1. What the Company Does

MTTSL is an investment holding company that, through its subsidiaries, operates as the largest domestic container liner shipping company in Malaysia (41-46% market share by cabotage volume).

Core Businesses

• Container liner shipping (domestic cabotage + regional/international routes)

• Vessel chartering (time charter to third-party liner companies)

• Container storage and depot services (via 71.4%-owned ICSD)

• Dry bulk shipping (via 39%-owned joint venture Lestari Maritime)

• Liquid bulk logistics (via 60%-owned United Bulk Logistics)

• Chemical tankers (via 90%-owned MTT Kenyalang)

• Harbour tug services (via 50% JV Harbour 360)

• Shipping agency (for Evergreen Group and others)

• Containerised automotive vehicle shipping

• Proprietary shipping software (iKapal, via 55%-owned Kapal Solutions)

Headquarters: Shah Alam, Selangor
Branch Offices: 9 locations across Malaysia (Johor Bahru, Kuching, Bintulu, Sibu, Miri, Tawau, Sandakan, Kota Kinabalu)
Route Coverage: 34 ports across Malaysia, Singapore, Thailand, Brunei, Indonesia, India, and China

2. IPO Details

Use of Proceeds

Previous IPO Attempt

The company applied to SC on 6 August 2021, was approved 9 December 2021, but withdrew on 22 December 2021 citing unfavourable equity market conditions. Now proceeding due to stronger financial performance.

3. Financial Performance

GP Margin by Segment (FPE 2025)

Balance Sheet (RM’000)

Cash Flow (RM’000)

Average annual operating cash flow (FYE 2022-2024): RM402.8 million.

Tax Advantage

Profits from operation of Malaysian-registered vessels are exempt from taxation under Section 54A(1) of the Income Tax Act. This is why the effective tax rate is only 1.2-2.9% — a major structural advantage, but also a regulatory risk if the law changes.

Pro Forma Post-IPO (RM’000)

4. Debt Profile

Total Outstanding Borrowings: RM796.1M (as at 30 September 2025)

• Sukuk Wakalah rated AA3/Stable/P1 (RAM Rating Services)

• No defaults or covenant breaches on any borrowings

• RM180M Sukuk tranche maturing August 2026 (near-term refinancing need post-listing)

• Personal guarantees from certain directors to be replaced with corporate guarantees upon listing

Maturity Profile

5. Dividend Policy

Historical Dividends

6. Ownership Structure (Pre-IPO)

• All shareholders have equal voting rights (no dual-class shares)

• Founding group + family collectively hold vast majority of pre-IPO shares

• No arrangement exists that may result in a change of control post-IPO

• Evergreen Taiwan (via Peony) holds 28.65% of subsidiary ICSD — strategic alignment

Key Family Relationships

• Dato’ Seri Ong is the father of Clarice Ong (Director of Corporate Affairs)

• Dato’ Seri Ong is the brother of Ong Guat Ee (substantial shareholder via MTTC and OCTSB)

7. Management Team

Concern: Co-founders are aged 59-69. Succession planning is critical. The CFO only joined in March 2025, less than one year before the IPO.

8. Competitive Strengths (Why to Invest)

8.1 Dominant, Protected Market Position

• Largest domestic carrier: 41% market share in 2024, rising to 46% in 2025

• Dedicated cabotage capacity of 11,608 TEUs — 76.4% more than the second-largest competitor

• 46.4% of Malaysian-flagged containership fleet by unit; 59.1% by TEU capacity

• No new domestic entrants since 2011 — high barriers to entry

• Cabotage policy restricts domestic routes to Malaysian-flagged ships

8.2 Youngest, Largest Fleet

• 24 owned + 2 chartered = 26 vessels, 29,149 TEU total capacity

• Average fleet age: 7.2 years vs. competitors at 22.9-25.0 years

• 21 vessels built between 2019 and 2025

• All vessels EEXI and CII compliant (ahead of competitors on environmental regulations)

