Wall Street financial giant JPMorgan's latest Retail Radar research report reveals that retail fund flows in the US stock market are showing a trend of cooling total retail capital, increasingly concentrated stock buying in a handful of AI computing leaders, and rising investment appeal of long-duration US Treasuries after record-selling pressure.
JPMorgan's exclusive retail flow tracking through September 30 reveals that retail investors have not expanded their buying of the broader US tech sector and Philadelphia Semiconductor Index components in sync with the long-term flow trends seen throughout the year. Recent advances in cutting-edge AI agents and large language models represented by Muse, Astra, and Anthropic Claude have provided an important technical foundation for large-scale commercialization of AI applications across industries and continued surge in AI computing demand, yet only Nvidia and SanDisk, the two leaders in the AI computing supply chain, remain favored by retail flows, while Intel, SpaceX, and some AI computing infrastructure stocks surprisingly saw net selling.
JPMorgan's calculated retail flows show that excluding the Magnificent Seven, tech remains the only sector receiving net buying, while AI-related hot stocks such as Intel, SpaceX, and Oracle were reduced, highlighting capital divergence within the same AI computing supply chain investment theme.
From September 24 to 30, JPMorgan's latest estimates show retail investors net bought $4.1 billion, about 40% below the average weekly level of the past 12 months, with individual stock net buying at only $700 million. However, Nvidia and SanDisk received $1.286 billion and $327 million in retail net buying respectively, totaling $1.613 billion, meaning that net selling of approximately $913 million across other individual stocks offset part of the buying.
Memory chip components for AI data center server clusters and AI GPUs remain the clearest supply bottlenecks in the AI computing supply chain. Market research firm TrendForce estimates that server DRAM contract prices will cumulatively rise about 270% in 2026, enterprise SSD prices will cumulatively rise about 235%, and HBM contract prices may still rise 70% to 140% in 2027, continuing to show doubling growth. These data reflect the combined effect of continued AI computing demand expansion and memory chip price increases. TrendForce calculations also show that NVL72 rack shipments covering Blackwell and Vera Rubin platforms are expected to grow more than 50% year-over-year in 2027. Its market research charts show that related system output value is expected to rise from approximately $226 billion in 2026 to $711 billion in 2027, a significant year-over-year increase of 214%.
Meanwhile, long-duration US Treasuries, which have been under price pressure since September, have seen significant contrarian accumulation by retail investors, with long-term Treasury ETFs experiencing record standardized buying skew. The long-duration Treasury ETF with ticker TLT became the most prominent focus of retail flows. Standardized buying skew refers to using historical volatility as a yardstick to measure how strong retail net buying is relative to historical norms. JPMorgan's latest calculation of TLT reaching plus 6.1z means retail investors' dip-buying tendency is unusually strong.
Dramatic narrowing of stock buying is perhaps the most fitting summary of retail flows: incremental retail capital is weakening while stock selection concentrates on a few AI computing super leaders, memory chip champions, and large tech names with strong cash flow fundamentals. On the bond market side, another clear signal has emerged: as intensifying Middle East geopolitics drives energy inflation higher and pushes long-term Treasury prices under significant pressure, ultimately keeping yields on 10-year and longer maturities at elevated levels, with Treasury yields and prices moving in opposite directions, long-term US Treasury ETFs saw a rare $260 million in retail net buying during the week.
Retail investors are simultaneously participating in the AI super bull market and the long-duration yield trade, with the former seeking strong earnings realization amid explosive AI computing industry expansion and the latter pursuing the highest US bond yields in over two decades plus price elasticity from future yield declines.
Retail Buying Narrows Fronts: Picking Leaders in Stocks, Choosing Beaten-Down Long-Duration US Treasuries in Bonds
JPMorgan's Retail Radar research report shows retail investors are still net buyers, but the incremental capital supporting the market has clearly weakened. Retail flows deserve increasingly close investor attention because they provide marginal buying in the spot market and also influence price elasticity through options trading. The brokerage channel transaction proxy indicators listed in JPMorgan's report accounted for approximately 25% of total US stock and ETF trading volume in the latest data period of June 2026, while retail options market participation remained near historic highs through the end of September.
