Meanwhile, I find a few themes worth exploring as the war continues into heating season.

European natural gas: Yes, TTF prices again. After a few months of ceasefire and no ceasefire, TTF prices reached EUR70/MWh, the highest since January 2023. The JKM Marker, Asia’s LNG benchmark, also reached its highest level since January 2023.

• EU gas storage is at 64%, compared to the average of 80% for the end of August in recent years. It is sitting at the lowest level in 13 years. The EU was waiting for lower prices, but it didn’t pull the trigger when prices were in the EUR40s, making it harder to refill at a low cost now.

• Britain is one of Europe’s largest gas consumers but has the smallest domestic storage capacity. Centrica CEO Chris O’Shea, whose company owns British Gas, said this week that Britain effectively has no gas stored for the coming winter.

• Germany, the country with the largest storage capacity in Europe, is only 50% full.

• Equinor, Norway’s largest oil and gas company, predicts the EU cannot reach its already-lowered 80% storage target. Norway is knowledgeable, given it supplies 30% of the EU’s natural gas.

• All Russian LNG will be banned starting in January 2027. Russian LNG comprises over 10% of EU natural gas supply.

• QatarEnergy extended its force majeure on LNG deliveries through October and early November.

Coal: Due to the natural gas situation, thermal and coking coal prices are seeing bids as countries scramble to maintain energy security. Coal remains the world’s largest source of power generation, producing nearly 11,000 TWh in 2026, while surging electricity demand is slowing efforts to displace the fuel with renewable energy. LNG disruptions also forced Asia to use more coal. Coking coal, used for steelmaking, is particularly strong, as Chinese supply declined 10% in June and July due to a Shanxi safety accident.

Crack spreads: With all the noise surrounding crude oil prices, Venezuela, and strategic oil reserves, the crack spread is quietly creeping up to record highs despite crude oil prices declining 20% from the peak.

• Crack spread measures the profits made by refiners, subtracting refined product selling prices from crude oil prices on a per-barrel basis. The most widely used benchmark is the 3-2-1 crack spread. It is calculated as [(2 × gasoline price + 1 × heating oil/diesel price) − 3 × crude oil price] ÷ 3, with all units being dollars per barrel. It calculates the per-barrel profit of refining 3 barrels of crude oil into 2 barrels of gasoline and 1 barrel of diesel.

Why are crack spreads at a record high?

• Low inventory: global refined product inventory (gasoline, diesel, jet fuel) is well below the 5-year average. As we get close to heating season and inventory depletes, product buyers, particularly heating oil (diesel) buyers, need to come back to refill inventory.

• Lack of refining capacity: US refineries are already running at 97% utilization in August, an 8-year high. Middle East and Russian refineries were damaged, which may take some time to come back. The Strait is not open, blocking refined product shipments. Product tankers face higher insurance costs, more complex scheduling, and greater destination risk.

• Lower crude oil prices, resilient product prices: crack spread increases if the cost of production is low (crude oil prices).

Before we dive in: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

Portfolio Update

Read more

Before we dive in: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

Tai Hing

Tai Hing (6811.HK) is a Hong Kong-based multi-brand casual dining restaurant group with over 215 locations across Hong Kong and Mainland China.

• Tai Hing: flagship brand. Signature dishes include chilled milk tea, roast pork, BBQ pork, barbecued and marinated platter, and roast goose.

• Men Wah Bing Teng: Hong Kong traditional “bing sutt” with Western-style fast-food snacks. Signature dishes include BBQ pork with egg and rice, egg in boiling water, and sa cha beef noodles.

• TeaWood: Taiwanese-style cuisine, such as noodles with stewed beef and mixed fruit tea.

• Asam Chicken Rice: Singapore and Malaysian cuisines, including Hainanese Chicken Rice, Laksa, and Penang Shrimp Noodles.

• Trusty Congee King: traditional Chinese cuisines, including various types of congee, noodles, and rice dumplings.

Business Segments (in HKD)

The company reports based on geographic segments.

#1 Hong Kong and Macau: $1.7b revenue, 93% of 1H26 total, +9% yoy. Please note that there is only one restaurant in Macau. The vast majority of restaurants and revenues are from Hong Kong.

• Tai Hing: $696m, +8%. New ad campaign to improve brand awareness and dining experience.

• Men Wah Bing Teng: $475m, +8%. Introduced new menus, added signature main courses, soups, and side dish combos for dinner. Increased customer traffic and table turnover rate.

• Asam Chicken Rice: $147m, +14%. Expanded footprint, launched new menu offerings, good marketing.

• Trusty Congee King: $106m, +31%. New, freshly prepared dishes, hospitality-driven dining experience. Awarded Michelin Bib Gourmand recognition for the 16th consecutive year.

#2 Chinese Mainland: $133m, 7% of 1H26 total, flat yoy. The number of restaurants and revenue have gradually declined over the last 5 years as the company scales back from the intensely competitive market.

Market Environment

Hong Kong’s catering industry faces intense local competition and northbound spending trends of Hong Kong residents. Business for restaurants was largely steady. Total restaurant receipts in the first half of 2026 increased by 0.8% to $55.5b, but down 0.1% by volume.

• Chinese restaurants (34% of 1H receipts): Stable, posting a +0.9% value increase in Q1 (though volume fell 0.2%), and improving to +1.5% in value (+0.7% volume) in Q2.

• Non-Chinese restaurants (34% of 1H receipts): Leading the industry growth, receipts grew by 2.8% in value (+2.1% volume) in Q1, and continued upward by +1.6% in value (+1.0% volume) in Q2.

• Fast food shops (22% of 1H receipts): Faced downward pressure, experiencing a drop of -0.6% in value during Q1 and worsening to -1.5% in value (-2.5% volume) in Q2.

• Miscellaneous Eating/Drinking Places (9% of 1H receipts): Recorded +0.5% value in Q1 before reversing into a -2.7% in Q2.

• Bars (1% of 1H receipts): The hardest-hit sector, dropping significantly by -4.0% in value in Q1 and continuing its decline by -3.8% in value (-4.6% volume) in Q2.

Financials, Valuation

Financials

• Margins: Gross margins maintained at 73% despite fluctuations in raw material prices due to the Middle East conflict. Cost efficiency measures and lease negotiations lowered staff and rental costs as a percentage of revenue by 1.5pct. Operating income increased by 59%, and earnings increased by 84%.

• Gradual Mainland exit: The number of restaurants in Mainland China decreased from 62 in FY2020 to 26 in 1H26. The Chinese Mainland segment has -1% operating margins, while the Hong Kong segment has 6% margins. As they continue to exit, margins should improve even if Hong Kong operations remain unchanged.

• Balance Sheet: $484m in cash, up $100m compared to year-end 2025. No financial debt, $950m in IFRS 16 lease liabilities. Over 95% of trade receivables and payables are settled within a month.

Valuation

• P/E: 10x on a LTM basis and 9x on a 1H26 annualized basis, even though 2H is historically the stronger half. If we assume the same earnings growth for 1H26 applied to FY25, then it is trading at a 7x forward multiple. Earnings in prior years were affected by impairment charges on property, plant and equipment due to the closing of Chinese Mainland locations.

• EV/EBITDA: 2x. Most of the EBITDA includes interest, depreciation, and amortization of leases.

• Dividend yield: 9% annualized based on interim dividends. Company historically pays more final dividends than interim dividends. Could reach 10% dividend yield based on past payout patterns and estimated earnings growth.

Largest Shareholder

• Founding family: 55% of S/O, second generation in management.

Risks

• Stagnant industry: there is no near-term catalyst that supports the growth in the Hong Kong catering industry. Tai Hing’s growth will come at the expense of market share gains from competitors. Cafe de Coral, one of Tai Hing’s major competitors, saw revenue decline by 4.5% in the fiscal year ended Mar/26. Fairwood, another competitor, saw 0% revenue growth in the fiscal year ended Mar/26.

• Intense competition: many Mainland brands have entered the Hong Kong market in recent years, often willing to operate at losses initially to gain a foothold.

• Operating costs: rental costs as a percentage of revenue have declined in the last 5 years due to a sluggish economy, as tenants were able to negotiate more favorable terms. Will landlords tolerate it if the economy recovers?

Thank You For Reading!

Hope you enjoyed the article. Please share, comment, subscribe, or pledge your support by donating to this page if you find it interesting!

Feel free to leave a comment on X as well! Capytal Management

For those who would like a brief overview of the company, please check out the link below.

Let’s dive in!

Before we continue: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

The Numbers and Year-on-Year Growth

• Revenue: $695m, +39%.

• Adjusted Profit Before Tax (PBT): $166m, +56%.

• PBT margins: 24%, up from 21%.

• ROE: 38%, up from 31%.

• Adjusted EPS: $1.72, +59%.

Segment Results

Clearing

• Revenue: $161m, +16%.

• Adjusted PBT: $79m, +12%.

• PBT margins: 49%, down from 51%. In line with 49-50% target.

• Drivers: higher average client balance ($19b), vs $16b in Q1 and $13b in Q2/25, partially offset by fewer contracts cleared (312m, vs 357m a year ago).

• My views: still impressive despite less contract volume. Clearing client balance growth appears structural as Marex earns more credibility.

Agency and Execution

• Revenue: $351m, +35%.

• Adjusted PBT: $117m, +69%.

• PBT margins: 33%, up from 26%.

• Drivers: prime (deeper institutional relationships), FX (expanding European FX client base, Hamilton Port acquisition), equities (strong momentum in derivatives), partially offset by lower energy revenues. Prime services becoming a bigger, higher-margin contributor.

• My views: not much to add except this segment is the stickier part of the business and is growing rapidly while margins expand.

Market Making

• Revenue: $118m, +106%.

• Adjusted PBT: $45m, +145%.

• PBT margins: 38%, up from 32%.

• Drivers: client activity across precious and base metals, aided by Middle East developments. Successfully integrated Winterflood, a leading UK equity market maker acquired last year. Energy benefited from hedging and trading conditions.

• My views: whether this part of the business can grow sustainably remains to be seen. Marex mentioned they achieved over $3m plus profits in 25% of trading days and only 4% of trading days are loss-making. Management believes market making performs the best when there are high levels of volatility, but not extremes. It will be interesting to see how they perform if markets indeed become extreme.

Hedging and Investment Solutions

• Revenue: $71m, +74%.

• Adjusted PBT: $25m, +295%.

• PBT margins: 35%, up from 15%.

• Drivers: commodities, FX, and structured products.

• My views: still a small part of the business and growing the fastest, but partially benefited from lower base comparison due to Apr/25 tariffs. Curious to see whether margin jump is temporary because it looks dramatic.

Management Outlook, Commentary

• 2025 M&A: $60m in total paid for Arna, Hamilton Court, and Winterflood for $16m total profits at acquisition (3x P/E purchase price). These companies now generate $60m in annualized profits, equaling the purchase price. Marex is not only good at purchasing companies at low multiples, but also at integrating them into the wider platform and generating synergies.

• BrightPoint International (BPI) acquisition: announced in Jul/26. BPI is a Singapore-based, Asian-focused clearing business with strong APAC and China client relationships. Offers commodities, FX, index futures and options, and digital asset derivatives with $800m of client balances. This was a “somewhat large purchase” bought at a low single-digit earnings multiple with identified synergies and high-quality, infrastructure-intensive earnings.

• M&A pipeline: not only strong, but the company is increasingly becoming the acquirer of choice. Repeatable strategy.

Other Notes

• Geographic diversification: majority of clearing client balance growth came from outside the US. Solutions business expanding distribution in Latin America, while BPI will strengthen its APAC presence.

• Client relationship deepening: 77 clients generated >$5m revenue, up from 49 in 2025 and 36 in 2024. Average revenue per client increased 34%. These clients include commodity producers, commodity consumers, banks, asset managers, hedge funds, and long-only funds.

• Organic/inorganic growth: 80%/20% split, expects annual growth target to remain 10-20% a year.

• Continuation of 1H26 trends in Q3 so far.

Overall Summary

• Thesis intact: Marex is not just a market volatility play, but a company with structural earnings growth due to its competitive positioning between the large banks that have balance sheet constraints and smaller competitors who are less competitive due to higher capital requirements. Delivered year-on-year profit growth in 19 of the last 20 quarters in a wide range of market conditions.

• Attractive financials: strong revenue growth, accompanied by even stronger earnings growth due to margin expansion. Management believes this can continue. Since its Q2/24 IPO, quarterly adjusted PBT has grown at an average rate of 48% year-on-year, with the upper quintile averaging 72% and the lower quintile averaging 21%.

• M&A strategy working: as demonstrated by the 2025 deals. Let’s observe its BPI integration and Asia strategy.

• Valuation still cheap: despite the 18% increase in share price, the company is still trading at 13x TTM P/E with 50% earnings growth.