• Specialised vessel designs (Bangkok-max, Kolkata-max, Rajang-max) creating route-specific advantages

8.3 Superior Profitability

• 14 consecutive years of profitability (FYE 2011-2024)

• PAT margin of 21-25% vs. competitors at 5.7-11.5% (nearly double)

• Extremely low effective tax rate (1.2-2.9%) due to vessel income tax exemption

• Average operating cash flow of RM402.8M/year

8.4 Multiple Growth Vectors

• 4 Integrated Fulfilment Facilities (Kota Kinabalu, Pulau Indah, Kuching, Bintulu) — ~RM444M investment

• 8 newbuild vessels (1,092 to 3,300 TEU) for delivery Dec 2026 to Jun 2028

• India Subcontinent expansion (Nhava Sheva, Mundra, Kattupalli, Chennai)

• Chemical tanker diversification (2 newbuild 12,500 DWT dual-fuel methanol tankers)

• Containerised automotive shipping growing at CAGR 25.6%

8.5 Improving Operational Metrics

• Back-haul utilisation improved from 14.9% (2022) to 42.0% (2025) — nearly tripled

• Transport volumes grew 44% from FYE 2022 to FYE 2024

• Bunker fuel costs trending downward, supporting margins

8.6 Favourable Industry Dynamics

• Malaysian container volumes projected CAGR 4.6% (2026-2030)

• Shortage of ships below 4,000 TEU (order book ratio only 14.6% vs. 35.9% global average)

• IMO environmental regulations will force scrapping of older non-compliant ships

• Gemini Cooperation (Maersk + Hapag-Lloyd) at Tanjung Pelepas driving feeder volume growth

8.7 Strategic Relationships

• Shipping agent for Evergreen Group since 2010

• Evergreen Taiwan (via Peony) holds 28.65% of ICSD — aligned interests

• Combined MTTSL + Evergreen Group = ~28% of East Malaysia port throughput

8.8 Strong Asset Base

• Fleet NBV: RM1,463,904,000

• Property NBV: ~RM206,000,000

• Clean audit opinion from KPMG with no qualifications

• No contingent liabilities

• 99.2% trade receivable collection rate

• AA3/Stable Sukuk rating

8.9 Vessel Disposal Track Record

• Historically disposed vessels at 195.1% above aggregate acquisition cost

• Most recent disposal (Oct 2025): MTT Pengerang sold for 249% above NBV

9. Risk Factors and Red Flags (Why NOT to Invest)

9.1 Post-COVID Normalisation

• FYE 2022 was an exceptional year with inflated rates from COVID supply chain chaos

• Average freight rates have declined 42% from RM2,932/TEU (FYE 2022) to RM1,694/TEU (FPE 2025)

• Average vessel charter rates fell from RM127,844/day (FYE 2022) to RM73,664/day (FPE 2025)

9.2 Massive Capital Expenditure Programme

• RM693.7M planned for FYE 2026 alone (mostly vessels under construction)

• RM812.3M in capital commitments at the LPD

• 8 newbuild contracts totalling RM793.5M

• 4 IFFs costing ~RM444M

• 2 chemical tankers at RM174.1M each

• This creates significant execution and leverage risk

9.3 Revenue Concentration

• Freight income = 69.4% of revenue; vessel chartering = 21.6%

• A downturn in container shipping would severely impact the business

• Dry bulk revenue is highly volatile (RM50.5M to RM157.7M across periods)

9.4 Founder/Family Dominance

• Founding group + family control vast majority of pre-IPO shares

• Dato’ Seri Ong’s daughter holds key executive position (nepotism risk)

• Risk Management Committee chaired by a non-independent director

• Significant related party transactions (7-8% of revenue)

9.5 Cabotage Policy Dependency

• Business heavily benefits from Malaysian cabotage laws protecting domestic routes

• Any relaxation of cabotage rules would be a material risk to the competitive moat

