Investor Summary
Fund Strategy
FUND PERFORMANCE AS OF 30th June 2026
| ANNUALIZED SINCE INCEPTION | QUARTERLY | YTD |
|---|---|---|
| - | 1.9% | - |
| ANNUALIZED SINCE INCEPTION | QUARTERLY | YTD |
|---|---|---|
| - | 1.9% | - |
Bristlemoon Global Fund returned 1.9% for Q2 2026, bringing the fund to its two-year mark. The past year has been challenging, with significant drawdowns driven by position sizing errors in stocks like Hemnet and PAR Technology where the fund averaged down without near-term catalysts. In response, the fund has implemented a compounding/conviction framework that decomposes positions into compounding and conviction exposures, concentrating capital only in stocks with positively inflecting narratives and accelerating growth profiles. This framework acknowledges permanent market structure changes driven by multi-manager firms and passive flows that create greater volatility and more extreme multiple oscillations. Early evidence suggests the framework is working, as demonstrated by the Timee position which was scaled from 5% to 8% following a growth inflection, subsequently rallying 40%. The fund's highest conviction remains in semiconductor capital equipment, particularly ASML, which is positioned to benefit from a multi-year WFE supercycle driven by unprecedented AI-related capex expansion. TSMC, Intel, Samsung, SK Hynix, and Micron are all significantly expanding capacity, with Korean memory giants announcing 800 trillion won in incremental investments. The fund's EUV and immersion DUV unit forecasts sit materially above consensus, and ASML remains a core position despite strong recent performance.
The fund invests in exceptional businesses with strong compounding potential, but has adapted its approach to acknowledge that market structure changes have created greater stock price volatility and more extreme multiple oscillations. The fund now employs a compounding/conviction framework that concentrates capital in stocks with positively inflecting narratives and accelerating growth profiles, not valuation cheapness alone, while being more tactical in timing position additions to exploit the multiple de-rate/re-rate cycle. The fund's largest conviction is in semiconductor capital equipment, particularly ASML, which sits at a critical bottleneck in the supply chain and is positioned to benefit from a multi-year WFE supercycle driven by unprecedented AI-related capex expansion across leading edge logic and memory.
The fund expects a WFE supercycle over the next several years driven by unprecedented leading edge logic and DRAM capacity expansion. ASML is positioned to benefit from this cycle as EUV tools become increasingly important across both leading-edge logic and DRAM, with the fund's unit forecasts sitting materially above consensus for both EUV and immersion DUV. The fund continues to see healthy upside for ASML and believes the stock is one of the cleanest ways to play the next leg of AI infrastructure buildout. The fund has implemented process improvements through its compounding/conviction framework to be more tactical in timing position additions and to concentrate capital in stocks with accelerating growth profiles, with early evidence that these changes are working as demonstrated by the Timee position.
| Date | Letter | Tickers | Keywords | Pitches | Quick Takes |
|---|---|---|---|---|---|
| Jul 14 2026 | 2026 Q2 | 000660 KS, 005930 KS, 8035.T, AAPL, AMAT, AMD, APP, ASML, GOOGL, INTC, KLAC, LRCX, MU, NOW, NVDA, PAR, TSLA, TSM | AI, market structure, Memory, Portfolio Management, Position Sizing, risk management, semiconductors |
APP ASML TIMEE.T |
Bristlemoon returned 1.9% in Q2 2026 after a difficult year that prompted implementation of a compounding/conviction framework to prevent over-sizing positions lacking catalysts. The fund concentrates capital in stocks with positively inflecting narratives, as demonstrated by scaling Timee from 5% to 8% following growth reacceleration. Highest conviction remains ASML, positioned to benefit from a multi-year semiconductor capex supercycle driven by unprecedented AI infrastructure buildout across TSMC, Intel, and memory manufacturers. |
| Apr 15 2026 | 2026 Q1 | APP, CSU.TO, DG, FICO, FND, GOOGL, GRDR, NOW, PAR, WDS.AX | Advertising, AI, E-Commerce, gaming, growth, software, technology, volatility | - | Bristlemoon Global Fund suffered a -25.5% drawdown in Q1 2026 from AI disruption fears hitting software holdings, despite stable fundamentals. The narrative-driven selloff created compelling valuations in quality businesses like AppLovin and FICO. Management refined position sizing to manage volatility while maintaining conviction in long-term earnings power of concentrated portfolio. |
| Jan 9 2026 | 2025 Q4 | ADBE, APP, ASML, EYE, FICO, FND, GOOGL, HD, LOW, NKE, PRTG, SBUX, SNPS, UBER | AI, growth, Housing, Restaurants, semiconductors, software, technology, value |