Thank You for Reading!

Hope you enjoyed the article. Please like, share, and subscribe if you find it interesting!

Feel free to leave a comment on X as well! @capytalmgmt

Before we dive in: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

This company is Huishang Bank (3698.HK), a city commercial bank headquartered in Hefei, Anhui Province. It is the first banking and financial institution in the PRC jointly established by city commercial banks and urban credit cooperatives. In 2005, it merged with 5 city commercial banks and 7 urban credit cooperatives in Anhui.

In China, a city commercial bank is a bank that operates primarily at the municipal or provincial level, with ownership tied to local governments and regional state-owned enterprises. City commercial banks rank below the giant nationwide state-owned banks such as ICBC or the Bank of China. Their main role is to serve the local economy (residents and small- and medium-sized enterprises) overlooked by the major banks.

Operating Segments

Huishang Bank’s business model is relatively straightforward.

Corporate banking: banking products and services to corporate customers, government authorities, and financial institutions. These products and services include current accounts, deposits, overdrafts, loans, credit facilities, foreign currency, and wealth management.

Personal banking: banking products and services provided to individual customers, including savings deposits, retail loans, credit and debit cards, payments and settlements, and wealth management products.

Business scale (all financials in RMB unless otherwise)

• Total assets: $2.3 trillion, +15% yoy

• Total liabilities: $2.1 trillion, +16% yoy

Loans

To evaluate a commercial bank where the majority of profits come from lending, the key balance sheet items are loans and deposits. Loans are how they make money, and deposits show their sources of funding.

Loans: $1.1 trillion, +13% yoy

• Corporate loans: $800 billion, +13%. 70% of total loans

• Personal loans: $331 billion, +11%. 30% of total loans

Reasons for loan growth: $210 billion in lending to sci-tech enterprises (+25% yoy), over $150 billion in loans to both the manufacturing and strategic emerging industries, both growing over 20%. Agricultural loans increased 16% to $206 billion.

One of the negatives of Chinese banks is their exposure to real estate, which has been on a multi-year downturn. Huishang’s overall exposure to the real estate and construction industries is insignificant (7% of loans), with non-performing loan (NPL) ratios of 0.74% and 0.87% respectively. None of the top 10 single borrowers are from those industries.

Note: residential mortgage loans comprise 12% of total loans (part of retail loans). Although the mortgage-level NPL ratio is unavailable, we know the overall retail loan NPL ratio increased from 0.7% in 2020 to 1.89%. Luckily, mortgages have declined from 19% of total loans since 2020.

The company-wide NPL ratio declined from 1.98% in 2020 to 0.98%, meaning the rise in retail loan NPL has decreased in significance. I should also note that Huishang Bank acquired Baoshang Bank’s assets and liabilities in Beijing, Shenzhen, Chengdu, and Ningbo in 2020 when Baoshang restructured. I believe this demonstrated Huishang’s credibility and the government’s trust in the company to assume assets from a bankrupt bank.

Deposits

Out of all major bank funding sources, deposits tend to be the most sticky (although sometimes not the case, as seen in the 2023 US regional banking crisis).

Deposits: $1.3 trillion, +11% yoy

• Corporate deposits: $539 billion, +8% yoy

• Personal deposits: $727 billion, +14% yoy

Both corporate and personal deposits are growing, with personal deposits becoming a greater share of total deposits. In fact, this trend has been the case in the last 5 years. In 2020, personal deposits were 33% of total deposits. In 2025, this number increased to 57%.

Other Major Balance Sheet Items

Assets: financial investments worth $951 billion, 41% of total assets. Mostly debt securities.

Liabilities: bonds issued worth $373 billion, 17% of total liabilities.

Profitability - FY 2025

Revenue: $38 billion, +1%

Earnings: $17 billion, +6%

Despite the growth in loans and financial investments, revenue did not grow much in 2025 due to lower net interest margin and net interest spread.

Net interest margin (NIM) calculates the net interest (interest income minus interest expense) earned by interest-earning assets, while net interest spread calculates the difference between the average interest yield a bank earns on its assets and the average interest expense on its liabilities.

• Net interest margin: 1.49%, down from 1.71%

• Net interest spread: 1.31%, down from 1.49%

Why are they declining?

• China’s central bank has been cutting interest rates to stimulate growth. This affects the Loan Prime Rate, the benchmark rate at which loans are priced.

• Low-cost lending to align with state policy of supporting the real economy.

• Chinese consumers are shifting toward higher-yielding, long-term time deposits due to economic uncertainty.

• Previously locked-up long-term deposit rates remain high.

Result: Asset yield is declining faster than liability interest cost.

NPL ratio: 0.98%, down from 0.99%.

Allowance to non-performing loan ratio: 279%, down from 286%. This ratio measures the amount set aside in anticipation of loan losses against NPLs. Huishang Bank appears conservative by provisioning 279% of actual NPLs.

Allowance to loans ratio: 2.73%, down from 2.82%. This means 2.73% of total loans are eventually expected to default.

• The NPL ratio measures actual defaulted loans, while allowance is the expectation based on management discretion.

Valuation Metrics

The company is trading at very cheap multiples.

• P/E: 3.6x

• P/B: 0.35x

• ROE: 11%

• Tier 1 capital adequacy ratio: 10%

• Capital adequacy ratio: 14%

• Dividend yield: 6%

Shareholder Overhang

One of the reasons why the valuation is where it is today is probably due to Huishang Bank’s second-largest shareholder, Zhongjing Group (中靜集團).

Zhongjing Group is a Shanghai-based investment company founded in 1995. From 2007 to 2015, the company gradually accumulated Huishang Bank shares.

Unfortunately, the relationship has been contentious throughout the years. Zhongjing consistently demanded higher dividend payments and higher payout ratios, while Huishang elected to keep them low. Huishang wanted to IPO in the A-share market, while Zhongjing voted against.

In 2019, Zhongjing grew frustrated and agreed to sell its shareholdings to Shanshan Group (杉杉集團), a conglomerate with textile origins that has since expanded into renewable energy and battery materials industries. However, the transaction was only partially completed, as they had a legal dispute regarding transaction proceeds and the terms of the agreement. Some of Zhongjing’s shares were frozen by the court, while Shanshan had financial and internal power struggles.

Zhongjing tried finding other buyers, but to no avail. In 2024, Zhongjing ran into financial difficulties and defaulted on its bonds. This partly explains why Zhongjing constantly demanded higher dividends, because they needed any liquidity they could get.

The lack of liquidity and the ongoing litigation may have cast a shadow of doubt on the company, but in March 2026, Zhongjing reportedly considered a full share sale.

CXMT - A Potential Catalyst?

What does Huishang have to do with CXMT, the Chinese memory giant?

Turns out Huishang is an early investor.

Remember Huishang is headquartered in Hefei? As the primary provincial commercial bank of Anhui, Huishang participated in the Hefei Industry Investment Holding Group’s industrial investment fund in 2017, which helped fund CXMT’s early fab buildout and equipment procurement. Over the years, Huishang provided financial support to CXMT through syndicated bank loans and additional funding rounds through equity investments in private equity vehicles from its wealth management arm. It is reported that Huishang owns a 0.134% indirect stake in CXMT, with the investment’s unrealized gain around 25% of its annual earnings post-IPO. I will be curious to see how Huishang reflects its CXMT investment in the future.

I believe it doesn’t stop here. Huishang can benefit from greater business collaboration with CXMT given its decade-old relationship. Huishang’s market share of corporate deposits and loans is the highest among all commercial banks in Anhui Province.

Mainland investors can invest in Huishang Bank: Huishang officially entered the Hong Kong Stock Connect in Mar/26 and has seen its share price steadily increase since.

Anhui’s Macro Environment

Over the years, Anhui has transitioned from agriculture to high-tech manufacturing. In the first half of 2026, it officially entered China’s top 10 provinces by GDP, with above-average growth rates.

1H26 Statistics

• GDP: 5.6% growth to $2.74 trillion, 3rd highest nationwide

• Industrial value-add: 12% increase, led by 45% growth in high-tech manufacturing, 23% growth in equipment manufacturing

• Trade value: $617 billion, +34% yoy

• Auto powerhouse: #1 province nationwide in auto production, NEV production, auto exports by unit and value

• Online retail: $320 billion GMV, +16% yoy

As Semianalysis mentioned, Hefei is one of the tech innovation hubs in China; its patient state-venture capital nurtured BOE (the world’s leading display panel maker), NIO (a leading EV maker), and CXMT. Here is an excerpt from their CXMT article describing how Hefei built industrial ecosystems:

“The Hefei municipal government did two things for CXMT. First, the Hefei government helped CXMT build a local supply chain around its fab. Hefei’s playbook is to take a large equity stake in an anchor “link-leader” and then draw the rest of the chain in around it. The city did that for BOE in displays and NIO in EVs, and from 2016 it replicated the same playbook for CXMT. Surrounding CXMT’s plant in Hefei’s airport-zone industrial park, the government produced a dense local cluster. Two packaging-and-test houses, Peyton and Xinfeng, are located within a street or a wall of CXMT’s fab, with Xinfeng booking over 99% of its revenue from CXMT. An on-site bulk-gas plant run by Guanggang supplies most of CXMT’s needs, while wafer-reclaim capacity from Zhiwei Semiconductors, a subsidiary of Zhichun Technology (至纯科技), is located in Hefei’s Xinzhan Hi-Tech District. The state-venture-capital vehicles have also taken outright control of an upstream chip-molding equipment maker, Wenyi Technology. Such a local supply-chain cluster gives CXMT a localized industrial base.”

Risk

The biggest risk for Huishang (and most banks in China) is if benchmark interest rates continue to decline. It seems to be stabilizing in 2026, but macro is not the easiest to predict. China’s commercial bank net interest margins have not fully bottomed out, but the rate of decline has significantly moderated. As term deposits reprice, lower interest expenses may offset some of the loan rates that adjusted more quickly in prior years.

Thank You For Reading!

Hope you enjoyed the article. Please like, share, and subscribe if you find it interesting!

Feel free to leave a comment on X as well! @capytalmgmt

Let’s dive in!

Before we dive in: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

Q2 2026 Recap

There are really no surprises here, as the company continues to benefit from the increase in option trading volume, new product releases, and non-transaction fee growth. Operating leverage is working as well, with earnings growing faster than the already-robust revenue growth.

The Numbers (Growth on a YoY Basis)

• Net revenue: $141m, +35%.

• Adjusted EBITDA: $77m, +57%.

  • Margins: 54%, up from 47% a year ago.

• Adjusted earnings: $53m, +41%.

• Cash: $660m, against $2m of debt.

Option Segment Performance

The Options Segment is by far the most important, since it comprises the majority of the company’s net revenue and its profits subsidize other segments.

• Net revenue: $124m, +34%.

• Average daily volume (ADV): 11m contracts, +25%.

• Revenue per contract (RPC): +6%, but management expects RPC to decline in future quarters as MIAX grows market share.

• Market share: 16.5%, vs 16.7% in prior period. Slight trade-off between RPC and market share

• Non-transaction fee: +37%, driven by more members connecting to the MIAX exchange, 2026 fee increases, expiration of certain MIAX Sapphire trading floor-related fee waivers, and new market data fee products.

• Adjusted EBITDA: $97m, +44% (Options business profitability subsidizing developments in other segments).

Other Segments

Equities

• Net revenue: $5.5m, +27% from improved pricing.

• Adjusted EBITDA: -$0.5m, vs -$0.9m in the prior period.

Futures

• Net revenue: $5.1m, flat yoy.

• Adjusted EBITDA: -$9.5m, vs -$9.0m in the prior period.

International

• Net revenue: $5.7m, +151% primarily due to the acquisition of The International Stock Exchange Group (TISE).

• Adjusted EBITDA: $2m, vs -$0.6m in the prior period.

Corporate Updates

• Volatile markets in Q2: geopolitics, trade, and AI-related stock swings led to elevated option volumes. Sustained market volatility leads to stronger demand for risk management tools.

• Bloomberg Financial Futures: market depth and volume in line with expectations. The next step is to engage retail by reaching out to trading firms and having them educate and market the products. MIAX will onboard liquidity providers over the next few weeks.

• Option market share: appears flat year-over-year but grew again in July to 17.1%. Market share continued to grow in early August.

• Nasdaq litigation: resolved and settled.

• Strong growth in non-transaction fees: increased market participant access fees on January 1.

• Cash balance: prioritize investments in the Bloomberg futures business and in new trading-floor functions. Member firms from around the world are coming to MIAX with new opportunities. No plans for dividends or buybacks (which makes sense given the company only listed a year ago).

Quick Thoughts

• As expected, volatile markets mean greater business prospects for hedging, speculation, and income generation. MIAX’s option volume grew 25% in Q2, and the industry grew 27%. As mentioned in the earnings call, MIAX is regaining market share in July and in early August.