9.6 Near-Term Debt Maturity

• RM180M Sukuk Wakalah tranche maturing August 2026 — needs refinancing shortly after listing

• Current ratio declined from 2.1x to 1.5x due to this reclassification

9.7 Dividend Uncertainty

• Historical payout only 10-13% despite 50% target

• Pre-IPO dividend of RM100M paid to existing shareholders before listing

• Heavy capex may prevent achieving the 50% payout target for years

• Sukuk Wakalah covenants may restrict dividend payments

9.8 Key Personnel Risk

• Co-founders aged 59-71; succession planning is critical

• New CFO (joined March 2025, less than 1 year before IPO)

• Deep industry relationships are personal to the founders

9.9 Foreign Currency Exposure

• ~38-40% of revenue in USD

• 10% RM strengthening vs USD would reduce PAT by RM7.4M-RM20.1M

9.10 Property and Regulatory Issues

• Multiple operating properties lack CCC/OC (Certificate of Completion and Compliance / Occupancy Certificate)

• Westports Depot lease expired at prospectus date; transition to Pulau Indah depends on obtaining CCC

• Provisional land title for Sarawak property (RM9.59M asset with title uncertainty)

• Several vessel certificates expiring in 2026, some very close to prospectus date

9.11 Other Concerns

• Previous IPO withdrawal in 2021 (market timing questions)

• 2 oldest vessels aged 29 and 31 years (increasing maintenance/compliance costs)

• Heavy reliance on foreign crew (regulatory and labour supply risk)

• Tug and barge operating on temporary licences only (expiring within 2 months of prospectus)

• Complex group structure with 40+ subsidiaries, many dormant

• ESOS could dilute shareholders by up to 5%, with post-listing options at up to 10% discount

• Limited IP portfolio (1 software copyright, 3 trademarks, no patents)

• No order book / forward visibility due to nature of business

10. Related Party Transactions

Revenue from Related Customers

Costs to Related Suppliers

• 9 related entities identified with overlapping directors/shareholders

• Board says loss of related customers would only cost ~RM18.6M PAT (7.3% of combined FYE 2024 PAT)

• All transactions at arm’s length on normal commercial terms (per Board)

• Audit Committee reviews all potential conflicts

Adviser Conflicts

• CIMB (Principal Adviser, Underwriter): Has RM267.4M in credit facilities to the Group; also Sukuk facility/security agent

• Affin Hwang (Joint Bookrunner, Underwriter): Has RM317.6M in credit facilities; holds RM94.5M of MTT Shipping’s notes

11. Fleet Summary

Current Fleet (26 vessels, 29,149 TEU total capacity)

• 24 owned vessels + 2 chartered

• Average fleet age: 7.2 years (youngest in Malaysia)

• 21 vessels built between 2019 and 2025

• 11 vessels currently chartered out at USD8,250-USD29,750/day

• 4 vessels fitted with EGCS scrubbers (only Malaysian-flagged vessels with scrubbers)

• All vessels EEXI and CII compliant

Orderbook (8 newbuilds)

• 1,092 to 3,300 TEU capacity

• Total cost: RM793.5M

• Delivery: December 2026 to June 2028

• Plan to acquire at least 4 additional vessels beyond these 8

Material Equipment (NBV as at 30 Sep 2025)

12. Key Operational Metrics

Average Freight Rates (RM per TEU)

Transport Volumes (TEUs)

Head / Back Haul Utilisation

Automotive Vehicles Transported

CAGR 25.6% (2019-2025), more than double the 9.6% total market growth.