ASML NA PAR ADBE BCG LN 215A JP FND |
Bristlemoon delivered 11.9% since inception despite Q4 weakness, benefiting from AI winners AppLovin and ASML while suffering from Hemnet and PAR declines. The manager initiated positions in Adobe, Floor & Decor, and others during market pessimism. PAR approaches potential McDonald's partnership while Floor & Decor awaits housing recovery. Portfolio positioned for opportunity-rich environment ahead. |
| Oct 10 2025 | 2025 Q3 | AAPL, ADI, AER, APP, ASML, BABA, CDNS, CRM, FICO, GOOGL, HEMN.ST, INTC, META, MSFT, NVDA, PAR, SNPS, TSM, UBER, V | AI, global, Long/Short, Quality, semiconductors, software, technology |
ASML NA GOOGL US SNPS US PAR UN |
Bristlemoon returned 5.0% in Q3 2025, adding positions in ASML, Alphabet, and Synopsys during market weakness while maintaining conviction in volatile holding PAR Technology. The fund targets high-quality businesses with competitive moats and mispriced earnings power, recently capitalizing on AI disruption fears and semiconductor equipment concerns to establish attractive entry points in monopolistic and oligopolistic market leaders. |
| Jul 14 2025 | 2025 Q2 | AER, APG, APP, CRM, PAR, UBER, UNH | AI, Concentration, healthcare, tariffs, technology, value |
HEM APG CRM UNH |
Bristlemoon underperformed due to defensive tariff positioning and concentrated portfolio volatility, but management expresses highest conviction ever in current holdings. New positions in APi Group, Salesforce, and UnitedHealth offer compelling risk-adjusted returns. Fund continues gradual deployment into idiosyncratic opportunities with 12-18 month catalysts while avoiding momentum chasing at market highs. |
| Apr 9 2025 | 2025 Q1 | 0700.HK, APP, HEM.ST, IBKR, META, PAR, PDD, PINS, XPOF | Cash Management, China, Restaurants, risk management, software, tariffs, technology, volatility |
HEM.ST PAR |
Bristlemoon delivered -3.2% in March 2025, hurt by Xponential Fitness exit but helped by China holdings. Fund proactively raised cash to 22% and hedged before Trump tariffs, then upgraded portfolio quality by exiting rate-sensitive and ad-dependent names. Largest positions Hemnet and PAR Technology offer significant pricing power upside. Defensive positioning preserves optionality for emerging opportunities. |
| Dec 1 2024 | 2024 Q4 | APP, META, MTCH, PAR, PDD, PINS, TCEHY | Advertising, gaming, Portfolio Management, Position Sizing, risk management, Social Media, technology | - | Bristlemoon delivered 10.4% in Q4 despite position sizing errors in top holdings. Pinterest remains largest detractor with revenue growth decelerating to 15-17% from 20%+. AppLovin drove performance with 148% gain as earnings forecasts increased 5.3x. Fund implements revised position sizing framework requiring exceptional risk/reward and conviction for large positions. Concentrated approach continues with 90.2% long exposure across 16 names. |
| Jan 10 2025 | 2024 Q4 | APP, META, MTCH, PAR, PDD, PINS, TCEHY | Advertising, gaming, Portfolio Management, Position Sizing, risk management, Social Media, technology | - | Bristlemoon delivered 10.4% in Q4 despite position sizing errors in top holdings. Pinterest remains largest detractor with revenue growth decelerating to 15-17% from 20%+. AppLovin drove performance with 148% gain as earnings forecasts increased 5.3x. Fund implements revised position sizing framework requiring exceptional risk/reward and conviction for large positions. Concentrated approach continues with 90.2% long exposure across 16 names. |
| Oct 10 2024 | 2024 Q3 | 0700.HK, 9988.HK, APP, EYE, META, PAR, PDD | Advertising, AI, China, Concentration, gaming, growth, Stimulus, technology | APP | Bristlemoon delivered 6.8% in Q3 through concentrated positions in Chinese internet stocks and AppLovin. Chinese stimulus drove significant gains in undervalued Tencent, Alibaba, and Pinduoduo holdings. AppLovin's AI-powered ad platform showed sustained 75% software growth. The fund maintains high conviction in quality businesses trading at attractive valuations with multi-year earnings compounding potential. |
| QUARTER | THEMES | TAGS |
|---|---|---|
| 2026 Q2 |
SemiconductorsThe fund views semiconductor capital equipment as the most attractive subsector within semiconductors, driven by unprecedented demand for compute and memory in a higher-for-longer wafer fab equipment capex cycle. ASML is the preferred exposure, with the stock nearly doubling since September 2025 despite material derating against peers. The fund expects a WFE supercycle over the next several years driven by leading edge logic and DRAM capacity expansion, with TSMC raising capex guidance and Intel Foundry experiencing a revival. |
ASML EUV WFE Foundries Memory |