The most valuable investment signal from this report is that amid continued expansion of overall AI computing industry demand, retail stock buying has instead become more concentrated in AI computing leaders, while long-duration US Treasuries have become another clear allocation direction. Micron's latest earnings report showed fourth-quarter fiscal 2026 revenue reached $54.229 billion, up approximately 379% year-over-year, with next-quarter revenue guidance of $61.5 billion plus or minus $1.5 billion. The Philadelphia Semiconductor Index subsequently rebounded 1.59% on October 1, providing earnings and price-level confirmation of the strong prosperity in the AI memory and computing industry.
More significantly, Micron executives stated directly on the earnings call that they cannot see the end of supply-demand balance. The company has signed 26 long-term agreements locking in approximately $150 billion in long-term orders and expects memory chip supply-demand markets in 2027 and 2028 to be even tighter than the record tight levels of 2026. With strong cash flow support, Micron announced capital expenditures of $25 billion in the first half of fiscal 2027 and committed to returning 100% of excess cash to shareholders in the future.
JPMorgan's calculated data shows that from September 24 to 30, retail net buying was $4.1 billion, about 40% below the $6.8 billion weekly average of the past 12 months, with ETF net buying at $3.4 billion and individual stocks at only $700 million. Weekly total retail flows in the US stock market were at the 12th percentile historically, with ETF and individual stock flows at the 3rd and 35th percentiles respectively. Daily buying roughly stayed at the 15th to 30th percentile, with limited response to weekly price changes, ultimately making September the weakest month for retail activity since December 2024. Weakness here refers to declining net buying intensity, while total capital remains a positive factor.
Within market allocation, large-cap broad market equity ETFs absorbed approximately $1.4 billion in retail capital, while covered call strategy ETFs, multi-cap broad market ETFs, and EAFE international equity ETFs absorbed $188 million, $169 million, and $148 million respectively. However, overall flows into international equities, crypto assets, and commodity ETFs weakened compared to before. JPMorgan's research report also listed non-retail category futures traders buying approximately $33 billion during the week, mainly concentrated in S&P 500 and Nasdaq futures, highlighting that observing rebound momentum requires considering different trading groups simultaneously.
In terms of market structure, declining retail net buying weakens support for market breadth expansion to more stocks, while indices may still be driven by large-cap weighted stocks and other capital channels. Long bonds have become a prominent direction for retail contrarian positioning, behind which is a repricing of interest rates and financing costs. Long-term US Treasury ETFs received $260 million in net buying during the week, with the long-end buying skew recorded at plus 5.4z and TLT at plus 6.1z. However, it should be noted that these figures measure abnormal strength relative to history and cannot be interpreted as records for yields, allocation ratios, or absolute dollar net inflows.
In the environment described in JPMorgan's research report where US Treasury yields are testing 22-year highs, buying TLT means increasing exposure to long-duration US Treasury yields and positioning for price rebounds in 10-year-plus US Treasuries. This can be understood as simultaneously seeking higher yields and price elasticity from future rate declines, though actual motivation cannot be determined solely from fund flows. The high-rate/high-yield curve also draws a financing cost dividing line within stocks: retail flows into small businesses more dependent on short-term, floating-rate bank financing are clearly under pressure, with the trend of retail net buying Russell 1000 and net selling Russell 2000 accelerating since March. Median short interest ratios for Russell 2000 and S&P 500 components stand at the 99.8th and 96.2nd percentiles historically, with idiosyncratic indicators for both stock categories also near the 92nd percentile, highlighting individual stock divergence.
The energy sector within stocks shows another kind of cooling: after the Saudi east-west pipeline restoration, energy stock and ETF buying turned to slight net selling. Although oil prices remain around $100, Middle East crude oil exports have recovered to 98% of pre-war levels, while refined product exports have only recovered to 58%. This indicates retail investors are separately trading AI growth, the sharp rebound expectations for US Treasury prices amid surging yields, and energy supply recovery, with clearly structured differences in capital direction.
Computing Demand Spreads but Retail Stock Buying Converges: Has AI Trading Entered the Name-Calling Era?
The AI computing theme that has supported the US stock market's super bull run since 2023 remains an important stock theme for retail investors, but having an AI tech label is no longer sufficient to explain the buying list. The report clearly states that retail investors prefer semiconductors and hardware, with software relatively lagging. Excluding the Magnificent Seven, the tech sector still saw $194 million in net buying during the week, while all other sectors saw significant net selling. The top five retail net buying names through the week were Nvidia at $1.286 billion, Tesla at $514 million, SanDisk at $327 million, Alphabet at $153 million, and Amazon at $146 million. Micron received $19.8 million in retail net buying on September 30 alone, with the report finding no abnormal pre-earnings accumulation.