13. Industry Outlook

Positive

• Malaysian container volumes projected CAGR 4.6% (2026-2030)

• Shortage of ships below 4,000 TEU (order book ratio only 14.6%)

• IMO environmental regulations forcing scrapping of older non-compliant vessels

• Suez Canal diversions since Dec 2023 increasing vessel demand

• National Transport Policy targets 90% general cargo in containers by 2029

• Current charter rates for 1,100-1,700 TEU ships still 161-225% above Jan 2020 levels

Negative

• Charter market rates are cyclical and declining from 2022 peaks

• Global trade tensions and tariff uncertainties

• Bunker fuel price volatility

• Terminal handling charges regulated by government (limited pricing flexibility)

14. Determining Factors of Profitability

These five factors are the primary levers that determine whether MTTSL makes more or less money. Each is backed by data from the prospectus.

14.1 Freight Rates

Freight rates are the single biggest driver of profitability. Freight income is 69.4% of total revenue (FPE 2025), and MTTSL has no long-term locked-in contracts — the prospectus states “no order book maintained due to nature of business”, meaning revenue reprices with the market.

Average Freight Rates (RM per TEU)

The blended average rate dropped 42% from FYE 2022 to FPE 2025. This directly flows into gross profit:

The same applies to vessel chartering (21.6% of revenue). Charter hire rates declined from RM127,844/day (FYE 2022) to RM76,334/day (FYE 2024) before partially recovering to RM78,734/day in FPE 2025. Charter contracts are renewable at prevailing market rates, with most charter lease receivables due within 1 year (RM258.8M vs only RM77.6M in year 2), confirming short-term repricing exposure.

Impact: When freight and charter rates were elevated in FYE 2022 (COVID-era), PAT was RM553M. As rates normalised, PAT fell to RM254M (FYE 2024) — a 54% drop despite volumes growing 44%. Rates overpower volumes.

14.2 Bunker Fuel Costs

Bunker fuel is the 2nd largest direct cost at 16.5% of total direct costs (FPE 2025), totalling RM106.8M for the 9-month period.

Average Bunker Fuel Prices (RM per Metric Ton)

Bunker fuel as % of direct costs over time

The declining fuel cost from FYE 2022 to FPE 2025 (MFO down 29%) contributed to the GP margin recovery from 27.6% (FYE 2024) to 32.7% (FPE 2025).

Mitigation — Bunker Adjustment Factor (BAF): MTTSL applies a fuel surcharge on freight rates to partially pass through fuel costs to customers. However, this is imperfect — there is a time lag in adjustments, competitive pressure may limit full pass-through, and no financial hedging (futures/swaps) is disclosed in the prospectus.

Scrubber advantage: 4 vessels are fitted with EGCS scrubbers (the only Malaysian-flagged vessels with scrubbers), allowing them to burn cheaper high-sulphur fuel oil while meeting IMO 2020 emissions requirements, giving a direct fuel cost advantage on those vessels.

Impact: Fuel is a cost MTTSL cannot fully control. When oil prices spike, margins compress until BAF catches up. When fuel prices fall (as in FPE 2025), margins expand. The FPE 2025 GP margin recovery to 32.7% was partly driven by MFO prices falling 18% from FYE 2024.

14.3 Asset Utilisation & The “Backhaul” Problem

A vessel sailing full in one direction and empty on the return is burning the same fuel, paying the same crew, and incurring the same port charges — but earning revenue on only one leg. The backhaul problem is a structural feature of MTTSL’s core cabotage route: manufactured goods flow from Peninsular Malaysia to East Malaysia, but far less cargo flows back.

Head Haul vs Back Haul Utilisation

Backhaul utilisation has nearly tripled from 14.9% (FYE 2022) to 42.0% (FPE 2025). This is one of the most significant profitability improvements in the prospectus because every additional backhaul TEU is almost pure incremental margin — the vessel is making the return trip anyway.

Volume evidence (TEUs)

Backhaul volumes grew at 49.8% CAGR while head haul was essentially flat. The revenue per TEU on backhaul (RM922 FPE 2025) is much lower than head haul (RM3,234), but the incremental cost of carrying that cargo is minimal since the vessel and crew costs are already incurred.