AIAI infrastructure buildout is driving massive capex expansion across the semiconductor supply chain. The rise of agentic AI is expected to drive rapid growth in the CPU market, with AMD forecasting the 2030 CPU TAM at $120 billion (doubled from six months prior) and Nvidia projecting $200 billion. ASML is positioned as one of the cleanest ways to play the next leg of AI infrastructure buildout, sitting at a critical bottleneck in the semiconductor supply chain. |
AI Infrastructure Agentic AI Data Centers Compute | |
Position SizingThe fund has implemented a compounding/conviction framework that decomposes positions into two exposures to prevent over-sizing positions lacking near-term catalysts or with negatively inflecting narratives. The framework seeks to concentrate capital in stocks with accelerating growth profiles that result in earnings beats, acknowledging that market structure changes mean stocks trade differently than historically. The fund will only average down and maintain large positions when there is a well-founded view of positively inflecting narratives, not on valuation cheapness alone. |
Risk Management Portfolio Management Conviction Catalysts | |
Market StructureThe fund observes significant market structure changes driven by the rise of multi-manager firms, retail investors, systemic strategies, and passive flows. These changes result in greater stock price volatility, with narrative shifts occurring more suddenly and stock multiples oscillating more extremely. The fund believes these changes are permanent and has adapted its approach to exploit the multiple de-rate/re-rate cycle, being more tactical in timing position additions and willing to bank gains when stocks rally sharply on narrative shifts. |
Volatility Multi-Manager Market Regime | |
MemoryThe fund sees a memory capex supercycle driven by unprecedented DRAM capacity expansion. Samsung and SK Hynix announced massive investment plans through 2040, with 800 trillion won of incremental capex for a new semiconductor cluster in southwest Korea. Both companies have pulled forward fab completion timelines, with Samsung's Pyeongtaek P5 Fab 1 and SK Hynix's Yongin Fab 1 targeting wafer production in 2027/28. Micron guided to north of $45 billion capex for FY27, exceeding cumulative capex from FY22-FY25, and pulled forward its Idaho fab timeline. |
DRAM HBM Samsung SK Hynix Micron | |
| 2026 Q1 |
AIThe fund experienced significant drawdowns due to AI disruption fears affecting software and digital businesses. The market bifurcated into perceived AI winners and losers, with the fund's holdings suffering from narrative-driven selloffs despite stable earnings. AppLovin is positioned as an AI winner through automated video ad generation and improved targeting algorithms. |
Disruption Software Algorithms Automation Targeting |
SaaSSoftware-as-a-service companies experienced brutal selloffs in Q1 2026, with the IGV software ETF declining 24%. The fund's asset-light, high-growth software holdings were particularly impacted by the SaaSpocalypse despite strong fundamentals and earnings growth prospects. |
Software Growth Multiples Earnings Volatility | |
GamingAppLovin's mobile gaming advertising business continues robust growth with 66% year-over-year revenue growth in Q4 2025. The company benefits from increased game creation through AI tools like Google's Project Genie, which should expand in-game advertising inventory and strengthen AppLovin's distribution advantages. |
Mobile Advertising Distribution Monetization Publishers | |
E-commerceAppLovin is rolling out e-commerce advertising capabilities with promising early results. The self-service platform automates advertiser onboarding, potentially creating explosive growth in smaller merchants. E-commerce ads can monetize previously wasted gaming ad impressions at 100% effective take rates. |
Advertising Merchants Automation Monetization Platform | |
| 2025 Q4 |
AIManager sees AI as creating both winners and losers, with Adobe positioned as an AI beneficiary despite market pessimism. PAR Technology leverages AI through Coach AI for restaurant management. The fund benefited from AI winners like AppLovin while avoiding perceived AI losers that proved resilient. |
Artificial Intelligence Machine Learning Generative AI Coach AI AI Tools |
SemiconductorsASML highlighted as a monopoly in the semiconductor industry during an AI boom. Manager took positions in Alphabet and ASML, with ASML stock rising from €600 to over €900 per share in a quarter as the market agreed with their investment thesis. |
ASML Semiconductor Equipment AI Boom Monopoly | |
RestaurantsPAR Technology positioned as the future operating system for global enterprise restaurants, potentially signing with McDonald's. The fragmented nature of legacy restaurant IT creates opportunities for unified platforms that enable AI deployment. |