• Operating leverage is fantastic, with EBITDA margins growing 700 basis points from a year ago. Mature exchange operators such as CME and ICE have 65-70% margins. MIAX still has room to grow without significant incremental investments.

• There is value in MIAX’s market data and exchange operator status. Not seeing SaaS-related issues and the impact of perpetual futures so far. I continue to believe perpetual futures is a product that serves a different client base.

• Even if more capital flows to the prediction market, MIAX’s 10% ownership of Rothera (derivatives exchange and clearinghouse joint venture between Robinhood and Susquehanna International Group) will benefit.

• The cash balance seems to be accumulating faster than they can spend. Hopefully it will be deployed in high return projects.

Thank You For Reading!

Hope you enjoyed the article. Please like, share, and subscribe if you find it interesting!

Feel free to leave a comment on X as well! @capytalmgmt

Let’s dive in!

Before we dive in: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

Q2 2026 Recap

To summarize the second quarter, I would simply say continuous execution according to plan, plus a favorable industry outlook heading into the winter heating season.

The Numbers

• Production: 17,125 boe/d, +6% qoq (quarter-over-quarter), despite major scheduled turnarounds. Production estimate for July is 23,000 boe/d.

• Funds flow from operations (FFO): $74 million, +15% qoq.

• Netback: $69.05/boe, +20% qoq. +47% yoy from higher EU natural gas exposure.

• Net debt: $378 million, -3% qoq due to cash flows generated in Q2. <1x 2026E FFO.

• Capex: $58 million, no change for FY26.

• Free cash flow: $16 million.

Operational Activities

• Production wells: 4 (1.9 net) wells drilled in Q1/26 now producing.

• Development wells: 2 (0.9 net) wells drilled.

• Maintenance: L2, L9, and N05 platforms, planned turnaround at the Den Helder plant.

• Took delivery of Triton-10 jack-up barge for 2H workover and beyond.

TEN

• K07-FB-103: exploration well at TEN, producing in Q2.

• K17-FA-103: second exploration well, completed drilling in Q2. Flowed at a rate of 15.7 MMcf/d (9.4 MMcf/d net) during clean-up. Being prepared for tie-in in Q3.

• K17-FA-102B: new sidetrack well, drilling underway.

GEMS

• N05-A-02X: completed drilling in Q2 and brought on production at a restricted rate of 74 MMcf/d (25 MMcf/d net) in early Q3. Similar reservoir characteristics to the N05-A-01 well, which continues to produce at 74 MMcf/d. These two are the highest-producing rate wells in the Netherlands. 100% success rate in the drilling campaign to date.

• Commenced drilling of the Noord extension to the N05-A pool.

Other Asset Updates

• Seasonal maintenance: L2 and L9 platforms, routine maintenance at Den Helder, Eni and GEMS planned maintenance. Production was reduced by 2,500 boe/d, which has been restored.

• Took delivery of Triton-10 jack-up barge for 2H26 workover and beyond. Multi-year program.

European Natural Gas Outlook

Perhaps the most important question regarding Tenaz’s outlook comes down to TTF, European natural gas prices.

• Q2 recap: Middle East conflict persisted, temporary ceasefire and brief resumption of limited LNG shipments through the Strait. TTF benchmark price averaged €45.64/MWh (+16% qoq) and ranged between €40 and €50/MWh for most of Q2.

• Post Q2: ceasefire violated, shutting down commercial shipments. LNG supply constraints put upward pressure on prices. QatarEnergy extended the force majeure on LNG exports through mid-Oct, signaling that exports will remain disrupted.

• EU storage levels: 58%, vs. a 5-year average of 75%. This is at a time when Europe needs to compete with Asian LNG demand since pipeline supply in Europe no longer has meaningful winter flexibility. The EU’s mandatory 90% storage target is now relaxed, allowing 80% for member states. Industry forecasts suggest that the EU will only reach 70-75% by the end of the injection season, meaning TTF prices should be supported. The TTF forward curve is reflecting this dynamic, showing €51.50/MWh through the 2026/27 heating season.

Corporate Updates

• Debt trading: Senior Unsecured Notes, with a total principal of $305m, are trading strongly in the secondary market with a 5.9% yield-to-worst pricing. Remember the two bonds, with the same call and maturity dates, were issued at yields of 12% and 9.5%. This means investors are seeing Tenaz’s improving credit quality and are willing to bid up the bonds to accept a lower yield. This also benefits Tenaz, as its future debt funding costs will be lower. $35m of the $250m revolving credit facility drawn in Q1 was repaid.

• Share repurchases: 143,100 shares repurchased at a weighted average price of $49.78 per share during Q2 (I should mention that 110,000 shares were repurchased in June). Unfortunately, there was no hint of whether they see buybacks at current prices as attractive. We shall continue to monitor the monthly buyback report.

• Hedging: 56% of expected 2026 TTF natural gas production, 41% for 2027, and 8% for 2028 are hedged at around €30-34/MWh. I was expecting more aggressive hedging to lock in higher prices. It appears they want to maintain flexibility to capture the upside, and the hedging program is simply for securing enough cash flows to satisfy capex requirements.

• No mention of M&A pipeline: probably less likely to do deals given the strong commodity pricing, as sellers would demand higher asset selling prices. Luckily, Tenaz’s existing assets have enough organic reinvestment opportunities.

Thank You For Reading!

Hope you enjoyed the article. Please like, share, and subscribe if you find it interesting!

Feel free to leave a comment on X as well! @capytalmgmt

As mentioned in my June update, I became increasingly worried about the debt-funded hyperscaler capex. I concluded I am unable to determine whether and when these investments will pay off. Hyperscalers’ credit and CDS spreads have now surpassed IG industrials, according to Apollo. They continue to accumulate massive, off-balance-sheet leases and purchase commitments that will eventually hit the financials (for those interested, check the leases and commitments section of the hyperscalers’ 10-Qs and 10-Ks; we are talking about many hundreds of billions for each company). If these gigantic investments pay off, I don’t mind being proven wrong. I simply think the odds are not on my side.

The rapid drawdown and eventual collapse of the Situational Awareness public fund caused a few days of extreme volatility in AI names. Of course, exiting the entire theme prevented me from losing capital. Some of these stocks rebounded when Citadel bought Situational Awareness’s entire public book, which I can in no way predict ahead of time. I have no issues missing out on that or if there is an eventual recovery.

"There are only three ways a smart person can go broke: liquor, ladies, and leverage." Charlie Munger was right. Warren Buffett noted that Munger added the first two for the alliteration, with leverage being the main point.

Before we dive in: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

Portfolio Update

Read more

Rather than forming a view on so many themes, why not benefit from the mood swings by owning companies that thrive on market fluctuations?

Here are my quick notes on two such ideas: Marex and Miami International.

Before we dive in: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

Marex Group (NASDAQ: MRX)

I have written a bit about Marex in my latest portfolio update. Here is the full summary plus additional research.

Marex (MRX.NASDAQ) is a financial services platform that specializes in clearing, agency and execution, market making, and hedging solutions to hedge funds, traditional investment managers, and family offices.

• Clearing: Processes, clears, and settles futures and options to over 60 exchanges worldwide.

• Agency and Execution: Independent broker to execute trades on behalf of mid-market institutional clients, either on-exchange or over-the-counter.

• Market making: Transactions using its own balance sheet.

• Hedging: Customized risk-management solutions and derivative products, bespoke derivative hedging.

Revenue model

• Clearing: Commission per trade (facilitate and clear trades for clients), interest income (interest earned from client collateral and cash balances).

• Agency and Execution: Commission per trade (fees per transaction matched/executed), product spread (price differences from buying and selling).

• Market making: Spread between buying and selling prices, with limited overnight risk.

• Hedging: Product pricing (profits embedded within the pricing structure of custom solutions).

The industry is characterized by declining competition amongst Futures Commission Merchants (FCMs). FCMs are regulated financial institutions that act as brokers for futures, options, and cleared swaps. They act as intermediaries by executing trades for customers and monitoring customer margins. If customers default, FCMs are required to step in and cover losses using their own capital, so the financial exchange doesn’t collapse. Marex sits strategically between the small, specialized boutique brokers and large investment banks. Here are the advantages of a mid-market FCM:

• Large investment banks retreat: Post-GFC, capital requirements severely penalized large investment banks from clearing and trading commodities. Many either focus on their largest clients or completely exit the physical commodity market. Marex is not a bank; it doesn’t have the same constraints.

• Small, boutique exits: Regulatory barriers force boutiques to invest a disproportionate amount into technology and compliance to achieve scale, making the business less profitable.

Marex took advantage of the industry backdrop by acquiring smaller FCMs and taking clients from large investment banks. Marex’s growth was led by JRJ Private Equity, a London-based PE firm specializing in the financial services sector.

Source: Marex Investor Day 2026

JRJ Involvement

• 2010: JRJ Group, along with co-investors, acquired a 74% controlling majority stake in Marex. JRJ Group’s co-founder, Roger Nagioff, became CEO of Marex, bringing a team of former Lehman Brothers bankers to scale the business.

• 2011: Acquired Spectron Group, creating one of the world’s largest independent commodity brokers across OTC energy and exchange-traded markets. Revenue at this stage was $230m.

• 2016: Launched Neon, an integrated trading, risk management, and data platform designed to enhance client market access.

• 2017: Created Marex Solutions, a custom hedging solutions and structured financial products platform.

• 2019: Launched Marex Clearing Services, acquired CSC Commodities (energy market-making), and Rosenthal Collins Group (North America expansion). Marex is a $352m revenue business.

• 2020-23: Made a series of acquisitions, including Tangent Trading (recycled metals trading), XFA (North American exchange-traded derivatives execution broker), ED&F Man Capital Markets (clearing and securities execution), and TD Cowen’s prime brokerage and outsourced trading business. Revenue reached $1.2b.

• 2024 IPO: Marex was listed on the NASDAQ in Apr/24, and JRJ has gradually reduced its stake.

M&A Strategy

Marex’s M&A criteria focus on bolt-on acquisitions at 5-6x P/E, targeting a 3-year payback.

• Acquired TD Cowen’s prime brokerage and outsourced trading business in Dec/23. Grew revenue from $80m in 2023 to $258m in 2025. Prime Services now comprise 25% of Marex’s group profits.

• Bolt-on 2025 deals: Aarna Capital (Abu-Dhabi clearing business, expand in the Middle East), Hamilton Court (UK FX specialist, UK + EU corporates), Winterflood Securities (UK equity market maker).

• Marex believes their total addressable market is over $80b, and they estimate they currently have 2% market share, with massive opportunities to grow organically and through bolt-on acquisitions.

FY 2025 Financials and YoY Growth

• Revenue: $2b, +27%

• Adjusted profits before tax: $418m, +30%

• Profit before tax: $412m, +39%

• EPS: $4.12, +39%

Q1 2026 Financials

• Revenue: $692m, +48%

• Adjusted profits before tax: $153m, +59%

• Profit before tax: $150m, +53%

• EPS: $1.52, +55%

The company is trading at a 14x P/E multiple on a TTM basis with accelerating growth. 75% of the profit increase in 2025 was driven by organic growth.

Risks

• Client activity drives earnings. If the market is stable, Marex will not grow as fast. If the market is too volatile and customers lose creditworthiness, Marex will be negatively impacted. I believe the customer credit risk is mitigated through Marex’s limited overnight business and infrastructure-focused revenue base.

• Near-term selling pressure from JRJ. They now own less than 10% of Marex.

Miami International Holdings (NYSE: MIAX)

MIAX is a holding company that builds, operates, and acquires financial exchanges, with a focus on options and futures. The company’s co-founder, Thomas Gallagher, identified the high-performance quoting demands of the US options trading industry and launched its first fully electronic options exchange in 2012. MIAX competed against incumbents such as CBOE, NASDAQ, and NYSE by investing in technology to offer high-speed, low-latency trading systems.

To build market share and attract traders, order flow providers, and market makers, MIAX established equity rights programs (“ERPs”). MIAX issued performance warrants with the right to buy shares of MIAX to participating traders when certain liquidity volume targets are met. A total of five ERPs were completed, which led to significant market share gains. MIAX currently has a 17% market share in US multi-listed options.

Before the company was listed, notable investors included Warburg Pincus, Hudson River Trading, and Hamilton Global Opportunities.

Source: June 2026 Investor Presentation

Following the success of the option business, MIAX expanded into cash equities, futures, and international listings through acquisitions and product launches.

Here are the markets they operate in today:

US Options

• Four fully automated electronic options exchanges in the US, with different allocation and pricing models.

• Market share: 2.4b contracts traded in 2025, +41% increase, and representing a 17% market share.