Fleet utilisation also matters for chartering. Of 26 total vessels, 11 are chartered out when not needed for liner operations. This dual deployment (liner when demand is high, charter when it’s not) maximises asset utilisation. However, the charter hire GP margin of 57.6% is far higher than container liner’s 28.6%, meaning fleet allocation decisions between own-use and chartering directly impact overall profitability.

Impact: The backhaul improvement from 14.9% to 42.0% materially contributed to profitability resilience even as freight rates declined. If backhaul continues improving, it provides an earnings floor. If it reverses, margins would compress further.

14.4 Port Turnaround & Congestion

Port-related costs are the largest and third-largest direct cost categories:

Port-related costs collectively represent ~38% of direct costs — more than double the fuel cost.

Terminal handling charges (THC) are particularly notable because they are regulated by the Malaysian government, meaning MTTSL has limited ability to negotiate them down. THC increased from RM95,330K (FYE 2022) to RM99,353K (FPE 2025, 9 months), growing significantly even as the company’s own freight rates declined.

MTTSL’s mitigation strategy is port frequency and network density:

• 25 port calls per week at Port Klang alone

• 8 ports in East Malaysia with at-least-weekly service

• Fixed weekly sailing schedules since 2011

• Specialised vessel designs (Bangkok-max, Kolkata-max, Rajang-max, Pengerang-max) built to maximise cargo capacity within specific port constraints (draft, beam, LOA limitations)

These purpose-built vessels reduce port turnaround time because they are optimised for the exact berth and channel constraints of their designated routes, allowing maximum cargo per call.

The IFF strategy also addresses this: The 4 planned Integrated Fulfilment Facilities (Kota Kinabalu, Pulau Indah, Kuching, Bintulu) are designed to reduce reliance on third-party depots and consolidate cargo handling, which should improve turnaround efficiency and reduce per-unit handling costs over time.

Impact: Port costs are large, partially government-controlled, and growing. MTTSL’s ability to manage this through vessel design, schedule density, and vertical integration (own depots/IFFs) is a key determinant of whether margins hold or compress further.

14.5 Regulatory Moats

The cabotage policy is the foundational reason MTTSL’s profitability is structurally higher than what a purely competitive market would allow. Their PAT margin of 21-25% versus competitors at 5.7-11.5% is not purely operational excellence — it is substantially underpinned by regulatory protection.

Cabotage policy: Malaysian law restricts domestic shipping routes (Peninsular Malaysia ↔ East Malaysia, inter-East Malaysia) to Malaysian-flagged vessels. This means:

• Foreign carriers cannot compete on MTTSL’s core domestic routes

• No new domestic entrants since 2011 — the cost of acquiring vessels + Malaysian reflagging requirements creates prohibitive barriers

• Even a 2017 relaxation allowing foreign carriers on certain routes failed to increase their market share

• MTTSL holds 46.4% of all Malaysian-flagged containerships by unit count; 59.1% by TEU capacity

Tax exemption: Profits from operation of Malaysian-registered vessels are exempt from taxation under Section 54A(1) of the Income Tax Act. This is why MTTSL’s effective tax rate is only 1.2-2.9%, compared to the standard 24% corporate rate. On FPE 2025 PBT of RM242.8M, the tax saving versus a normal rate is approximately RM51M — this alone adds roughly 5 percentage points to PAT margin.

IMO environmental regulations (EEXI/CII): All 26 MTTSL vessels are EEXI and CII compliant. The IMR report cited in the prospectus notes that these regulations will force scrapping of older non-compliant ships, further tightening vessel supply. Competitors’ fleets average 22.9-25.0 years old versus MTTSL’s 7.2 years, meaning competitors face greater regulatory obsolescence risk.

Combined regulatory effect on market position

Risk: These regulatory advantages can be removed. If the cabotage policy is relaxed further or the Section 54A tax exemption is repealed, MTTSL’s margin superiority would narrow significantly. The prospectus identifies cabotage dependency as a risk factor but provides no mitigation beyond the company’s scale and service quality advantages.