PAR Technology Point of Sale Restaurant Technology McDonald's Unified Platform | |
Real EstateFloor & Decor positioned to benefit from housing market recovery. Manager expects existing home sales to recover from current depressed levels as mortgage rates normalize and the lock-in effect from low-rate mortgages gradually fades. |
Floor & Decor Housing Market Mortgage Rates Home Sales Flooring | |
SoftwareAdobe viewed as beneficiary of AI through workflow orchestration despite market fears of displacement. The manager believes AI will increase demand for editing and management tools rather than replace them, with Adobe capturing value in both creation and composition. |
Adobe Creative Software Workflow SaaS Digital Content | |
| 2025 Q3 |
AIThe fund views AI as a transformative force across multiple holdings. Google has shipped market-leading AI products including Gemini models and AI Mode, while benefiting from cost advantages through its TPU hardware. The AI infrastructure boom drives demand for ASML's lithography equipment and creates opportunities for Synopsys in custom ASIC design tools. |
Artificial Intelligence Machine Learning AI Infrastructure Custom ASICs AI Models |
SemiconductorsASML represents a monopoly position in critical lithography equipment for advanced chip manufacturing. The fund sees temporary weakness in orders as creating attractive entry points, while long-term demand remains strong driven by AI infrastructure needs and memory requirements. EUV and DUV tools remain essential for leading-edge production. |
Lithography EUV DUV Chip Manufacturing Wafer Fabrication | |
Enterprise SoftwarePAR Technology represents the fund's bet on restaurant technology transformation, despite recent volatility. The company offers unified POS and software solutions with potential for mega-deal wins including McDonald's. Synopsys provides mission-critical EDA tools for chip design in a stable oligopoly market structure. |
SaaS Point of Sale Restaurant Technology EDA Tools Software Platforms | |
QualityThe fund focuses on high-quality businesses with competitive advantages, seeking both obvious quality at reasonable prices and underappreciated quality companies. They look for businesses with predictable cash flows, high reinvestment returns, and mispriced earnings power, maintaining flexibility to pivot where opportunities exist. |
Business Quality Competitive Moats Predictable Cash Flows High Returns Mispriced Assets | |
| 2025 Q2 |
AISalesforce is positioned to capitalize on agentic AI through its data advantage and distribution via the app layer. Agentforce reached $100 million ARR with 5x quarter-over-quarter growth. The agentic AI market could reach $155 billion by 2030 according to Bank of America estimates. |
Agentforce Automation Data Cloud Enterprise |
Trade PolicyThe fund was defensively positioned against tariff risks but was too slow to redeploy when Trump announced a 90-day pause on reciprocal tariffs. Markets quickly wrote off tail risk as the TACO trade emerged - Trump Always Chickens Out. |
Tariffs Trump Defensive Policy | |
Managed CareUnitedHealth faces challenges from elevated Medicare Advantage utilization and V28 risk adjustment model implementation. The stock offers asymmetric risk/reward after a 50% drawdown, with potential recovery to $290-$375 range. |
Medicare Utilization Value-based Healthcare | |
| 2025 Q1 |
Risk ManagementFund increased cash weighting to 22% during March due to heightened market uncertainty and Trump's tariff policies. Adopted maximum loss framework for position sizing and reduced concentration to manage downside risk during volatile periods. |
Cash Volatility Hedging Concentration Uncertainty |
ChinaPDD Holdings and Tencent Holdings were key contributors to performance as top five positions. Both China-based companies provided positive returns during the quarter despite broader market challenges. |
PDD Tencent Technology E-commerce Internet | |
TechnologyPortfolio includes significant technology exposure through companies like Meta, AppLovin, PAR Technology, and Interactive Brokers. Focus on mission-critical software platforms and digital advertising businesses with strong competitive moats. |
Software SaaS Digital Platforms Enterprise | |
RestaurantsPAR Technology represents major position as provider of mission-critical POS software to enterprise restaurant chains including Burger King, Arby's, and Wendy's. Company has enormous latent pricing power and resilient customer base. |
POS QSR Enterprise Software Foodservice | |
| 2024 Q4 |
Social MediaThe fund holds significant positions in social media platforms including Pinterest and Meta Platforms. Pinterest represents the fund's largest performance detractor since inception, with management acknowledging position sizing errors. The company is transitioning to lower-funnel ad products but has experienced revenue growth deceleration from 20%+ to 15-17% guidance. |