Particularly noteworthy is that Nvidia's net buying alone exceeded the $700 million total net buying across all individual stocks in the market, indicating that selling in other stocks offset a considerable portion of AI computing leader buying. On the other hand, Intel, SpaceX, Oracle, Nebius, and Bloom Energy were respectively net sold $199 million, $171 million, $88 million, $60 million, and $54 million by retail investors last week. Apple, Microsoft, and Meta also saw net selling. Within AI investment theme baskets, AI computing core infrastructure supply, data center construction and computing leasing, data center electrification, growth stocks, AI software monetization, and US companies with high China revenue exposure remain areas of long-term retail interest.
In derivatives markets, the one-month rolling average of retail options trading share, while off its high, remains at the 93.5th percentile historically, with an actual share of approximately 23% still at historically elevated levels. Tesla, Meta, Micron, Nvidia, AMD, and SanDisk remain options trading focal points. Nvidia topped the spot net buying list but also ranked high in options Delta selling. SpaceX saw spot net selling but remained active on options Delta buying lists, showing divergence between trading instruments. JPMorgan noted that the latest social media discussions among retail investors focused on names including CIFR, WULF, HUT, AKAM, and CoreWeave, but actual retail selling in high short-interest stocks surged, reducing overall short squeeze risk. ALOY, EU, and EVTL are event cases requiring separate observation.
Clear divergence also exists in non-AI areas: Carnival's positive earnings drove a 13% stock price increase, yet retail investors net sold $14.4 million. Veeva gained a major pharmaceutical client win and its stock rose, yet still saw net selling. KOD's successful clinical trial attracted $5 million in buying, with abnormal strength reaching 11.4z. MGM's consideration of acquiring PPLI, which holds approximately 27% of its shares, activated an event trade with potential buyback characteristics.
Computing demand spreads while stock buying converges is JPMorgan's summary phrasing for retail fund flows. Nvidia and SanDisk, as AI computing core infrastructure and memory chip beneficiaries amid explosive AI computing demand expansion, became representatives of concentrated buying, with Nvidia corresponding to AI computing and SanDisk to NAND memory. Meanwhile, long-duration US Treasuries under sustained price pressure have seen significant contrarian accumulation by retail investors.
OpenAI is in talks to raise funding at approximately $1.4 trillion pre-money valuation. Potential IPO investors for Anthropic have given valuation judgments of $1.8 trillion to $2 trillion, with expectations of matching or exceeding SpaceX's offering scale, though these remain financing negotiations and Silicon Valley venture capital IPO expectations. Compared to AI application valuation narratives, the AI computing supply chain has more concrete strong demand evidence: media disclosures show Anthropic's approximately $518 billion infrastructure arrangements over the next decade, of which about 80% has non-cancellable or take-or-pay characteristics. Anthropic's 2025 revenue grew to approximately 12 times the prior year at nearly $4.6 billion, with operating losses exceeding $8 billion. Computing and infrastructure spending reached $7.33 billion, approximately three times that of 2024, accounting for about 58% of total operating expenses of $12.65 billion. Nvidia's latest quarterly data center revenue was $89 billion, up 117% year-over-year. South Korea's September semiconductor exports were $60.3 billion, up 262.8% year-over-year, with officials noting increases in both memory export volumes and contract prices.
From the perspective of massive AI inference workloads, Muse's continuous background execution and Astra's improved capabilities on complex computer tasks have expanded the scope of work AI can undertake. Multi-step tasks, tool calling, and parallel agents may also increase model calls and context processing per user. Anthropic has observed in its research systems that Token usage for multi-agent tasks is approximately 15 times that of regular chat. What can be deduced is that with comprehensive penetration of cutting-edge AI agents like Muse, AI computing demand will spread along GPU computing, HBM and DRAM capacity and bandwidth, KV cache and SSD storage, CPU tool execution, and network transmission. Total resource demand ultimately depends on the combined effect of task volume growth and per-task efficiency improvement. This also explains why AI hardware still has fundamental appeal but cannot guarantee that all related stocks simultaneously receive incremental capital.
JPMorgan's data supports that retail investors continue selectively buying AI computing and memory leaders while increasing long-term US Treasury exposure. For a comprehensive return to tech, high-beta small caps, or all AI infrastructure stocks, retail fund flow evidence does not yet hold.