Impact: The regulatory moat is arguably the most important factor of all five. It protects pricing power (limited competition), ensures near-zero tax (Section 54A), and creates barriers that prevent new entrants from diluting market share. Without it, MTTSL would be a fundamentally different investment proposition.

14.6 Oil Price Spikes & Geopolitical Disruption — How MTTSL Is Positioned

The effect of oil price spikes on shipping stocks depends on the type of shipping company and the cause of the spike. The March 2026 Strait of Hormuz closure following US/Israel strikes on Iran provides a real-time case study:

• Crude tanker operators (e.g. Frontline, up 51% YTD) benefit massively — route disruptions force longer voyages, tightening vessel supply and spiking tanker day rates far faster than bunker costs rise. Frontline secured 7 VLCC time charters at USD76,900/day, well above historical averages.

• Container liners (e.g. Maersk, Hapag-Lloyd) get mixed results — higher bunker costs squeeze margins, with only lagged pass-through via freight surcharges.

• Dry bulk operators are generally hurt — their cargo demand (iron ore, coal, grain) is tied to global economic health, and oil spikes are stagflationary.

Post-COVID research (ScienceDirect, 2025) found the oil-freight relationship has flipped in supply-shock scenarios: higher oil prices now signal broader supply constraints that push freight rates up faster than costs, reversing the pre-2021 pattern of pure margin compression.

Where MTTSL sits in this framework

What helps MTTSL in a disruption:

• 11 of 26 vessels are chartered out — the charterer pays for fuel, so 21.6% of revenue is insulated from bunker cost spikes

• Route disruptions tighten global vessel supply, supporting charter rates on renewal. The prospectus already notes Suez Canal diversions since Dec 2023 increased vessel demand

• The shortage of ships below 4,000 TEU (order book ratio only 14.6%) means supply cannot respond quickly to higher demand

What hurts MTTSL in a disruption:

• Higher bunker costs hit their own liner operations directly (69.4% of revenue), with BAF pass-through that is periodic and lagged, and no financial hedging disclosed

• Their core cabotage routes (Peninsular ↔ East Malaysia) are domestic — not affected by Hormuz or Suez disruptions, meaning they bear the fuel cost increase without the corresponding freight rate spike that international routes see

• Dry bulk segment (5.3% of revenue) would weaken if global trade slows from stagflationary pressure

Net assessment: MTTSL is not a play on oil disruption or geopolitical risk. Their domestic cabotage dominance means they are somewhat insulated from global route disruptions, but they also don’t benefit from them the way tanker companies do. In an oil spike scenario, the primary impact is cost inflation on bunker fuel eating into liner margins, partially offset by the charter segment where fuel cost sits with the charterer.

15. Property Portfolio

Owned Properties: 14 material properties, ~RM206M total NBV

• 1 freehold (HQ in Shah Alam, RM26.0M)

• 13 leasehold (29-94 years remaining)

• 6 vacant land parcels (~RM77.5M) for future development

• Heavy concentration in Pulau Indah, Klang (~RM62.2M across 6 properties)

Rented Properties: 3 material leases

• Westports Depot lease expired at prospectus date (transitioning to owned Pulau Indah site)

• Pasir Gudang Depot (RM90,822/month, expiring Mar 2027, new 3-year agreement from Apr 2027)

• Perai Depot (RM86,320/month, expiring Jun 2026)

Property Red Flags

• Multiple properties operating without CCC/OC

• 1 property on provisional land title (Sarawak)

• Most properties charged as security for bank borrowings

16. Corporate Governance

Positive

• Board committees appropriately constituted with INED majorities

• No bankruptcy/criminal proceedings against any director in past 10 years

• Boardroom diversity target: at least 30% women directors

• Clean compliance record across all maritime regulations

Concerns

• Risk Management Committee chaired by non-independent executive director

• No service contracts with termination benefits for KSM (less formal structures)