Pinterest Meta Advertising Revenue Growth |
GamingAppLovin Corporation was a key contributor to quarterly performance with 148% stock price increase. The company exemplifies uncertainty creating opportunity, with analyst earnings forecasts increasing 5.3x from $1.14 to $6 per share for FY25 in just two years, driving the stock up more than 5x since initial investment. |
AppLovin Earnings Forecasts Growth Mobile | |
AdvertisingDigital advertising is a core focus with positions in Pinterest and Meta Platforms. Pinterest's transition from brand-awareness to performance advertising has been slower than expected, with management unable to explain revenue growth deceleration despite product improvements and advertiser adoption initiatives. |
Digital Performance Revenue Growth Platforms | |
| 2024 Q4 |
Social MediaThe fund holds significant positions in social media platforms including Pinterest and Meta Platforms. Pinterest represents the fund's largest performance detractor since inception, with management acknowledging position sizing errors. The company is transitioning to lower-funnel ad products but has experienced revenue growth deceleration from 20%+ to 15-17% guidance. |
Pinterest Meta Advertising Revenue Growth |
GamingAppLovin Corporation was a key contributor to quarterly performance with 148% stock price increase. The company exemplifies uncertainty creating opportunity, with analyst earnings forecasts increasing 5.3x from $1.14 to $6 per share for FY25 in just two years, driving the stock up more than 5x since initial investment. |
AppLovin Earnings Forecasts Growth Mobile | |
AdvertisingDigital advertising is a core focus with positions in Pinterest and Meta Platforms. Pinterest's transition from brand-awareness to performance advertising has been slower than expected, with management unable to explain revenue growth deceleration despite product improvements and advertiser adoption initiatives. |
Digital Performance Revenue Growth Platforms | |
| 2024 Q3 |
AIAppLovin's AXON 2.0 AI-powered ad targeting algorithm has driven exceptional growth, with software revenues growing 75% year-over-year. The neural network utilizes recent AI breakthroughs to improve ad targeting beyond previous capabilities. AXON 2.0's targeting improvements allow AppLovin to deliver superior return on ad spend outcomes at scale. |
Machine Learning Ad Tech Targeting Neural Networks Performance |
ChinaChinese internet companies Tencent, Alibaba, and Pinduoduo were purchased as exceptionally cheap, high-quality businesses. Chinese government stimulus measures in September 2024 provided significant multiple expansion. The fund has since trimmed China positions after the rally but maintains exposure to these quality franchises. |
Stimulus Internet Valuation Government Policy Multiple Expansion | |
GamingAppLovin operates in mobile gaming advertising and owns a portfolio of over 100 mobile games. The company's ad network connects advertisers with in-game ad slots, utilizing AI to show the right ad to the right person at the right time. Mobile gaming represents a core vertical for AppLovin's advertising technology platform. |
Mobile Gaming Ad Network Monetization User Acquisition In-App Advertising | |
AdvertisingAppLovin's advertising technology platform reaches 1.4 billion users daily through its AppDiscovery and MAX mediation products. The company benefits from network effects and switching costs in the mediation business. App install advertising is performance-driven with uncapped budgets, creating a favorable market dynamic. |
Ad Tech Mediation Performance Marketing Network Effects Real-Time Bidding |
| Date | Pitch Type | Author | Ticker | Company | Industry | Sub Industry | Bull / Bear | Exchange | Keywords | Action |
|---|---|---|---|---|---|---|---|---|---|---|
| Jul 14, 2026 | Fund Letters | The Bristlemoon Global Fund | APP | AppLovin Corporation | Advertising Agencies | Interactive Media & Services | Bull | NASDAQ | Ad Tech, compounder, earnings acceleration, growth, Mobile Advertising, multiple expansion, software platform, US | Login |
| Jul 14, 2026 | Fund Letters | The Bristlemoon Global Fund | ASML | ASML Holding NV | Semiconductor Equipment & Materials | Semiconductor Equipment | Bull | NASDAQ | AI infrastructure, Capital equipment cycle, DRAM, EUV lithography, Logic semiconductors, Monopoly, Netherlands, Pricing power, semiconductor equipment, WFE supercycle | Login |
| Jul 14, 2026 | Fund Letters | The Bristlemoon Global Fund | TIMEE.T | Timee Inc | Other | Human Resource & Employment Services | Bull | Japan Exchange Group (Tokyo Stock Exchange) | Food service, Gig Economy, Growth Reacceleration, Japan, Labor platform, marketplace, Monopoly, operating leverage | Login |