Source: MIAX website as of July 7, 2026

US Equities

• Fully automated electronic cash equities trading platform. Launched in Sep/20.

• All U.S. equity securities designated under the Regulation National Market System (“NMS”) qualifying for Unlisted Trading Privileges (“UTP”).

• Market share: 45.8b shares traded in 2025, -8%, representing a 1% market share.

US Futures

• Fully electronic futures exchange and clearinghouse. Acquired in Dec/20.

• Growing demand for 3rd party central counterparties to add stability to the financial system and reduce market risk.

• Market share: 3.3m contracts traded in 2025, +2%, representing <1% market share.

International Listings

• Bermuda Stock Exchange: Specializes in the listing and trading of Insurance-Linked Securities (ILS). ILS are financial instruments that allow insurance and reinsurance companies to transfer risks to the capital markets. One example of ILS is catastrophe bonds. Insurance companies sometimes issue bonds linked to natural disasters to investors. If the qualifying event doesn’t occur, investors get both interest and the principal. If it occurs, the bond issuer (insurance company) keeps the principal to pay claims, and investors take losses. Acquired in Dec/20.

• The International Stock Exchange (TISE): Leading market in Europe for listing high-yield bonds, private equity debt, asset securitization, and listed UK Real Estate Investment Trusts. Acquired in Jun/25.

Others

• Bloomberg Partnership: Futures and options on the Bloomberg 100 and Bloomberg 500 indexes, providing investors with an alternative way to manage US equity market exposure. Bloomberg indexes market themselves by being transparent, rules-based, and allowing high-profile IPOs to be eligible for inclusion sooner, rather than having committees decide on constituents. MIAX began launching Bloomberg-related products in May/26.

• MIAXdx (Rothera): Sold 90% of MIAXdx, a CFTC-regulated prediction-market exchange and clearinghouse, to a joint venture between Robinhood and Susquehanna International Group. Rothera began operations in Jun/26.

Industry Drivers

• Trading volume: Market volatility, hedging demand, interest rates, corporates managing commodity exposure, geopolitics, and retail participation.

• Product launches: Bloomberg, short-dated options (0DTE), structured products such as leveraged ETFs with embedded options, and prediction markets.

Revenue model

• Transaction and clearing (61% of net revenue): Fees charged to exchange member firms for executing and clearing trades on MIAX exchanges.

• Access fees (25% of net revenue): Fees for allowing customers to connect their networks to MIAX exchanges.

• Market data fees (10% of net revenue): Direct market data subscription fees or distributions from US Tape Plans. Also includes revenue earned from selling proprietary market data directly to customers.

• Others (4% of net revenue): Initial and annual listing fees.

FY 2025 Financials and YoY Growth

• Net revenue: $431m, +56%

• Adjusted EBITDA: $199m, +142% (46% margins, up from 29%)

• Adjusted earnings: $165m, +267%

Q1 2026 Financials

• Net revenue: $129m, +40%

• Adjusted EBITDA: $66m, +66% (51% margins, up from 43%)

• Adjusted earnings: $45m, +51%

• Cash and cash equivalents: $551m

• Debt: $2m

Valuation: 19x TTM EBITDA, or 15x 2026E EBITDA

Risks

• Slowdown in trading volume.

• Competitor technological investments.

• Alternative trading products (stock price dropped due to fears surrounding perpetual crypto futures).

Thank You For Reading!

Hope you enjoyed the article. Please share, comment, subscribe, or pledge your support by donating to this page if you find it interesting!

Feel free to leave a comment on X as well! Capytal Management

Mr. Market is as moody as always. So many headlines, so much fear and greed, and AI is what everybody is talking about right now.

The summary below includes datapoints I find interesting.

Before we dive in: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

AI Data Center Buildout

United States

According to various investment banks, US hyperscalers are expected to spend over $700B on capital expenditures this year. Amazon is guiding $200B, Alphabet $185B, Meta $135B, and Microsoft $190B. This doesn’t include Oracle and the so-called neoclouds like CoreWeave and Nebius, adding another $100B. Big Tech’s business models are rapidly changing from capital-light to capital-intensive, often deploying 30-40% of annual revenue to capital expenditures.

A few years ago, the majority of their capital expenditures were funded by operating cash flows, meaning they could support these spending plans without relying on external funding sources. Starting last year and increasingly so this year, they are tapping capital markets, utilizing their investment-grade credit ratings to raise debt and, in some cases, issue shares as well.

• Amazon issued $54B of bonds.

• Alphabet raised $35B through bonds and another $85B (over time) through equity. This doesn’t include the billions in backstop guarantees they provide to companies that are leasing data centers from 3rd parties.

• Meta raised $30B of bonds in late 2025 and another $25B this year. This doesn’t include joint venture data center project financings.

For the first time in a while, debt investors can allocate significant sums to these dominant, trillion-dollar tech giants.

Companies borrow when they spend more than their current resources allow. Perhaps AI is so transformational that all the capex is warranted and the debt is well-supported. Are they all going to win, or is the market winner-take-all? What happens to the value of all the other investments if only a select few companies win? Maybe all will win?

China

Due to restrictions on advanced chips and semiconductor equipment, China’s AI data center buildout lags materially behind the US. China’s advantage is a more coordinated power system because, as we know, the US grid is congested. As domestic chips and AI models catch up, we are seeing spending plans heat up.

China, led by the National Development and Reform Commission, is preparing to allocate $300B building data centers across the country over the next 5 years. This, of course, doesn’t include numbers from hyperscalers such as Alibaba, Tencent, and ByteDance. Deepseek recently raised $7.4B and plans to double its workforce. According to Goldman Sachs, Chinese hyperscalers are expected to spend $100B on capex in 2026, much less than their US counterparts. They have been raising money as well, although at a much smaller scale.

EU Gas Storage

As of the time of writing, TTF natural gas prices retreated from >EUR60/MWh to EUR43/MWh with the Middle East conflict ending in sight. European storage facilities are currently around 48% full, below the five-year average of 61% at this time last year. At the same time, an intense heatwave across Europe is boosting demand from electricity providers to meet increased cooling needs. On the other hand, Qatar is resuming production, aiming to reach 80% of pre-war production in a month.

Portfolio Management

Unfortunately, many of my picks have not worked out since the end of Q1. My buy-and-sell decisions have not helped either. I have no idea whether those stocks are temporarily out of favor or if I made mistakes evaluating them. Only time will tell. I think I did a better job cutting losses and not letting them snowball, but at the same time, my position sizing was subpar. All of the YTD gains have been contributed by Tenaz, which I exited at the $60-62 range. Losses in Nam Cheong, Harbin, Chongqing, Bloom, etc., were slightly offset by Precision Tsugami, China semiconductor ETF, and SK Square.

On the brighter side, this allowed me to reevaluate my portfolio. In the current market environment, knowledge as an investment edge is rapidly commoditized, while emotional discipline is as vital as ever. Luckily, I am not reporting to any allocators, and investing is a marathon, not a race.

I hope my subscribers performed better with the same names.

Holdings

Read more

Before we dive in: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

Kaspi (NASDAQ: KSPI)

Kaspi operates a “Super App” in Kazakhstan. For those unfamiliar with this concept, a Super App is an all-in-one app that integrates services used on a daily basis in one place. A prime example is WeChat. WeChat is China’s most widely used mobile app, with 1.4 billion monthly active users as of 2026. WeChat facilitates payments, operates a social networking platform called Moments, and offers wealth management products such as deposit accounts and investment funds. Merchants accept WeChat for a cashless purchase, and advertise on WeChat’s massive social media platform.

Kaspi has a similar Super App. Although Kaspi doesn’t have a social media platform like Moments, it offers payment, fintech, and e-commerce services to consumers and merchants in Kazakhstan.

• Dominant market share: As of 2025, Kaspi has 15.7 million MAUs (monthly active users) in Kazakhstan, processing approximately 80% of all e-commerce payments and over 70% of all non-cash retail transactions nationwide.

• Sticky customer base: 77 monthly transactions per active consumer.

Business Segments

Payments (~15% of Revenue)

Facilitates transactions between merchants and consumers.

• Consumers: Shop online, pay regular household bills, and make money transfers.

• Merchants: Accept online payments, issue and settle invoices, and pay suppliers.

Marketplace (~50% of Revenue)

E-commerce platform, integrated 3rd-party logistics partners.

• Consumers: Buy general goods, groceries, travel services, and even used cars.

• Merchants: Sell products through product listings and advertisements.

Fintech (~35% of Revenue)

Provides consumers and merchants with Buy-Now-Pay-Later (BNPL) solutions, financing, and deposit products.

• Consumers: BNPL (unsecured financing generally up to 3 months, or 6-24 months for consumer purchases on the Marketplace Platform), general loans (consumer day-to-day purchases outside the Marketplace Platform), car loans, and deposit accounts.

• Merchants: Working capital financing, Buy-Inventory-Now-Pay-Later (BINPL), and deposit accounts.

Turkey Expansion

After establishing a dominant position in Kazakhstan, Kaspi has its eyes set on Turkey. Turkey has over 85 million digitally-engaged consumers, more than four times the size of Kaspi’s home country, but with much lower e-commerce penetration. It entered the market through two investments:

• Hepsiburada acquisition: In early 2025, Kaspi invested US$1.3 billion for an 87% stake in one of the largest e-commerce platforms in Turkey called Hepsiburada. After significant investments, e-commerce orders growth improved, trending from -11% in Q1/25 to +7%, +16%, and +19% in Q2-Q4/25. Kaspi Co-Founder and CEO Mikheil Lomtadze highlighted the Turkish e-commerce opportunity below.

• Rabobank’s Turkish subsidiary: Although Rabobank doesn’t have borrowing or depositing clients nor a branch network in Turkey, this acquisition is intended to obtain a banking license in order to offer financial services in Turkey. It was approved on June 24, 2026, and the deal is expected to be closed next month.

2025 and Q1 2026 Results

2025 was a decent year for Kaspi, with double-digit growth rates across Payments, Marketplace, and Fintech, despite a challenging macro environment. The company highlighted regulatory impact on smartphone registration, 10% tax on government bonds, increased minimum bank reserve requirements, and base rate increase from 15.25% to 18%.

FY2025

• Revenue: +19%

• Net income: +10%

Q1 2026

• Revenue: +31%

• Net income: -1% due to higher interest expense and Turkey e-commerce platform investments

KPIs to Monitor

Payments: TPV (total payment volume) and take rates.

Marketplace: E-commerce GMV (gross merchandise volume), advertisement.

Fintech: Loan portfolio growth, yield, net interest spread.

Valuation, Management

Valuation: Kaspi is currently trading at 7.7x P/E, with an 8% dividend yield and a 64% payout ratio. The company reinstated the KZT 850/share dividend in Q1/26 after completing the Turkey e-commerce transaction and believes it is sustainable.

2026 guidance looks a bit underwhelming on the surface, but this year’s heavy investments should pay off with future earnings growth.

Co-Founder and CEO Mikheil Lomtadze: Georgian-born, grew up with limited financial resources. He once said, “Everything we had went into our education. We didn’t have food on the table, but I always had a teacher.” He has extensive business experience and studied at Harvard Business School.

• Founded Georgia Consulting Group Audit, later sold to EY.

• Partner at private equity firm Baring Vostok Capital Partners, one of the largest PE companies in Russia and the CIS.

• Stepped in as Kaspi CEO after Baring Vostok invested in 2006, when the company was struggling.

• Pivoted away from corporate lending to a consumer-focused ecosystem, including a digital marketplace, QR code payment systems, and banking services.

Recent investment: In Apr/26, Kaspi announced that Tencent, Mikheil Lomtadze, key members of the senior management team, and long-term institutional investors purchased 6 million ADS (3% shares outstanding) from Baring Fintech Venture Funds. Mikheil Lomtadze already owns 23% of Kaspi. As of Q1/26, Tencent’s investment is purely financial with no strategic announcements yet.

Risks

Emerging markets: Both Kazakhstan and Turkey have significant inflationary pressures and high interest rates. Kazakhstan’s latest inflation and interest rates are 10% and 18%, while Turkey’s rates are 33% and 37%. These macro indicators can also affect FX rates, which can affect USD-based dividend payments.

Execution risk: Kaspi’s success in Kazakhstan might not be easily replicable in Turkey, given the different culture and consumer habits. WeChat’s super app business model is not adopted globally.

Government policies: Kazakhstan increased the corporate tax rate from 20% to 25% this year. The central bank continues to maintain high capital requirements and high benchmark rates. These policies can affect consumer spending and merchant activities.

Thank You For Reading!

Hope you enjoyed the article. Please share, comment, subscribe, or pledge your support by donating to this page if you find it interesting!

Feel free to leave a comment on X as well! Capytal Management

Before we dive in: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

Portfolio Positioning

Read more

Let’s dive in!

