
September 30 brought a 5.6% rebound to $22.02, leaving the month down roughly 29%. The distinction worth drawing is that unless a transaction ultimately changes Gen’s capital structure or business mix, a substantial portion of this decline could reverse if deal probability falls. That is very different from FICO, where the competitive architecture of the market had already changed.
Equifax ($EFX) -27.45%
Equifax, one of the three nationwide credit bureaus, fell roughly 27% in September, which initially looks counterintuitive. Equifax co-owns VantageScore alongside Experian and TransUnion, so it should in theory benefit from FICO’s troubles. The September selloff shows why that reading is incomplete.
Regulators were not only changing which score lenders use. They were attacking the overall cost architecture of mortgage credit reporting, including whether lenders need reports from all three bureaus at all. Equifax fell alongside FICO on the same September 4 announcement, because the agency director simultaneously criticized credit bureau costs. On September 9, the risk became explicit when reports indicated a single-report model was among the cost reduction options under consideration. A shift away from the traditional three-bureau tri-merge toward fewer reports would reduce bureau transaction volumes even as VantageScore gains share.
Management addressed the tension directly at a September 14 investor conference, where the CEO said VantageScore pricing would remain around $1 through 2027 compared with roughly $10 for FICO, and that broader adoption could reduce credit file costs by about 45% and save the industry more than $1 billion. He simultaneously defended the value of tri-merge reports, citing differences across bureau files and the additional borrower coverage combined data provides. Those two statements capture the problem precisely. VantageScore adoption helps the ecosystem Equifax partly owns, but at $1 per score it generates little direct revenue, while any reduction in the number or price of bureau reports strikes at the core U.S. mortgage credit file economics.
Renewed pressure arrived with the unified pricing framework late in the month, with the stock falling roughly 3% on September 29 and another 2.55% to $137.24 on September 30. Context matters here. Equifax had reported revenue of $1.700 billion in its most recent quarter, up 11%, with mortgage revenue growing 25%, so the business was not entering September in distress. Management noted at the same September conference that mortgage activity remained roughly half of historical levels, and with rates above 7%, the regulatory shock landed on an already depressed volume base.
What the Month Really Said
September divided almost perfectly along a single line: whether the rules governing a company’s market were being rewritten, or whether the definition of its addressable market was expanding.
Intel and AMD benefited from the second. The month strengthened the belief that AI demand is broadening beyond accelerator training into CPUs, inference, memory, systems, and physical AI. AMD had the clearer evidence, with an announced acquisition and a large commercial order, while Intel’s case rested more on analyst validation, reported pricing power, and exploratory discussions. That difference is why Intel’s re-rating remains more sensitive to expectations being confirmed.
FICO and Equifax experienced the first. Regulators and customers directly attacked the economics of businesses that had been protected by entrenchment rather than by ongoing competitive advantage. Those are changes capable of affecting volumes, pricing, share, and margins simultaneously, which is why neither decline should be dismissed as volatility. Gen Digital belongs in neither category. Its business was performing well, and the entire move reflected what investors thought a reported acquisition might do to its balance sheet.
Moderna sits somewhere in between, and is the clearest illustration of why monthly performance tables can mislead. Its September gain was an aftershock from an August trial result, and the September 30 downgrade marked the moment the debate shifted from clinical risk to valuation risk. For a month in which the index fell only half a percent while 78% of its members declined, that is a fitting note to end on. The headline number told you almost nothing about what was happening underneath it.
This article is for informational purposes only and is not investment advice or a solicitation to buy or sell securities. The content is based on publicly available information and reflects the author’s opinions as of the publication date, which may change without notice. All investments carry inherent risks, including the potential loss of principal, and past performance is not indicative of future results. Readers should conduct their own research or consult a financial advisor before making investment decisions. BBAE holds no position in the securities mentioned, nor are they compensated by the companies mentioned.