| Jan 9, 2026 | Fund Letters | George Hadjia | ASML NA | ASML Holding N.V. | Information Technology | Semiconductor Equipment | Bull | Euronext Stock Exchange | AI, CapEx, Euv, Monopoly, semiconductors | Login |
| Jan 9, 2026 | Fund Letters | George Hadjia | PAR | PAR Technology Corporation | Information Technology | Application Software | Bull | New York Stock Exchange | AI, Catalysts, platform, POS, Restauranttech | Login |
| Jan 9, 2026 | Fund Letters | George Hadjia | ADBE | Adobe Inc. | Information Technology | Application Software | Bull | NASDAQ | AI, Retention, Software, valuation, Workflows | Login |
| Jan 9, 2026 | Fund Letters | George Hadjia | BCG LN | Baltic Classifieds Group PLC | Communication Services | Internet Services & Infrastructure | Bull | New York Stock Exchange | Classifieds, Insiders, Monopoly, Networkeffects, rerating | Login |
| Jan 9, 2026 | Fund Letters | George Hadjia | 215A JP | Timee, Inc. | Industrials | Human Resource & Employment Services | Bull | New York Stock Exchange | growth, Japan, Labor, Marketplaces, Networkeffects | Login |
| Jan 9, 2026 | Fund Letters | George Hadjia | FND | Floor & Decor Holdings Inc. | Consumer Discretionary | Specialty Retail | Bull | New York Stock Exchange | Cyclicality, Housing, Operatingleverage, Procustomers, Specialtyretail | Login |
| Oct 10, 2025 | Fund Letters | George Hadjia | ASML NA | ASML Holding NV | Information Technology | Semiconductor Equipment | Bull | Euronext Stock Exchange | AI, Euv, growth, Lithography, Monopoly, semiconductors, technology, valuation | Login |
| Oct 10, 2025 | Fund Letters | George Hadjia | GOOGL US | Alphabet Inc | Communication Services | Interactive Media & Services | Bull | NASDAQ | advertising, AI, cloud, Hardware, innovation, monetization, Search, technology | Login |
| Oct 10, 2025 | Fund Letters | George Hadjia | SNPS US | Synopsys Inc | Information Technology | Electronic Design Automation | Bull | NASDAQ | AI, Design, Eda, growth, Margins, semiconductors, Software, valuation | Login |
| Oct 10, 2025 | Fund Letters | George Hadjia | PAR UN | PAR Technology Corp | Information Technology | Enterprise Software | Bull | NYSE | acquisition, ARR, growth, McDonald's, Restaurant, SaaS, Software, turnaround | Login |
| Jul 14, 2025 | Fund Letters | George Hadjia | HEM | Hemnet Group AB | Communication Services | Interactive Media & Services | Bull | NASDAQ | growth, Portals, Pricing, Property, Subscriptions | Login |
| Jul 14, 2025 | Fund Letters | George Hadjia | APG | APi Group Corporation | Industrials | Specialty Services | Bull | New York Stock Exchange | consolidation, Inspections, Margins, Recurring, services | Login |
| Jul 14, 2025 | Fund Letters | George Hadjia | CRM | Salesforce, Inc. | Information Technology | Application Software | Bull | New York Stock Exchange | AI, Automation, Data, SaaS, Software | Login |
| Jul 14, 2025 | Fund Letters | George Hadjia | UNH | UnitedHealth Group Incorporated | Health Care | Managed Health Care | Bull | New York Stock Exchange | healthcare, Insurance, Medicare, turnaround, valuation | Login |
| Mar 1, 2025 | Fund Letters | The Bristlemoon Global Fund | HEM.ST | Hemnet Group AB | Communication Services | Interactive Media & Services | Bull | Stockholm Stock Exchange | EBITDA margins, marketplace, Monopoly, Pricing power, Property Portal, Real Estate, SaaS, Sweden, Take rate, Vendor-Paid Model | Login |
| Mar 1, 2025 | Fund Letters | The Bristlemoon Global Fund | PAR | PAR Technology Corporation | Information Technology | Application Software | Bull | NYSE | ARR growth, cross-selling, Enterprise software, M&A strategy, Mission-Critical, platform, POS Software, Pricing power, QSR, Restaurant technology, SaaS | Login |
| Sep 30, 2024 | Fund Letters | The Bristlemoon Global Fund | APP | AppLovin Corporation | Communication Services | Interactive Media & Services | Bull | NASDAQ | adtech, AI, Data Moat, growth, high margins, machine learning, Mediation, Mobile Gaming, network effects, software platform | Login |
| TICKER | COMMENTARY |
|---|---|
| ASML | We believe semiconductor capital equipment (semicap) is currently the most attractive subsector within semiconductors because the tremendous demand for compute and memory will drive a much higher-for-longer wafer fab equipment (WFE) capex cycle. Our preferred semicap exposure remains ASML, which we first wrote about in our September 2025 letter. In that letter, we outlined and debunked the main bear arguments levied against the business, and the stock price has nearly doubled since. Despite the strong stock price performance, ASML has nonetheless derated materially against its large semicap peers. The historical premium that ASML used to command over the other four semicap majors (Applied Materials, Lam Research, KLA Corp and Tokyo Electron) has fully closed and the stock now trades at a discount to peers. We think ASML is one of the cleanest ways to play the next leg of the AI infrastructure buildout. The company sits at a critical bottleneck in the semiconductor supply chain, its EUV tools are becoming increasingly important across both leading-edge logic and DRAM, and our unit forecasts sit materially above consensus for both EUV and immersion DUV. While the stock has performed strongly, we do not believe expectations yet reflect the scale or duration of the coming WFE cycle, nor the potential for ASML to capture more of the value it enables. We continue to see healthy upside from here, and ASML remains a core position in the Fund. |
