Cycle on May 22, 2026

I will now list the open threads for topic `019e4cae-c731-7725-a446-fc1b9c574535` to see if there are specific directions or questions that need to be addressed in this research cycle.

I will now search for recent analyses and indicators regarding AI capex returns (specifically around the $800B-$1.16T estimate) and whether hyperscalers are beginning to see adequate returns.

I will now fetch the Allianz PDF on the "AI capex cycle: war-proof for now" and the Substack article "What To Watch in 2026 To Evaluate The AI Bubble" to understand the key early indicators that would confirm or refute whether AI capex will produce adequate returns. Let's start with those.

The fetch for Allianz PDF timed out. I will try the Investing In AI Substack link first, and also search for other high-quality articles discussing the indicators of AI capex returns. Let's fetch the Substack article first.

This Substack article is extremely rich in details! It outlines specific early indicators to confirm or refute whether AI capex will produce adequate returns. Let's summarize the key indicators mentioned in the article: 1. **Refuting indicators (Bubble bursting / inadequate returns):** - Hyperscaler capex decelerating more than 20% year-over-year (reverting toward historical norms). - Enterprise AI production deployment rates stalling below 15% (currently only 8.6% of enterprises have AI agents deployed in production, and many are in "pilot purgatory"). - Nvidia's revenue concentration from its top four customers exceeding 70% (currently four customers contribute 61% of Nvidia's revenue, suggesting a tight circular loop/capex echo chamber). 2. **Confirming indicators (Solidifying floor / adequate returns):** - Non-tech sector AI spending (specifically healthcare and manufacturing) exceeding $20 billion annually. - Enterprise AI deal conversion rates sustaining above 45%. - Fortune 2000 AI budgets growing 30% or more with documented productivity gains (such as healthcare reports of 3.2x ROI, financial services 4.2x returns, manufacturing 61% reporting decreased costs). Let's do some more searches to see if we can find additional perspectives or recent (May 2026) reports from Goldman Sachs, Morgan Stanley, or other financial institutions regarding AI capex returns, hyperscaler monetization, or indicators. I'll search for Goldman Sachs or other analyst views on AI Capex ROI in 2026.

This is great! Let's fetch the Goldman Sachs article: "The Assumptions Shaping the Scale of the AI Build-Out" or "Why AI Companies May Invest More than $500 Billion in 2026", and also the Fortune piece from May 6, 2026: "Goldman Sachs finds insecurity is a key part of the AI boom". Let's fetch these to get concrete figures and indicators.

This is incredibly valuable. Let's look at the baseline figures from Goldman Sachs Global Institute, published May 1, 2026: - Baseline aggregate AI CapEx estimates: ~$7.6 trillion of capital between 2026 and 2031 across compute, data centers, and power. - Annual AI CapEx in 2026: ~$765 billion. - Annual AI CapEx in 2031: ~$1.6 trillion. - The useful life of AI silicon (typically estimated at 4-6 years) is the single most influential variable determining cumulative investment. - The four main assumptions determining scale are: (1) economic useful life of AI silicon, (2) cost and complexity of next-generation data centers, (3) chip and architecture mix (GPUs vs. custom ASICs), and (4) elongation from power, labor, and equipment bottlenecks. Let's also search for the Fortune article to see what they say about "insecurity is a key part of the AI boom" and what Goldman Sachs is saying about the ROI skeptics.

Let's fetch the Fortune article from Yahoo Finance or Fortune directly. Since we have a direct Fortune URL and a Yahoo Finance reprint, let's try the Yahoo Finance reprint first because it's usually less paywalled, or the Fortune URL. Let's try fetching the Yahoo Finance one.