Before we dive in: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

Beng Kuang Marine

Financials (in SGD)

• Revenue: $26m, +8% yoy

• Gross profit: $7m, -22% yoy

• Profit before tax: $3m, -28% yoy

• Net income: $3m, -13% yoy

Revenue growth: FPSO infrastructure engineering services.

Gross profit and margins: Lower due to more revenue from shipbuilding and other earlier-stage work, earning less than FPSO services.

Orderbook: $56m as of Mar/26, $51m FY26 revenue secured. Additional opportunities expected from West Africa contract renewals.

Other news

• Co-founder Chua Meng Hua divested his stake to institutions (Amova Asset Management under Sumitomo Mitsui Trust, Tokio Marine Life Insurance, and Tan Kim Seng from Tembusu Partners), and management (Chua Beng Yong, Executive Chairman, and Yong Jiunn Run, CEO). Not particularly worried since he has not been active in managing Beng Kuang.

• AGM: Mentioned the latest JV with Epsilon Navigation will involve just 1-2 vessels initially, oversee construction progress, followed by chartering and eventual vessel sale. Upskilling employees to expand the service scope to be retained for the West Africa FPSO.

Personal Views

• Profits and margins declined more than I expected because of higher revenue contributions from lower-margin segments. Management claims more non-FPSO projects are underway. I would still want the higher-margin FPSO service segment to be the majority of the company’s revenue.

• Cost-conscious: admin costs down 13%.

• Happy to see the CEO pick up more shares at $0.48/share from the co-founder. He is the person who engineered the business turnaround in recent years. Having some institutional investors doesn’t hurt either.

• Monitor West Africa FPSO service contract updates and ASOM acquisition closing.

Marco Polo Marine (MPM)

Financials (in SGD)

• Revenue: $74m, +40% yoy

• Gross profit: $31m, +45% yoy

• EBITDA: $29m, +87% yoy

• Adjusted net income: $14m, +44% yoy

Revenue Growth Drivers

• Ship chartering revenue grew 38% due to a larger fleet beginning in 2H25. Average fleet utilization was 65%, unchanged yoy.

• Ship building and repair revenue grew 43% due to higher volume of ship repair projects and higher contract value. Volume growth is supported by expanded shipyard capacity.

Gross profit and margins: higher due to revenue growth and resilient day rates. Securing long-term charters to provide a durable and recurring earnings base.

Company Outlook

• Oil prices are supportive of OSV demand, expect day rates and utilization to be stable across SE Asia.

• Two additional AHTS vessels to be delivered in FY26. The company is renewing its fleet to replace old vessels.

I was very satisfied reading up to this point, until they simultaneously announced a reverse takeover (RTO) transaction by Fuji Offset Plates Manufacturing (FOPM).

RTO Key Points

• $139m valuation, comprising a base consideration of $120m and a deferred consideration of $19m structured as an earn-out.

• Source of funding: Share issuance by Fuji Offset Plates Manufacturing, a Catalist-listed company at $0.701/share.

• Marco Polo Marine to remain a controlling shareholder, owning 74% of FOPM, which may rise to 77%.

• Rationale: An independent, separately listed shipyard business whose value the market can recognize from fully reportable financials. Easier to conduct future capital-raising.

My concern is, why do MPM shareholders need to absorb FOPM? Then I found out that Apricot Capital, a 12% shareholder of MPM, is a substantial shareholder of FOPM as well. FOPM is a tiny $3.3m-revenue business that sells printing cycliners and owns a bunch of rental properties. This means Apricot is using MPM’s shipyard business to boost the value of its FOPM holdings.

Personal Views

• MPM’s fundamentals are solid, with good exposure to oil and gas and offshore wind. Utilization and day rates remain solid according to the company’s investor presentation.

• I don’t like the reverse takeover because it seems to be a way to monetize FOPM by diluting the shipyard business. Apricot might have too much control over MPM to facilitate such a deal, unless they plan to end their legacy business. If the shipyard business should be independent and needs capital, a spinoff would be more shareholder-friendly. This to me is a red flag.

Nam Cheong

Financials (in RM)

• Revenue: $118m, +1% yoy

• Gross profit: $50m, -12% yoy

• Operating profit: $32m, -28% yoy

Revenue: Only up 1% due to a smaller fleet from fleet disposal. Vessel utilization increased from 48% in Q1/25 to 58% in Q1/26. The company expects utilization to improve in Q2.

• 5 new vessels are scheduled to join the fleet in 2026.

Gross profit: Higher operating costs from a vessel deployed in the Middle East.

Other income: I excluded $61m in other income from operating profit calculations. This is a gain from the disposal of one vessel during the quarter.

Balance sheet: $250m cash, up $48m, $405m of borrowings, down $20m. Deleveraging continues.

Personal Views

• I was expecting a much higher utilization rate (similar if not better than MPM) due to Nam Cheong’s younger fleet and Middle East tensions.

• Don’t see much commentary on day rates or future outlook. May signal that the spot rates for oil and gas OSVs in SE Asia are not as strong as some of the other regions.

• I don’t think many vessels were deployed in the Middle East, gross margins declining from 48% to 42% is quite dramatic.

MPM seems to be a much better offshore play, given its exposure to offshore O&G and wind. The recurring ship repair business has sticky customers. I may be wrong on Nam Cheong’s fleet attractiveness.

Thank You For Reading!

Hope you enjoyed the article. Please share, comment, subscribe, or pledge your support by donating to this page if you find it interesting!

Feel free to leave a comment on X as well! Capytal Management

Before we dive in: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

Portfolio Positioning

Read more

Let’s dive in!

Before we dive in: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

2025 Results

To analyze Harbin Electric, I like to focus on a few key metrics (in RMB)

• Revenue

• Gross margins

• New orders

• Other key financial items

Revenue: $46b, +19% yoy

• Coal: $18b, +16%

• Hydropower: $4b, -8%

• Nuclear: $4b, +4%

• Clean and efficient industrial systems: $5b, +2%

• EPC and trading: $8b, +101%

• Modern manufacturing and service: $3b, +89%

Gross profit: $6b, 14% margins (up 2pct)

• Coal: $3b, 18% margins (up 6pct)

• Hydropower: $700m, 18% margins (up 4pct)

• Nuclear: $747m, 17% margins (down 14pct)

• Clean and efficient industrial systems: $315m, 7% margins (up 3pct)

• EPC and trading: $-17m, -0.2% margins (down 4pct)

• Modern manufacturing and service: $1b, 31% margins (down 4pct)

New order: $65b, +14% yoy

• Coal: $24b, +39%

• Hydropower: $9b, -2%

• Nuclear: $9b, +82%

• Clean and efficient industrial systems: $3b, -37%

• EPC and trading: $11b, -4%

• Modern manufacturing and service: $6b, +27%

Other key financial items

• Net cash: $13b

• Contract liability: $34b, +28%

• OCF: $5b, up from -$240m

• Provision for bad debt: $3.6b, down from $3.9b. Harbin previously had issues collecting receivables from overseas clients. Reduced bad debt is a good sign.

• Earnings: $2.6b, +54% from improved margins and large fixed-cost base.

My Views

• Coal remains the primary revenue driver, with earlier equipment orders now converting into deliveries and recognized sales.

• EPC is driven by the large Saudi Arabian power plant project. Harbin does not make money from EPCs, so we should not cheer for massive growth.

• Gross margins are improving in general due to higher-value orders in prior years and better cost management. Nuclear is a weak spot due to minimal newbuild (only 1.65 GW of nuclear power capacity added in China in 2025). Given strong nuclear orders growth and a good start to 2026, I expect profitability to improve.

• Orders remain strong in coal and nuclear, but hydropower is relatively stagnant. I am expecting this to improve in the coming years. Two major initiatives (the pumped storage buildout and the Yarlung Tsangpo Hydropower Project) should generate robust demand for Harbin’s hydropower equipment.

  • Pumped storage: State Grid forecasts 30GW of new construction and 30GW of new installed capacity in the next 5 years. As a reference, there were 66 GW of pumped storage capacity at the end of 2025.

  • Yarlung Tsangpo Hydropower Project: 60GW of hydropower to be added in Tibet, the largest hydropower project in the world, and three times the size of the Three Gorges Project. Turbine procurement should begin in 2027-2028.

• Contract liability: The reported new orders are likely real, since contract liabilities grew twice as fast as new orders (companies are paying deposits upfront and/or milestone payments as projects progress).

Power Plant Equipment Industry

Harbin Electric operates in the power plant equipment sector. It is vital to know the pace and the type of power China is installing.

China 2025 new power capacity installed: 550GW

• Wind and solar: 440GW

• Natural gas: 20GW

• Pumped storage: 7.5GW

• Coal: 78GW, the highest in a decade

• Nuclear: 1.65GW

China 2025 total capacity installed: 3,890GW

• Thermal: 1,539GW (coal: 1,260GW)

• Solar: 1,202GW

• Wind: 640GW

• Hydro: 448GW (pumped storage: 66GW)

• Nuclear: 62.5GW

2026E new capacity: 400GW, including 300GW of renewables

• Coal: drop to 31% of installed capacity, implying around 1,333GW and 75GW of new additions

2026 Updates

#1 Power Capacity Additions

Every month, China’s National Energy Administration summarizes key statistics regarding the country’s power system. We have two months of data so far (January and February), and the data is encouraging.

Thermal: 20GW added, vs 6.6GW in 2M25. In monthly updates, coal and gas-fired thermal power capacities are reported together. It is estimated that coal additions were 15GW, +178% year-over-year, and gas was 5GW, +700% year-over-year.

Hydropower: 1.2GW added, down 0.7GW year-over-year.

Nuclear: 1.2GW added, compared to 0 in the same period last year. Remember, only 1.65GW of nuclear capacity was added in 2025, so this is a strong start.

#2 Company News

Frequent visits: Harbin Electric has had a busy start to the year, visiting numerous companies. Here is a list of companies they met:

• CHN Energy Investment Group

• State Grid (State Grid Corporation of China)

• China Huaneng (China Huaneng Group)

• China Datang (China Datang Corporation)

• China Huadian (China Huadian Corporation)

• SPIC (State Power Investment Corporation)

• Three Gorges Group (China Three Gorges Corporation)

• Sinopec (China Petrochemical Corporation)

• CNOOC (China National Offshore Oil Corporation)

• SDIC Group (State Development and Investment Group)

• China Resources Group (CR Group)

• China Coal (China National Coal Group)

• PowerChina (Power Construction Corporation of China)

• CGN (China General Nuclear Power Group)

• Guangdong Energy Group

New investment: On April 2, China Yajiang Group, the company created for the Yarlung Tsangpo Hydropower Project, formed a new entity called Yajiang Hydropower Technology and Innovation. Both Harbin Electric and Dongfang Electric own 2% each. They are the only two domestic hydropower turbine supplier capable of taking on a project of this scale. This signals the project is ongoing.

SASAC capital injection: On April 9, it was reported that Harbin Electric’s registered capital increased by $5b, from $2b to $7b. This is notable, since Harbin Electric makes around $2.5b of net profits a year. The capital injection came from the ultimate parent, 70% owned by the State-owned Assets Supervision and Administration Commission of the State Council (SASAC). Unless Harbin Electric (the listed company) is in major financial trouble (which does not seem to be the case), the capital injection is likely for financing major projects in the future.

Thank You For Reading!

Hope you enjoyed the article. Please share, comment, subscribe, or pledge your support by donating to this page if you find it interesting!

Feel free to leave a comment on X as well!

Let’s dive in!

Before we dive in: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

Beng Kuang Marine Ltd

Beng Kuang Marine (BKM) was founded in the 1990s as a subcontractor providing corrosion prevention services to shipyards. It has steadily expanded into the marine and offshore energy industry over the last 30 years. The company today has two major divisions: Infrastructure Engineering (IE) and Corrosion Prevention (CP).

Infrastructure Engineering (80% of FY25 Revenue)

• Repairs and maintenance of floating production platforms

• Marine fabrications (repair metal components and structures) for vessels

• Production and supply of customised pedestal cranes and deck equipment

Corrosion Prevention (20% of FY25 Revenue)

• Surface preparation

• Protective coatings

Industry Overview

Beng Kuang mainly performs services for FPSOs and FSOs. What are they?

Floating, Production, Storage, Offloading (FPSO) Units: Offshore vessels that receive hydrocarbons from subsea wells, process them onboard, and store the output before offloading to shuttle tankers. Their versatility stems from ultra-deepwater operating capabilities without the multi-billion-dollar subsea pipeline investments needed for onshore transport. These vessels can be redeployed to different regions upon project completion, unlike fixed pipelines. FSOs are similar but lack production/processing capabilities.