| APP | One example of this re-rate and de-rate cycle in play is AppLovin, which remains the Fund's largest position. The stock is the largest contributor to the Fund's returns since inception, but it has also been a significant source of month-to-month volatility for the portfolio. With the benefit of hindsight, the optimal decision would have been to reduce the position size, perhaps aggressively so, when the stock price breached $700 and the forward P/E multiple exceeded 50x. However, it is worth highlighting that while the stock is currently in a 28% drawdown from all-time highs and trading at 28x forward P/E, the stock price is still higher than it was in February 2025 when the forward P/E exceeded 70x. In other words, the company's explosive earnings growth was more than enough to offset the significant earnings multiple compression. When AppLovin staged its face-ripping rally from late 2024 to early 2025, our earnings estimates for the business remained well ahead of sell-side consensus (and buyside consensus as best we could gauge), so the stock was always trading at a much lower multiple to our numbers than the consensus numbers on stock screeners. By the time the stock reached $700 in late 2025 and early 2026, sell-side consensus had caught up to our estimates. So, at this point, sharp upward estimate revisions were no longer a driving force for the stock price. In hindsight, the optimal decision was to reduce the position size once sell-side and buyside estimates caught up to or exceeded our own, and the stock price continued to appreciate. |
| PAR | Hemnet and PAR Technology are two examples of this, with both stocks having been significant underperformers that have contributed to the bulk of the Fund's drawdown over the past year. While we believe these stocks will recover, we have been wildly wrong on the timing of when that recovery would occur, and the path with which those stocks would take before staging a recovery. These unforced errors in position sizing and averaging down have been very costly, and the only solace we can take from them is that they have prompted a lot of reflection that has resulted in process improvements that will hopefully reduce the likelihood of making similar position sizing mistakes in the future. |
| TSM | TSMC raised its 2026 capex guide to the high end of the $52-56 billion range and disclosed that capex over the next three years will be significantly higher than the past three years ($101 billion). Sell-side consensus currently estimates $197 billion in capex over 2026-28, but we believe that well over $200 billion cumulative capex is plausible considering TSMC is now playing catchup from a capacity perspective. The company has implemented three rounds of leading edge wafer price increases since 2025, with the latest set to occur in 2H 2026 at a reported +15%. These price increases clearly support higher capex plans. An unintended consequence of TSMC's conservative capital investment stance, especially during 2023-24 (remember leading edge fabs take 2-3 years to build and ramp to volume production), is that Intel Foundry has experienced a revival. Based on its qualitative capex outlook, we assume TSMC will continue to add around 100kwpm of leading edge capacity annually, which would require mid-40s EUV units. |
| INTC | In May, it was reported that Intel had reached an agreement with Apple to manufacture some of the latter's chips, though the product family and process node are currently unknown. This followed an earlier commitment from Tesla to utilise Intel 14A for some of its AI chips. Intel also recently secured orders for 3 million Google TPUs, although the announcement was ambiguous as to whether this was packaging only or includes front-end wafer foundry as well. In any case, these external customer deals are not expected to enter production before 2028, so Intel's near-term capex likely remains subdued (albeit with an upward bias). In the interim, 18A yields continue to improve towards mass production levels (90%+) and 18A-P – the performance-enhanced external customer node – entered risk production in June ahead of schedule. As far as 14A is concerned, when Intel CEO Lip-Bu Tan assumed the role in early 2025, he asserted that Intel was not going to invest in 14A capex until firm customer commitments had been secured. Tan's recent capex commentary on the podcast circuit has become far more constructive, particularly as it relates to 14A, which strongly suggests more customer commitments are to be expected. For Intel to support the growth of its own CPU business and have credible capacity to run an external foundry business, it likely needs to tool up at least three 40kwpm 18A/14A fabs between 2027 and 2030. For Intel, 120kwpm of new leading edge capacity by 2030 (three 40kwpm fabs) will likely require 50+ EUV tools, most of which will be backloaded to end-of-decade to support 14A external foundry volume ramp. |