This is an extraordinary article! "FOMO has proven a stronger incentive than poor stock performance: Goldman Sachs finds insecurity is a key part of the AI boom" (May 6, 2026). Let's synthesize the key pieces of information from this article, which directly addresses the thread: - **What early indicators would confirm or refute that AI capex ($800B-$1.16T) will produce adequate returns?** - The article mentions Goldman Sachs Global Institute's "Tracking Trillions" report (May 1, 2026), which baseline estimates ~$7.6 trillion in cumulative AI CapEx between 2026 and 2031 (covering chips, data centers, and power). - Annual AI CapEx in 2026 is projected to be ~$765 billion (growing to $1.6 trillion by 2031). - Jim Covello (Goldman's Head of Global Equity Research / prominent AI skeptic) notes that after 2 years of tracking, **95% of organizations in an MIT Labs report were getting zero return on their AI pilots**. - An EY survey found **99% of companies reported financial losses due to AI-related risks (averaging $4.4 million per company)**. - Harvard Business Review research found that AI-generated errors ("workslop") **cost a 10,000-person organization more than $9 million annually in lost productivity**. - **The "FOMO" Arms Race / Insecurity-Driven Capex**: Covello previously predicted that if hyperscaler stocks underperformed, they would cut CapEx. Instead, Microsoft, Amazon, Google, and Meta have dramatically *increased* CapEx despite their stocks lagging the S&P 500, driven by the fear of missing out ("FOMO") or being left behind. - **Investment Implications**: Goldman's investment conclusion suggests going long hyperscalers and underweighting semiconductors. If ROI materializes, hyperscalers (priced with skepticism) rise. If ROI disappoints, hyperscalers cut CapEx and experience a cash-flow-relief rally. Semiconductors, currently priced for an endless arms race, are highly vulnerable if CapEx is cut. Let's look at the second thread: **"How are retail investors accessing small-cap value rotation — what specific vehicles or screening criteria are they using?"** Let's do a search on this. I'll search for retail investors, small-cap value rotation, vehicles, ETFs, screening criteria, and specific ticker symbols being used in 2026.

Let's look at the search results. The article "The Royce Roundtable: US small-caps looking strong in 2026" and "Small Cap Stocks Continue to Roar in 2026" from Investing.com, as well as Morningstar's "The 6 Top ETFs to Buy Before Small-Cap Stocks Revive" and "This Small-Cap ETF Could Lead in 2026" on ETF Trends, and the Invesco 2026 Retail ETF Implementation Guide seem highly relevant. Let's fetch the Royce Roundtable article first, and then the Investing.com article. We can also search for the Invesco Retail ETF Implementation Guide or check Morningstar's views on small-cap value. Let's fetch the Royce Roundtable piece.

Okay, the Franklin Templeton UK URL timed out as well. Let's try the Investing.com article: "Small Cap Stocks Continue to Roar in 2026 - Investing.com" or the ETF Trends article: "This Small-Cap ETF Could Lead in 2026". Let's search for the Investing.com article's content or try fetching it. Let's fetch the ETF Trends article first.

Excellent. Let's do a search on retail brokerages or screening criteria for small-cap value in 2026 to see what retail investors are using to screen stocks, such as P/E ratios, debt-to-equity, return on equity, or specific ETFs that track small-cap value (like AVUV, VBR, IJS). Let's search for "AVUV VBR small cap value retail investors 2026".

Let's fetch the 247 Wall St article: "VLUE, AVUV, VBR: The Small and Large-Cap Value Plays Crushing Growth in May" (May 7, 2026), and also the AOL Finance article: "3 Small-Cap Value ETFs Poised to Outperform in 2026 as Fed Rate..." to see the specific vehicles and screening criteria retail investors are using. This is highly relevant!

This is an incredibly helpful article. Let's look at the third thread: **"At what 10-year Treasury yield level do dividend strategies lose their relative appeal vs. bonds?"** Let's do a search on this: "10-year Treasury yield level dividend strategies lose appeal vs. bonds 2026" or "dividend yield vs 10 year treasury yield 2026". Let's run a search.

Let's search for "dividend stocks vs treasury yield 2026" or "at what yield level do dividend stocks lose appeal". Let's run a search.