Fleet age: According to Lim and Tan, there are 217 FPSOs globally, with over 70% of them >10 years old and over half of them >30 years old. New supply is limited, with 3 units expected to come into the market each year from 2026 to 2028. Newbuild is also very expensive (US$1 billion to US$3 billion per unit).

• If we include 118 FSOs, there are 335 FPSOs and FSOs operating today (BKM’s total addressable market). BKM serves 19 FPSOs, capturing only 6% of the market today.

• Clarksons Research forecasts 66 FPSO and FSO newbuilds entering service between 2022 and 2027, increasing the total fleet to over 400 assets. Every newbuild today will eventually become a potential BKM client within 5-10 years.

Demand: The global FPSO fleet is primarily concentrated in Southeast Asia, South America (notably Brazil and Guyana), and West Africa. BKM has a presence in all three regions. These locations exhibit robust demand driven by substantial deepwater reserves and limited pipeline infrastructure. Deepwater oil and gas investment is expected to remain resilient, with a potential upturn around 2027 as earlier deepwater discoveries transition into drilling and production. Meanwhile, fading peak-oil concerns, aided by supply disruptions related to the Iran conflict, support continued capital deployment in offshore projects and FPSO operations.

• Southeast Asia: Demand remains stable amid maturing basins, with operators extending FPSO lifespans via brownfield developments and life-extension projects. BKM maintains established operations in the region, though it is not the primary growth driver amid broader global opportunities.

• Brazil: Petrobras, Brazil’s state-owned oil and gas giant, committed US$78 billion to upstream investments in its 2026-2030 Strategic Plan, with a major focus on FPSO deployments in the prolific pre-salt fields (ultra-deepwater reserves beneath thick salt layers requiring specialized large-scale FPSOs). The plan requires ~20 FPSO units from 2026 through the early 2030s, with additional units planned beyond. Only 9 are currently contracted, signaling near-term opportunities. ASOM benefits from an established relationship with Petrobras, having executed work on the FPSO Capixaba.

• Guyana: One of the fastest-growing offshore oil and gas regions globally. From zero production in 2015 to 900,000 barrels per day a decade later, driven by the massive Stabroek Block. 6 FPSOs are currently operating, with 10 expected in the early 2030s. ExxonMobil is the operator (45% interest), a company that ASOM served before and has direct engagement with.

• West Africa: Renewed interest in deepwater projects, driven by Angola.

  • Angola’s offshore oil production is expected to stabilize at around 1.1 to 1.2 million bpd, after dropping below 1 million in Jul/25. Reforms encouraging investments in mature assets and new projects aim to reverse declining output, such as exploration campaigns by TotalEnergies and ExxonMobil. The formation of Azule Energy (BP-Eni joint venture) and subsequent oilfield investments led to more FPSO tie-backs. This extends the useful lives of FPSOs, which means more repairs and maintenance opportunities.

  • Ghana and Guinea are seeing stable demand due to their maturing fields. Both regions are focusing on increasing production capacity to offset natural field depletions.

Investment Thesis

This does not seem to be a rapidly growing industry. Why do I find Beng Kuang interesting? The answer is their recent acquisition, the nature of the industry, and the CEO’s turnaround initiatives.

Acquires 49% of ASOM

On February 26, 2026, Beng Kuang announced that it would acquire a 49% equity stake in Asian Sealand Offshore and Marine (ASOM) for S$60m. ASOM is their infrastructure engineering arm. BKM already owns 51% before the acquisition. ASOM will be a wholly-owned subsidiary after the deal closes.

The S$60m will be funded by the following:

• S$20m by the issuance of 57m shares at $0.35 per share, a 13% premium at the time of the deal announcement.

• S$20m in cash.

• S$20m in contingent consideration, subject to ASOM’s FY26 and FY27 annual profit reaching at least S$15m.

This deal is attractive for several reasons:

• Purchase multiple: 8x P/E

• EPS accretion: 84% increase despite the equity dilution.

• Management alignment: ASOM management will own 20% of BKM.

• Earning quality: 100% consolidation of earnings and cash flows, without a 50% leak to minority shareholders. Potential re-rating given full ownership.

Nature of the FPSO Industry

• Aging fleet: As mentioned, the FPSO fleet is growing older. This creates demand for recurring repair, maintenance, and inspection services to extend the vessels’ operating lives. To minimize downtime, FPSOs are not brought back to shipyards when they need to be repaired. Instead, companies like BKM go out to the rig to fix it while it continues to produce.

• High newbuild costs: Strong incentives to maximize FPSO fleet utilization through extended operations, despite elevated repair and maintenance requirements.

• Opex-driven: Repair and maintenance are operating expenses, not discretionary capital expenditures. Oil majors incur them regardless of oil prices and capex budgets.

• Mandatory compliance: Regulatory bodies, the American Bureau of Shipping (ABS) and Det Norske Veritas (DNV), are the world’s leading Classification Societies for FPSOs. Their role is to verify whether the FPSO is sound, seaworthy, and safe for the crew. Without a Class Certificate from one of the two, an operator cannot get insurance and operate the vessels. Both bodies require mandatory inspections and surveys every 2-5 years.

CEO Yong Jiunn Run

• 2021 hiring: Former head of commercial banking at CIMB Singapore, spent 30 years in corporate and commercial banking.

• BKM 2.0: Strategic review in 2021 to prioritize cost minimization and deleveraging, while focusing on monetizing fixed assets and high-potential business segments.

  • Exited the capex-heavy livestock carriers business in 2022.

  • Batam shipyard: spans 328,956 sqm. Conducted a strategic review in 2021, sold 58% of the area for S$18m in 2025, booking a $13m gain.

  • Deleverage: S$25m net debt and negative earnings at year-end 2020. S$7m net cash today post ASOM acquisition.

  • Redeemed bonds and initiated dividends in 2024.

  • Margins: Gross margins improved from 12% in FY21 to 37% in FY25.

  • Revenue: Increased from $53m in FY21 to $98m in FY25, despite exiting non-core businesses.

Financials (in SGD)

Revenue: $98m, -12% due to project timing delays in Infrastructure Engineering (IE).

• IE: $79m, -14% from logistical challenges in deploying barges to support FPSO projects in West Africa, not a demand issue. Work has resumed.

• CP: $20m, -4% from lower work order volumes.

Gross margins: 37%, up 2.6pct despite a 12% decrease in revenue, due to a shift to higher-margin maintenance services.

Administrative expenses: $21m, flat yoy.

Free cash flow (FCF): S$21m, up 73% yoy despite the revenue decline.

Net cash: $37m (pre ASOM acquisition), $7m post acquisition.

Valuation

• Market cap: $143m

• Net cash: $7m

• Contingent consideration: S$20m

• Enterprise value: $156m

• 2025 FCF: $21m

• EV/FCF: 7.4x

• 2025 Earnings: $13m

• P/E: 11.4x

Reminder: 7.4x FCF and 11.4x PE are on depressed earnings due to temporary delays in contract execution. Work has already resumed. Enterprise value conservatively incorporates the S$20m contingent consideration.

Thank You For Reading!

Hope you enjoyed the article. Please share, comment, subscribe, or pledge your support by donating to this page if you find it interesting!

Feel free to leave a comment on X as well! Capytal Management

I believe no matter how the Iran War ends, it will lead to an increased focus on energy security, particularly in energy-import-dependent regions such as Asia. Marco Polo Marine is in a prime position to benefit, just like Nam Cheong. Let’s dive in!

Before we dive in: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

Quick Overview

Marco Polo Marine (MPM) is an Asian marine logistics company that engages in shipping and shipyard activities.

Shipping: Chartering of OSVs, including anchor handling tug supply (AHTS) vessels, commissioning service operation vessels (CSOV), and crew transfer vessels (CTV) in the Gulf of Thailand, Malaysia, Indonesia, and Taiwan. It also owns tugboats and barges.

• The ship chartering business comprise 65% of FY25 revenue.

• Please note MPM’s fiscal year ends in September.

Shipyard: Shipbuilding and ship maintenance, repair, outfitting, and conversion services in its shipyard in Batam, Indonesia. The shipyard occupies 34 hectares with a 650-meter seafront, capable of undertaking projects involving mid-sized and complex vessels.

• The shipyard business comprise 35% of FY25 revenue, with the majority in ship repairs (28% of FY25 revenue) and the rest in shipbuilding and others.

Vessel Types

Offshore oil and gas: MPM owns a fleet of 11 AHTS vessels serving the offshore oil and gas industry. AHTS is an OSV used to anchor offshore oil and gas platforms and tow them to drilling locations, designed to operate in harsh sea conditions. Two additional AHTS will join the fleet in 2026 to support oil and gas activities in Southeast Asia.

Offshore wind vessels: The CSOV and three CTVs:

• CSOV is a specialized vessel to support the installation, commissioning, and maintenance phase of offshore wind farms. In Sep/22, MPM announced its plan to build, own, and operate a CSOV, the MP Wind Archer, to accommodate the growing offshore wind industry. The ship was designed and built in-house in MPM’s Batam shipyard. In Apr/25, MP Wind Archer successfully commenced operations under its maiden charter with Siemens Gamesa Renewable Energy (SGRE), Siemens Energy’s offshore wind division. The vessel was deployed to the Hailong Wind Farm, one of the largest offshore wind projects in the East China Sea. After the maiden charter, the MP Wind Archer will be deployed in Taiwan under a 3-year agreement with Vestas Taiwan.

• CTV is a specialized, high-speed vessel to transport technicians and equipment to and from wind turbines and substations. In 2025, all three CTVs were contracted to Siemens Gamesa. One CTV completed its maiden deployment in South Korea, with the other two CTVs supporting the Greater Chunghwa offshore wind project in Taiwan, one of the key projects under their energy transition initiative. MPM is building a new CSOV to serve both wind and O&G customers. Construction will commence in 2026, with delivery expected in 2028.

Notable Contract Wins

• August 21, 2025: Secured a three-year Master Service Agreement with Cyan Renewables, a Singapore-based offshore wind vessel owner. Marco Polo Shipyard will provide comprehensive ship repair, maintenance, and conversion services for Cyan Renewables’ fleet of offshore wind vessels. Contract value is undisclosed.

• November 13, 2025: MPM secured a S$198m contract to construct an oceanographic research vessel (ORV) for Taiwan’s National Academy of Marine Research. The ORV will be built over 4 years in their shipyard, with design support from the renowned Norwegian firm Skipteknisk AS. The project is fully self-financed with no project-specific debt required.

• March 18, 2026: MPM secured a 15-year emergency towage and salvage services charter contract with Taiwan’s Marine Port Bureau, valued at $118m. MPM will provide dedicated services within Taiwan’s territorial waters and the surrounding offshore wind farm areas. MPM is expanding beyond CSOV and CTV chartering in the attractive offshore industry.

Financials and Valuations

Although FY25 was a relatively lackluster year for revenue growth, Q1/FY26 is seeing acceleration driven by new vessel deployment and a recovery in utilization rates. The key financial metrics below are for both FY25 and Q1/FY26, in SGD.

Revenue

• FY25: $123m, flat yoy.

  • Ship chartering: $80m, +12% driven by fleet expansion.

  • Ship repair: $34m, +14% driven by higher repair contract values.

  • Shipbuilding and others: $7m, -64% driven by a reduction in the number of shipbuilding projects for 3rd party customers.

• Q1/FY26: $33m, +27%.

  • Ship chartering: $23m, +53% driven by fleet expansion and improved utilization (76%, up from 71%, reflecting sustained vessel demand).

  • Ship repair + shipbuilding (no separate reporting): $10m, -9% driven by reduced shipbuilding activities, partially offset by ship repair projects.

Gross profits

• FY25: $54m, +12%.

  • Margins: up from 39% to 44%, driven by a greater portion of revenue from the higher-margin ship chartering segment.

• Q1/FY26: $14m, +32%.

  • Margins: up from 41% to 43%, driven by a greater portion of revenue from the higher-margin ship chartering segment.

EBITDA

• FY25: $50m, +17%, driven by improved gross margins and operating leverage despite flat revenue growth.

• Q1/FY26: Undisclosed, but should be higher than +32% growth in gross profits.

KPI Trends

• Charter Rates: maintained steady growth, driven by CSOV and CTV deployment.

• Utilization: higher year-over-year, has been averaging 70-80% in the last few quarters.

Valuation

• Market cap: $535m

• Net cash: $31m (post Jan/26 $21m equity raise).

• Enterprise value: $504m

• EV/EBITDA: 10x (FY25), or 7.7x (FY26E) if 30% growth is maintained throughout the year.

Why is it Interesting?

Tightening OSV market: Due to the decade-long offshore oil and gas downcycle, the industry saw few new OSVs; the new vessel orderbook is minimal, and older vessels began to retire. This leads to an aging fleet and a lack of replacement vessels. It is estimated that 40% of Asia’s OSV fleet could exit over the next 5 years. According to Tidewater, Southeast Asia is seeing increased tendering activity for OSVs in 2026, after a relatively soft 2025.