| 005930.KS | At the end of June, Korean memory giants Samsung Electronics and SK Hynix both announced massive investment plans stretching through 2040, likely coerced by the strong invisible hand of the Korean government, with a mandate to spread their newfound memory-driven wealth across the nation. Samsung announced plans to invest 2,450 trillion won (US$1.6 trillion) between 2026 and 2040, of which 2,100 trillion is related to semiconductor production (though this seems to consolidate previously announced long-term investment plans). The most concrete incremental capital investment appears to be 800 trillion won (400 trillion each) to develop a new semiconductor cluster in the southwest of the country. Spread over a decade, this translates to 80 trillion won of incremental capex per year. Sell-side consensus already has Samsung and SK Hynix spending north of 150 trillion won in each of the next two years – figures that are likely to increase – and these new investment plans should be largely incremental. Samsung and SK Hynix have both pulled forward the completion of currently under construction fab projects: Pyeongtaek P5 Fab 1 for Samsung and Yongin Fab 1 for SK Hynix, both targeting wafer production in 2027/28, which means equipment installation commencing in 2027. |
| 000660.KS | SK Group, the chaebol under which SK Hynix sits, announced 2,100 trillion won of investments, of which 1,100 trillion would be for semiconductor production and 1,000 trillion for AI data centre investments. The most concrete incremental capital investment appears to be 800 trillion won (400 trillion each) to develop a new semiconductor cluster in the southwest of the country. Sell-side consensus already has Samsung and SK Hynix spending north of 150 trillion won in each of the next two years – figures that are likely to increase – and these new investment plans should be largely incremental. Samsung and SK Hynix have both pulled forward the completion of currently under construction fab projects: Pyeongtaek P5 Fab 1 for Samsung and Yongin Fab 1 for SK Hynix, both targeting wafer production in 2027/28, which means equipment installation commencing in 2027. |
| MU | Micron, the third member of the memory oligopoly, guided to north of $45 billion capex for FY27, which would exceed the cumulative capex spent from FY22-FY25. The company also pulled forward first wafer output from its new Idaho fab ID1 to mid-2027 from H2 2027, as well as meaningful volume contribution from its recently acquired Tongluo, Taiwan facility in mid-2027, a quarter earlier than scheduled. |
| AMD | Thanks to the rise of agentic AI, the mature CPU market is also expected to undergo rapid growth over the next few years. AMD – the other x86 CPU design company – upped its forecast 2030 CPU TAM to $120 billion in May, compared to $60 billion it projected just six months prior. |
| NVDA | Not to be outdone, Nvidia then came out with a $200 billion TAM forecast. When we look at ASML's customers, we can see for example TSMC earning 65% gross margins (vs 50-60% historically) and still pushing through 5-10% semi-annual price increases, memory vendors earning 85% gross margins (vs 20-30% through-cycle historically), and their customers, such as Nvidia, earning a 75% gross margin on top of that. |
| AAPL | In May, it was reported that Intel had reached an agreement with Apple to manufacture some of the latter's chips, though the product family and process node are currently unknown. |
| TSLA | This followed an earlier commitment from Tesla to utilise Intel 14A for some of its AI chips. |
| GOOGL | Intel also recently secured orders for 3 million Google TPUs, although the announcement was ambiguous as to whether this was packaging only or includes front-end wafer foundry as well. |
| Ticker | Put/Call | Amount Bought | Shares Bought | % Change | Weight % |
|---|---|---|---|---|---|
| No Recent Buys Data | |||||
| Ticker | Put/Call | Amount Sold | Shares Sold | % Change | Weight % | Status |
|---|---|---|---|---|---|---|
| No Recent Sells Data | ||||||
| Industry | Prev Quarter % | Current Quarter % | Change |
|---|---|---|---|
| No industry data available | |||