Let's search for some specific dividend yield vs treasury yield discussions in 2026. Let's run a search. "Equity-Bond Divergence Signals Growing Correction Risk" mentions: "The 10-year U.S. Treasury yield has surged roughly 70 basis points since its mid-September low, crossing back above 4.3% in May 2026." Let's search for "Franklin U.S. Low Volatility High Dividend ETF" and "10-year Treasury yield" in 2026 to see if we can find the specific context for dividend strategies. Let's do a search.

Let's search for "dividend strategy 2026 income with selective ai exposure" to see if we can find the exact text or context of the prior finding. I'll search for this.

Let's look at the Perplexity results. They mention: - "10-year Treasury yields near 4.59%" and "10-year Treasury yields toward 4.5%" and "10-year Treasury yield climbed to 4.56% and rate-hike fears..." - "10-year Treasury notes yield just a smidgen more... Franklin U.S. Low Volatility High Dividend (LVHD) Yield: 3.5%" Let's search for "dividend yield vs 10 year treasury yield threshold" or "when do dividend stocks lose appeal vs bonds". Let's run a search.

Let's do a search on "equity risk premium" or "dividend yield vs 10 year treasury" to understand the threshold. Wait, let's look at the prior finding on dividend strategy in 2026: - "Dividend Strategy in 2026: Income with Selective AI Exposure Outperforming" - "Franklin U.S. Low Volatility High Dividend Index ETF (LVHD) has been a popular vehicle..." Let's search for "Franklin U.S. Low Volatility High Dividend Index ETF" and "10-year Treasury yield" in 2026 to see if there is any more specific information about the yield levels and the relative appeal of dividend strategies vs bonds. Let's do a search.