Renewables angle: Both Taiwan and South Korea are adding offshore wind capacity.

• Taiwan currently has 4.5GW, and the government wants to install 1.5GW per year. On March 27, they launched an offshore wind project auction for up to 3.6GW, with softer localization requirements that could attract foreign companies like Orsted, of which Siemens Gamesa is a supplier to.

• South Korea is aiming to increase its capacity from 12 GW to up to 20 GW by 2030. The new “Offshore Wind Power Special Act” consolidates 29 permitting procedures into a single office, aimed at reducing approval times from 71 to 31 months.

High-margin ship repair: An aging OSV fleet needs more repair and maintenance. MPM expanded its shipyard capacity by 25% in FY25 to serve the offshore and general shipping customers. I expect the ship repair business to be less cyclical than the chartering business.

Contract visibility: MPM is building vessels because of clear demand indicators, rather than speculatively with the help of leverage (industry norm pre-2016 collapse). Recent contracts solidify MPM’s ability to compete and win the renewables market and government support. The lack of competition also means MPM should be able to secure a share of many future contracts in the rapidly growing offshore wind sector. I like participating in the growth through vessel chartering, rather than the money-losing project construction angle.

Equity Issuance and Shareholder Transactions

March 4, 2026: Raised $21m through an equity private placement, a 4% share dilution. The placement was well-received by institutional investors. The net proceeds will be used to fund expansion in offshore wind and enhance its fleet.

• Notable investors: Ginko-AGT Global Growth Fund (already an existing shareholder), Value Partners, Lion Global Investors.

• I consider this a slight negative. I hope they won’t issue more shares in the future.

January to March 2026: Halom Investment, controlled by Michael Kum (47th on Forbes Singapore list with an estimated US$1.2b net worth), acquired 5.8% of MPM by buying from Apricot Capital and Penguin International, two existing substantial shareholders. He was in the OSV industry before; he founded Miclyn Offshore in 2004, grew a fleet of 16 OSVs to 49 by 2008, and sold the fleet to Macquarie Group in the same year. Sean Lee, the CEO of MPM, acknowledged the investment and looks forward to a collaborative relationship.

Shareholder Base

Lee Family (21% ownership): MPM was founded by Lee Wan Tang, a Chinese-Indonesian businessman with diversified interests in real estate, mining, and shipping. His son, Sean Lee, is the CEO and has led the company since the early 2000s. He established the shipyard operations, took the business public, and entered the offshore oil and gas chartering business. He also secured backing from investors to restructure MPM’s debt after the 2016 oil price crash and pivoted into offshore wind farm services. His sister, Liely Lee, is a former CFO and a current non-executive director.

Apricot (12% ownership): Private investment vehicle of the family of Mr David Teo, a founding member of the Super Group Ltd. Led the $60m equity financing during MPM’s restructuring.

Halom Investment (6% ownership): Recent shareholder, controlled by Michael Kum.

Penguin International (5% ownership): Strategic investor and board oversight during MPM’s restructuring. Integrated designer, owner, and operator of high-speed aluminium vessels.

Thank You For Reading!

Hope you enjoyed the article. Please share, comment, subscribe, or pledge your support by donating to this page if you find it interesting!

Feel free to leave a comment on X as well!

Before we dive in: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

War and Impact So Far

There are likely thousands (if not millions) of articles written about the war and its impact on the energy markets. I will try to keep it short.

Oil and gas: 20% of the global oil and LNG supply is shipped through the Strait of Hormuz. Oil-producing countries have cut production, while Qatar, one of the largest LNG exporters globally, declared force majeure (unable to meet obligations due to unforeseen events) through May. Oil and gas prices jumped, of course.

• Immediate effects: Fossil fuel-importing countries scrambling for fuel. Asian countries are paying higher prices than the EU to secure fuel, causing many LNG tankers originally heading to Europe to reroute. Refined product and petrochemical product prices skyrocketed. Some countries are switching back to coal power, while many others have enacted numerous measures to limit fuel usage.

Potential Second-Order Effects

China AIDC delay: As I mentioned in my paid subscriber chat (for people who are interested in my portfolio moves), I became very cautious and sold the China AIDC diesel engine block. I worry that the war’s impact on fuel supply and pricing could potentially delay (not end) the China AIDC buildout as the country prioritizes social stability, such as transportation, farming, and industrial activities, over the data center buildout in the near term. China’s power system is perhaps the most resilient in the world in dealing with the energy crisis, but it is still a major oil importer (11 million barrels a day, the world’s biggest). In fact, Beijing ordered state-owned refineries to cease all exports of refined products to ensure domestic supply. I don’t mind revisiting the theme once things stabilize (or even missing out on the recovery) if I am too slow to react.

• Despite the export ban, it was reported that China exported diesel to Southeast Asian countries to alleviate local fuel shortages and demonstrate its diplomatic leverage.

Europe: Natural gas storage is the lowest since 2022 due to colder-than-normal weather. Although 80% of Qatar’s LNG exports go to Asia, the cutoff is forcing Asian buyers to bid for LNG from suppliers elsewhere, reducing gas availability for the EU (and others). It has been mentioned that several LNG tankers have been diverted from the EU to Asia due to higher selling prices. Russian gas doesn’t seem to be returning, with the EU rolling out a ban on spot imports of gas and LNG over the next two years, beginning on March 18, 2026.

EU countries are asked to start refilling earlier for next winter, with some suggesting a lower storage target 90% to 80%. Yet, the EU is caught in a dilemma due to TTF’s backwardation structure. TTF is the benchmark for European natural gas prices.

• In simple terms, backwardation means that TTF prices today are higher than the future. Market participants expect TTF prices to be lower during the upcoming winter (the expected future price, also known as the forward price) than today, betting the war would end, and supply would resume then. This pricing structure disincentivizes early refilling (buying today) because countries will be paying a higher price to fill up and store gas compared to what they expect the same gas to be sold for in the upcoming winter. If the war ends, delaying the refill is the correct policy.

• If the war drags on, there will be less time for refilling, and the lack of gas storage could increase winter prices. The Netherlands, the heart of Europe’s natural gas trading, is sitting on 6% storage, a record low in a decade. EU’s gas storage is 28%, unusually low for this time of the year.

Energy security back into focus: The Russia-Ukraine War in 2022 ultimately resulted in a trade-flow issue. Oil and gas were still produced, just for different buyers. Prices spiked for a while, but supplies were generally available. The current war created a global physical chokepoint. In 2022, Asia was largely insulated from the European gas crisis, but not in 2026. I believe the current crisis may bring energy security back into focus, particularly in emerging economies such as Southeast Asia. This should lead to more investments in all types of power, such as offshore oil and gas (lower breakeven prices compared to shale, longer-term projects), renewables, battery storage, and nuclear (including restarts).

AI supply chain risk: Japan, Korea, and Taiwan are pivotal to AI. Japan is a critical supplier (sometimes the only supplier) of specialized semiconductor materials and equipment, South Korea dominates the HBM market, and Taiwan manufacturers 90% of global AI servers (TSMC alone consumes 10% of Taiwan’s electricity). They are also extremely reliant on energy imports. Japan and Korea generate 60% of their electricity from coal and gas. If we include oil, 80% of their overall energy demand is from fossil fuels, almost entirely imported. For Taiwan, 80% of electricity is generated by coal and gas, nearly 100% imported. I consider AI supply chain disruption due to energy constraints to be a remote possibility, a tail risk, rather than a high probability event, but it is something I would be mindful of.

What Is Overlooked: Credit Markets

With all the news coverage about the war, I think a relatively quiet corner of the market is slowly unraveling: private credit (PC). I came across the Steve Eisman podcast, and he did a brilliant job summarizing the current private credit landscape.

• US private credit market size: Up from $300 billion a decade ago to $1.8 trillion.

• Direct lending: 50-60% of the US PC market is direct lending. 80% of direct lending is to fund private equity (PE) buyouts.

• Circular lending: AI circular financing is often talked about, but private credit/private equity circular financing is not mentioned as much. The PC lenders, the business development companies (BDCs), are mostly owned by PE companies. This means when PE companies conduct buyouts, they would ask PC funds for funding, essentially partially funding their own deals.

• Software exposure: 25% of US direct lending loans are lent to software companies. PE companies were buying software companies aggressively during the 2018-2022 period at high valuations, using higher leverage and lower interest rates. Publicly traded software companies have declined significantly due to AI disruption fears (or Claude disruption fears). However, because of the private nature of private equity, public market valuation changes of comparable companies and their loan values are not immediately reflected in some of the PE funds’ software holdings. As the maturity wall begins in 2027 and beyond, the ability to either 1) sell the private software companies through IPO or to another buyer, or 2) refinance software loans at higher interest rates (and potentially lower valuations) will be tested.

• Redemption requests: Investors are starting to notice the problems and want their money back. Due to the illiquid nature of private credit loans, private credit funds are unable to meet the outsized investor redemption requests and need to set redemption caps. Many large funds’ redemption requests are twice the amount of the cap.

Second-order effect: My worry is the spillover effect to AI capex. Morgan Stanley estimated that $2.9 trillion will be spent on global AI capex from 2025 to 2028. Only $1.4 trillion will be funded by hyperscaler cash flows, with the remaining $1.5 trillion funded by a combination of private credit, corporate debt, asset-backed financing, and private equity. A shortfall in private credit/private equity funding could potentially slow the data center buildout.

I Dislike Doom and Gloom

Before we continue, I would like to make myself clear. I do not like being pessimistic. I am only pointing out risk factors that market participants should be mindful of. Last night, there appeared to be light at the end of the tunnel regarding the war. I can think of several companies to buy/re-enter if ceasefire rumors become true.

Portfolio Positioning

Read more

Let’s dive in!

Before we dive in: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

2025 Achievements

To say that TNZ made substantial progress in 2025 would be a major understatement.

NOBV: acquisition closed in May/25, with a cash consideration of $9 million (dollar amounts in Canadian dollars unless otherwise stated). I repeat: Mr. Marino and crew acquired NOBV, an entity with nearly $1 billion of (likely understated) 2P reserves, for $9 million.

GEMS: acquired non-operated interest from Hansa Hydrocarbons for $339 million ($323 million in cash, $17 million in TNZ common shares, a minor 3% dilution). This asset contributes to near-term organic growth and substantial inventory for long-term development and exploration.

For this article, I will focus on the TTF gas assets since they comprise 90% of TNZ’s production and 92% of 2026E revenue.

NOBV (now TEN, Tenaz Energy Netherlands)

Multi-year organic growth: began drilling and workover activities after operations were handed over, without any health, safety, or environmental incidents. TNZ expects 3-4 gross wells in 2026.

• Workover: restored 13 MMcf/d (million cubic feet per day), or ~2,200 boe/d (barrels of oil equivalent per day), of gross production from previously shut-in or intermittent wells at the K15-K platform (45.6% working interest). Workover activities will continue in 2026.

• Exploratory discovery: first well at K07-FB-103 (45.6% working interest), drilled on budget and on schedule for EUR43 million ($62 million). The initial well test had a peak gross production rate of 23 MMcf/d (~4,000 boe/d) with some condensates as well. The well is expected to be online in Q2/26, although at a lower production rate than the test rate.

  • Perhaps more importantly, this well was included in the 2025 prospective resource report rather than in the year-end reserves report because the well was tested in 2026. As TNZ said, this speaks to the quality of the projects available in the contingent and prospective resource portfolio. We should look forward to more “speculative” resource positions converted into reserves and eventually cash flows if this level of execution continues.

• Development: another well, K17-FA, could come online in Q4/26 if the data is positive. This well offsets an existing producer.

GEMS (Gateway to the Ems)

Multi-year organic growth: near-term high rate of production plus multi-year drilling activities. Expects 4 gross wells (1.4 net to TNZ) in 2026.

• Currently producing well: began production in Mar/25 at a rate of 75 MMcf/d (25 MMcf/d net to Tenaz, or ~4,300 boe/d) from a single well (the highest production rate well in the Netherlands), able to maintain similar rates today. Powered by a wind farm called Riffgat in German waters, making GEMS one of the lowest-emission oil and gas projects in the world.

• N05-A-03 development well: gas deposits seem to be at a shallower depth (“higher on structure”) with more gas than expected (“modestly increased gross thickness of the gas column”). Production expected to start in Q2/26.

• Other wells: expects extension and exploratory drilling in 2H26.

Brief update on the L10 license: This Eni-operated asset will test a well in the coming weeks before being tied into the platform for production. Tenaz owns a 21% working interest.