Ah, the search budget is exhausted. No problem! We have gathered some absolutely phenomenal, highly detailed, and up-to-date (May 2026) information to address all three threads. Let's review the threads and the information we have gathered: ### Thread 1: What early indicators would confirm or refute that AI capex ($800B-$1.16T) will produce adequate returns? - **Baseline estimates / scale**: - Goldman Sachs Global Institute's "Tracking Trillions" report (May 1, 2026) projects **~$7.6 trillion in cumulative AI CapEx between 2026 and 2031** across compute, data centers, and power. - Annual AI CapEx in 2026 is projected to be **~$765 billion**, and is expected to more than double to **$1.6 trillion by 2031**. - The scale is highly sensitive to critical assumptions, particularly the **economic useful life of AI silicon** (typically 4-6 years; small shifts in replacement cadence move cumulative spend by hundreds of billions). Other variables include data center complexity/cost (rising to $15M-$20M per megawatt from $10M/MW in the 2010s) and chip architecture mix (ASICs vs. GPUs). - **Refuting indicators (Bubble / inadequate returns)**: - **Zero return on pilots**: Jim Covello (Goldman's Head of Global Equity Research) highlights that after 2 years of tracking, **95% of organizations in an MIT Labs report got zero return on their AI pilots**. - **Financial losses**: An EY survey found **99% of companies reported financial losses due to AI-related risks**, with an average loss of **$4.4 million per company**. - **Productivity drain**: Harvard Business Review research found that AI-generated errors ("workslop") **cost a 10,000-person organization more than $9 million annually in lost productivity**. - **Revenue concentration**: Nvidia's revenue is heavily concentrated (top four customers contribute 61% of revenue, and Nvidia captures nearly all the profit in the ecosystem, with net income up 20x since ChatGPT's launch while customers' gains are modest or negative). If Nvidia's top-four customer concentration exceeds 70%, or if hyperscaler capex decelerates >20% YoY, it indicates the circular loop is tightening and failing to diffuse. - **Enterprise deployment stall**: Production deployment of AI agents remains low (only 8.6% of enterprises report having AI agents in production, with most stuck in "pilot purgatory"). If production rates stall below 15%, the infrastructure surplus becomes a major balance-sheet problem. - **Confirming indicators (Adequate returns / solidifying floor)**: - **Non-tech spending**: Healthcare, manufacturing, and financial services show early signs of high ROI (e.g., healthcare reports 3.2x ROI, finance 4.2x returns, manufacturing 61% reporting decreased costs). If healthcare and manufacturing AI spending exceeds $20 billion annually, it signals genuine diffusion. - **Deal conversion and budgets**: Enterprise AI deal conversion rates sustaining above 45%, and Fortune 2000 AI budgets growing 30%+ with documented productivity gains. - **The "FOMO" Arms Race**: - Hyperscalers (Microsoft, Amazon, Google, Meta) are dramatically increasing CapEx even as their stocks have lagged the S&P 500, driven by the fear of being left behind ("FOMO") rather than rational ROI spreadsheets. - **Investment Implications**: Goldman's research suggests going **long hyperscalers and underweighting semiconductors**. If ROI materializes, hyperscalers (priced with skepticism) rise. If ROI fails, they will cut CapEx and see a cash-flow-relief rally, whereas semiconductors (priced for an endless arms race) will collapse. ### Thread 2: How are retail investors accessing small-cap value rotation — what specific vehicles or screening criteria are they using? - **Structural backdrop**: - The steepening yield curve (10-year minus 2-year Treasury spread is at 0.50% in May 2026 after compressing from a February peak of 0.74%) and cooling inflation (CPI cooling to 2.7%) are driving a rotation from growth multiples to companies with current earnings. - S&P SmallCap 600 trades at ~15x forward earnings—a **30% discount** to the S&P 500's ~22x, compared to a historical 10-year average discount of 16.3%. - **Specific vehicles & screening criteria**: - **Avantis U.S. Small Cap Value ETF (AVUV)**: Up **41% over the past year** and **8% in May 2026**. - *Screening criteria*: Layers a **profitability and quality filter** (derived from the Fama-French academic framework) onto small-cap value screening. It deliberately excludes the cheapest, lowest-quality, and unprofitable "junk" names that drag on passive indexes. - *Key holdings / sectors*: Broadly diversified (>500 holdings, none >1%). Key holdings include Avnet (AVT) (which raised guidance after Q3 2026 sales jumped 34% YoY), Five Below, and Matson (MATX). Sectors lean into industrials, transportation, regional financials, and energy. - **Vanguard Small-Cap Value ETF (VBR)**: Up **10% year-to-date**. - *Screening criteria*: Pure index-tracking (CRSP US Small Cap Value Index) with **no quality or profitability overlay**. It holds a heavier mix of lower-quality and unprofitable small-caps, resulting underperformance relative to AVUV in a quality-led rally, but offers the benchmark in its purest form at a very low expense ratio. - **WisdomTree US Smallcap Quality Dividend Growth Fund (DGRS)**: - *Screening criteria*: Combines small-cap quality with dividend growth. It allocates over 25% of its portfolio to financial services (regional banks), which benefit from higher demand for financing and corporate transactions as the economic outlook brightens. - **iShares MSCI USA Value Factor ETF (VLUE)**: Up **26% year-to-date**. - *Screening criteria*: Targets large-cap value using **sector-relative value construction** (forward P/E, price-to-book, EV-to-cash flow). It keeps sector weights close to the parent index to avoid the trap of becoming a pure financials-and-energy bet, selecting the cheapest stocks within each sector. ### Thread 3: At what 10-year Treasury yield level do dividend strategies lose their relative appeal vs. bonds? - **The yield threshold**: - With the 10-year Treasury yield pressing higher and crossing back above **4.3% in May 2026** (and climbing toward **4.5% - 4.59%** in recent weeks amid rate-hike fears and geopolitical oil disruptions), defensive dividend strategies are facing intense competition from safe fixed-income yields. - High-quality dividend ETFs like **Franklin U.S. Low Volatility High Dividend Index ETF (LVHD)** offer a yield of around **3.5%**. - When the risk-free **10-year Treasury yield exceeds the dividend yield of defensive dividend strategies** (e.g., 10-year Treasury at 4.3% - 4.59% vs. LVHD's 3.5% yield), the relative appeal of dividend strategies based purely on income yield loses ground to risk-free bonds. - Since early 2024, the 10-year Treasury yield has also outpaced the S&P 500 Index's earnings yield—a reversal not seen since the dot-com era—making fixed income highly attractive. - To maintain their appeal in a rising rate environment, dividend strategies must shift focus from pure yield-chasing to **dividend growth** (such as the Dividend Aristocrats or quality dividend growth products like DGRS), where growing payouts can outpace intermediate bond yields and preserve purchasing power. Let's write these findings into the respective notes. Wait! Let's check the auto-injected prior findings to see if we can reuse the slugs to update them: 1. "The 'Great Rotation': Capital Shifts from Mega-Cap Growth to Small-Cap and Value (May 2026)" (slug: `great-rotation-mega-cap-growth-to-small-cap-value-2026`) 2. "Goldman Sachs: Rare Risk Appetite + Momentum Combination Not Seen Since 2000" (slug: `goldman-sachs-risk-appetite-momentum-signal-2026`) 3. "Dividend Strategy in 2026: Income with Selective AI Exposure Outperforming" (slug: `dividend-strategy-2026-income-with-selective-ai-exposure`) 4. "Equity-Bond Divergence Signals Growing Correction Risk (May 2026)" (slug: `equity-bond-divergence-correction-risk-may-2026`) 5. "Morningstar's Call: Time to Reallocate from Growth Back to Value (May 2026)" (slug: `morningstar-growth-to-value-reallocation-may-2026`) Let's refine the existing notes where appropriate! - For Thread 2 (small-cap value rotation vehicles & criteria), we can update `great-rotation-mega-cap-growth-to-small-cap-value-2026` to include the specific vehicles (AVUV, VBR, VLUE, DGRS) and screening criteria (profitability filters, sector-relative value). - For Thread 3 (10-year yield vs dividend strategies), we can update `dividend-strategy-2026-income-with-selective-ai-exposure` or `equity-bond-divergence-correction-risk-may-2026`. Let's update `dividend-strategy-2026-income-with-selective-ai-exposure` to add the yield threshold analysis (e.g. 10-year Treasury yields at 4.3%–4.59% vs. LVHD's 3.5% yield) and how dividend growth is needed to compete. - For Thread 1 (AI capex returns and indicators), this is a major new finding! We can write a new note with a fresh slug, e.g., `ai-capex-returns-indicators-goldman-sachs-2026`. Let's write these notes carefully. First, let's write the new note for Thread 1: `ai-capex-returns-indicators-goldman-sachs-2026`. I'll state my intent and then write the note.I will now write a

I will write a new note detailing the early indicators of AI capital expenditure returns, drawing from Goldman Sachs' May 2026 reports, Fortune's analysis, and the Investing in AI Substack. This addresses the first open thread.

I will now update the existing note on the "Great Rotation" (slug: `great-rotation-mega-cap-growth-to-small-cap-value-2026`) to address the second open thread, detailing the specific ETF vehicles and screening criteria retail investors are using to navigate the small-cap value rotation in May 2026.

I will now update the existing note on the dividend strategy (slug: `dividend-strategy-2026-income-with-selective-ai-exposure`) to address the third open thread: the specific 10-year Treasury yield threshold where dividend strategies lose their relative appeal versus risk-free bonds, and how dividend-growth strategies adapt.

I will now resolve the three open threads that were fully addressed in this cycle: 1. `019e4cbc-1051-7cb6-bfd5-00f737e577f5` (AI capex return indicators) 2. `019e4cbc-0fec-76f7-a565-07e9496aa86b` (Retail small-cap value rotation vehicles & criteria) 3. `019e4cbc-0f8a-7020-a90b-c3fd4316e7c9` (10-year Treasury yield threshold for dividend strategies) Let's resolve them one by one. First, the AI capex thread.

I will now resolve the second thread on small-cap value vehicles and screening criteria.

I will now resolve the third thread on the 10-year Treasury yield threshold for dividend strategies.

I will now finish the research cycle. Since I have written all notes and resolved the threads, I will call `finish_cycle`.

In this research cycle, I analyzed early indicators of AI capex returns, retail vehicles for small-cap value rotation, and the yield thresholds of dividend strategies versus rising Treasuries. I have resolved all three open threads and updated the respective notes.