The Numbers

Before we continue, I would like to remind readers that the metrics only reflect post-acquisition. Q4/25 is the most up-to-date since it is the first quarter that includes both TEN and GEMS.

• Production: 15,556 boe/d in Q4/25, 9,609 boe/d in FY25 (257% higher than FY24). Management guides 19,500 to 22,500 boe/d in FY26, with an exit rate of 27,000 boe/d.

• Funds flow from operations (FFO): $62 million in Q4/25, $120 million in FY25 (nearly 5x higher than 2025).

• Net debt: $345 million. 1.4x net debt / FFO ratio, or 1x 2026E FFO at current commodity strip.

• Capex budget for 2026: $250 to $275 million

Other Interesting Numbers

Gain on acquisition: $192 million, resulting from the NOBV acquisition. What happened here was the value of the assets acquired (mainly the natural gas reserves), net of liabilities, exceeded the consideration to close the acquisition. The gain is an accounting entry to reflect the difference. This is driven by the following:

• More reserves to be recovered from increased investments (increased asset value).

• Longer useful lives of key infrastructure and gas fields, reducing the value of decommissioning liabilities (the timing of decommissioning gas fields is extended, resulting in lower liabilities).

Contingent earn-out liabilities: reduced from $87 million upon closing in May/25 to $20 million. As part of the NOBV deal, future payments to the original owners (Shell and Exxon) are dependent on NOBV’s free cash flows (“FCF”) from 2025 to 2027. Tenaz plans to accelerate capital expenditures during this period, which will reduce near-term free cash flows available for distribution to Shell and Exxon. Therefore, this estimated liability account was revised lower.

• This is, of course, a rational use of capital. By spending today, TNZ’s FCF and production in future years will be substantial and attributable mostly to TNZ shareholders, rather than paid out to Shell and Exxon.

Ocean Bottom Node advanced payment: $22 million, accelerated final payment for the seismic survey to be used for future drilling activity. Avoids interest and premium charges if paid later. Excited to see how this translates into reserves and cash flows in the future, since Marino spoke highly of this data in many of his industry conference appearances.

Reserves

Probably the most important report for any oil and gas company, and Tenaz did not disappoint.

• PDP: 32.3 million boe (mmboe), +839% from 3.8mmboe, 913% reserve replacement ratio (“RRR”), 6-year reserve life. $620 million NPV.

• 1P: 57.4mmboe, +499% from 10.2mmboe, 1,471% RRR, 10-year reserve life. $1.07 billion NPV.

• 2P: 91.7mmboe, +427% from 16.6mmboe, 2,358% RRR, 16-year reserve life. $1.68 billion NPV.

• Recycle ratio: 2.8-3.4x for 1P and 2P.

Unless similar deals are made in 2026, I highly doubt these growth rates will be repeated. Nonetheless, the sheer amount of reserves sets the foundation for TNZ for the coming years.

I tend to ignore speculative metrics such as contingent and prospective resources for valuation purposes. However, the success of TEN’s exploration drilling suggests some of the numbers below can eventually translate into reserves and cash flows.

• Contingent Resources (15 fields): 24.1, 38.0, 53.4mmboe for 1C, 2C, 3C. Risked best estimate of 24.1mmboe. 11 out of 15 fields were economically valued at $150 million NPV. The term “Risked” means on a probability-weighted basis based on estimated geological success.

• Prospective volumes (96 fields): 128.3, 329.2, 603.5mmboe for low, mean, and high estimates. Risked mean estimate of 101.1mmboe based on chance of discovery and development on a prospect-by-prospect basis. 23 of the 96 fields were economically valued at $734 million NPV.

Natural Gas Market Commentary from the President

With the 24/7 news cycle in today’s world, information overload is unavoidable. Here are some key points from Mr. Marino regarding the natural gas market (a highly fluid situation that is subject to change, of course).

2025 Recap

• The EU had high seasonal demand and heavy reliance on LNG, especially from the US Gulf Coast. Marino thinks this source of supply may not be necessarily reliable, depending on future geopolitical conditions. EU’s storage saw some of the largest withdrawal rates on record during winter; current storage is the lowest since 2022.

• China’s LNG demand growth slowed due to Russian pipeline gas and domestic production.

2026 Conditions and Outlook

• Middle East conflict: Qatar’s LNG production suspension reduced 20% of global LNG supply. TTF spot and near-term prices surged to reflect increased competition for available LNG cargos. The EU relies on spot and short-term LNG purchases as opposed to long-term contracts.

• EU storage: 90% storage target was not met in 2025 by some member countries. If the target is enforced for the 2026 winter season, it provides support for this summer’s gas prices. If there is no mandate, storage going into next winter may be inadequate, because the current backwardated term structure of TTF prices does not incentivize storage.

• Russian gas phase-out: EU elected to cease Russian gas imports in 2027. This phase-out might occur earlier, based on Putin’s recent commentary.

Tenaz’s Value Proposition

• Domestic reserves: safe, clean, reliable, sustainable.

• Economic benefits: employment, investments, tax contributions, and lower prices relative to LNG imports.

Corporate Updates

• Hired Kyle Preston as Vice President of Investor Relations. Most recently served as Vice President of Investor Relations at Vermilion Energy (that’s another high-level employee moving from Vermilion to Tenaz).

• Reserve-based loan: commitment increased from $115 million to $150 million. Added DNB Bank ASA (Norway’s largest bank) to the lending syndicate. The loan remains undrawn.

• Share repurchases: 336,000 shares repurchased at a weighted average price of $18.25 per share. Repurchased 43,600 shares at $35.84 per share in 2026 thus far. Unfortunately, share repurchases will be less accretive due to a higher share price, but I am not complaining.

• Hedging: 54% of TTF exposure and 43% of 2026 revenue are currently protected. They will “layer in additional revenue protection for 2026 and beyond to protect payout of capital projects and manage cash flow volatility”. This is a good opportunity to lock in cash flows at these benchmark prices.

Thank You For Reading!

To conclude, I will include a section from the press release:

To free readers: Hope you enjoyed the article! Please share, comment, subscribe, or pledge your support if you find it interesting! Feel free to leave a comment on X as well!

To my paid subscribers: don’t leave just yet! Please continue to scroll below for my personal views on this largest holding in my portfolio.

Personal Views

Read more

I don’t think I should spend too much time addressing market developments since the last portfolio update. All I will say is it is a deliberate choice of mine not to invest in private equity/private credit names, BDCs, and software simply because they are too complex for me to understand. I believe the easy financing environment and lack of regulations prolonged imprudent underwriting activities. Software’s recurring ARR model is bound for disruption due to complacent management and infinite price hikes without corresponding increases in incremental product value. Both of these industries were trading at expensive multiples to begin with and are interconnected, making the multiple compression justified. I might be wrong about the eventual outcome, but I won’t feel like I missed the opportunity by not investing today, even if they eventually recover.

Before we dive in: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

Portfolio Update

Read more

Before we dive in: This is not investment advice. Please do your own research. And if you would like to comment, please be respectful to capybaras; they deserve it.

Yuchai Marine IPO Prospectus

I have summarized Yuchai Marine’s IPO prospectus in the article below. The key takeaway is that China’s power generator engine market is accelerating, driven by surging demand for data center computing power. As more advanced chips are deployed, power density per rack increases. More AI servers are expected to be delivered to data centers. Provinces in China have specifically prepared for clusters ranging from 10,000 to 1 million GPUs.

Side note: CYD’s earnings seem to have missed estimates due to higher R&D expenses and lower government grant contributions to earnings. The share price might have run ahead of itself pre-earnings, but I am not worried at the moment. There is also a rumor suggesting CYD secured the majority of ByteDance’s latest diesel engine tender order value.

Company Updates

ByteDance: ByteDance is developing an AI inference chip, in manufacturing talks with Samsung, sources say. The chip project, codenamed SeedChip, is part of ByteDance’s broader push to channel resources into AI development, from chips to large language models, betting the technology will transform its business portfolio spanning short video, e-commerce, and enterprise cloud services. This news comes after the company released the viral Seedance 2.0.

Alibaba: Alibaba Cloud is reportedly considering increasing its planned three-year capex investment in AI infrastructure to $69.05 billion (480 billion yuan). This would be an increase on the previously promised $52.4 billion (380 billion yuan), stated by the company in early 2025. During the company’s November earnings call, Alibaba revealed that in the last 12 months it had spent around US$16 billion on AI-related purchases, including data centers, suggesting that we can expect capex to continue to ramp up over the next couple of years - even more so should the company expand its plans.

AXTI Earnings Call: AXTI produces compound semiconductor substrates for data center applications. “In China, the data center build-out is early in its ramp. But we are seeing rapid growth as China moves to accelerate its data center expansion and AI capabilities. Our revenue related to the data center market in China is expected to grow by more than 60% in Q1 over Q4, highlighting both increased investment in these Tier 1 data centers as well as the strong desire for Chinese domestic suppliers to secure local stores at every level of the AI infrastructure supply chain.” “Irrespective of permits, we mentioned we are seeing growth in this business in China as well. And looking at it quarter-to-quarter, Q4 was probably double revenue in China than Q1 in 2025. And we would expect -- we’re looking at potentially doubling again through 2026.”

GLM-5: (Allegedly) trained entirely on Huawei Ascend chips, not needing US AI chips and hardware. Deep domestic chip integration for inference (Huawei Ascend, Moore Threads, Hygon, Cambricon, Kunlunxin, MetaX, and Enflame). The model was so popular that the company had to apologize for operational difficulties. They are seeing external partners to expand computing capacity.

GDS: monetized part of its DayOne stake, with the proceeds used to reinvest in domestic data center opportunities.

2026 Outlook

In the previous article, Inner Mongolia and Shanxi were mentioned as prime hubs for China’s AIDC buildout ambitions. Lucky for us, each provincial government releases an annual Report on the Work (工作報告), offering insight into where AIDC stands among its policy priorities. Since Shanxi’s current scale is relatively limited compared to other national hubs, it does not explicitly mention data center capacity targets. Fortunately, many major provinces and cities released their 2026 plans.

Starting with China: 1,590,000 PFLOPS (computing unit) as of 2025. Expects at least 24% CAGR for the next 3 years according to a Feb/25 forecast. This growth rate is likely understated given China’s AI development in 2025.

Provinces

Inner Mongolia (15% of China’s computing capacity): The top intelligent computing province in China; 85% of the Horinger County AIDC cluster is powered by green energy. Capacity reached 220,000 PFLOPS in 2025. Aiming for 48% growth yoy

Guizhou (10%): Accelerating the buildout of computing infrastructure. Capacity reached 160,000 in 2025. Aiming for 20% growth yoy

Ningxia (8%): 7.4 billion yuan of AIDC investments in 2025, with 10.3 billion yuan planned for 2026. Computing capacity reached 130,000 PFLOPS in 2025. Aiming for 100% growth yoy (再翻番)

Shanghai (8%): 120,000 PFLOPS in 2025, expects to reach 156,000 by 2026 and 236,000 by 2028. Strong demand for low-latency compute in finance and manufacturing, key compute provider for the broader Yangtze River Delta region. Already collaborating with the three SOE telcos, Alibaba, and Sensetime.

Gansu (7%): Capacity reached 114,000 PFLOPS in 2025. Aiming for 32% growth yoy. First domestic 100,000 chip training cluster in Qingyang, a 36.6 billion yuan project. First GW-scale data center cluster in China powered by green energy. In fact, this project is 90% powered by renewables.

Beijing (4%): 60,000 PFLOPS in 2025, expects to reach 200,000 PFLOPs by 2027 (182% 2-year CAGR). This makes sense since Beijing is home to China’s advanced AI research universities and talents, 50% of AI unicorn startups, and the headquarters of top AI model and chips companies.

Beijing will build super nodes and industry nodes in new energy‑rich areas and mountainous areas, including Yanqing–Mentougou–Fangshan, Changping–Haidian, and Yizhuang–Tongzhou within its administrative area. Beijing’s computing needs will also be supported by the Beijing-Tianjin-Hebei Region (specifically mentioned Hebei Zhangjiakou and Langfang, and Tianjin’s Wuqing). Beijing will accelerate the development of West Compute in Inner Mongolia, Ningxia, and Qinghai.

Portfolio Positioning

I have decided to concentrate my investments on the diesel engine providers (CYD, Chongqing M&E, and Weichai Power) and sell everything else. Cloud, chips, land bidding, and data center construction all look too competitive for me to determine the winner(s). Not only are the diesel engine manufacturers more modestly valued, but they also operate in an oligopolistic industry without price wars. If you have other stock ideas, please kindly comment below.

There will be another portfolio update soon!

Thank You For Reading!

Hope you enjoyed the article. Please share, comment, subscribe, or pledge your support by donating to this page if you find it interesting!

Feel free to leave a comment on X as well! @capytalmgmt