• Heavy founder/family concentration in ownership and management

• ESOS allows up to 70% of options to Directors and senior management

• Board can extend ESOS from 5 to 10 years without shareholder approval

17. Quick Reference — Numbers That Matter

18. Summary Verdict Framework

Bull Case

• Dominant market position in protected cabotage market with no new entrants since 2011

• Youngest, largest fleet with superior profitability (2x competitor margins)

• Near-zero effective tax rate is a massive structural advantage

• Strong cash generation funds aggressive growth programme

• Favourable supply-demand dynamics for small containerships

• Multiple growth engines (IFFs, fleet expansion, automotive, chemical tankers, India)

• At ~8x P/E, reasonably priced for a market leader

Bear Case

• Post-COVID rate normalisation has compressed revenues and margins significantly

• Enormous capex programme (RM800M+ commitments) creates execution and leverage risk

• IPO proceeds go entirely to vessel purchases, not de-risking the balance sheet

• Historical dividend payout (10-13%) far below 50% target; heavy capex may persist for years

• Founder/family dominance with key-man risk (ages 59-71)

• Cabotage policy change would destroy the competitive moat

• Charter rate cyclicality means earnings could fall further if global shipping weakens

• RM180M Sukuk maturity in August 2026 needs immediate refinancing

This analysis was compiled from all 672 pages across three parts of the MTTSL Prospectus dated 26 March 2026. It is intended as a summary of investment-relevant information only and does not constitute investment advice. Investors should read the full prospectus and conduct their own due diligence.

Silicon Valley VCs have distilled it into two questions they ask every founder who walks through the door: Why you? Why now?

Strip away the politeness, and what they’re really asking is: where is your advantage?

And here’s where it gets uncomfortable. If you’re someone without capital, without connections, without a safety net — having an “edge” sounds like a cruel joke. Because if you actually had one, you wouldn’t be broke. You wouldn’t be starting from zero. You wouldn’t need someone to believe in you.

But I think most people are looking for their edge in the wrong place entirely.

I. The Mistake Everyone Makes

Edge is a deceptively vague word.

Money is an edge. Time is an edge. Connections are an edge. Even recklessness can be an edge if applied at the right moment.

But here’s where the majority of people go wrong: they spend their entire lives trying to build an advantage inside a game where someone else wrote the rules.

Think about it. You go to the best school you can get into — a game with rules set by admissions boards. You climb the corporate ladder — a game with rules set by HR departments and performance reviews. You invest in index funds or property — games with rules set by institutions that have been playing for decades longer than you.

In every one of these games, the rules are clear, the scoring system is established, and the people who designed the system have already won.

You are not going to out-compete Goldman Sachs at their own game. You are not going to out-network someone born into the right family. You are not going to out-save your way to wealth when the system extracts value from you at every turn.

II. Where Edge Actually Comes From

I used to believe that the dividing line was between “mediocre” and “excellent.”

That if you were just good enough at something — good enough at your job, good enough at investing, good enough at building — you’d be safe. And that AI would only replace the mediocre. The average performers. The people who weren’t trying hard enough.

I was wrong.

In an environment of exponential machine capability, the entire KPI system that defines “excellent” is being rewritten in real time. The metrics that made someone a top performer in 2023 are becoming table stakes by 2026. The goalposts aren’t just moving — the entire field is being redesigned.

When a machine can do in seconds what took a team of “excellent” people weeks, past definitions of excellence become meaningless. Not gradually. Overnight.

III. The Unwritten Game

So where does that leave someone starting from zero?

I’ve come to believe that the only real opportunity for a breakout — the kind that changes your trajectory, not just your income — exists in the moment before the rules are written.

I call this The Unwritten Game.

Every great wave of wealth creation in the last thirty years follows the same pattern:

The internet, early 1990s. No one knew what a “website” was worth. No one had written the playbook for e-commerce, digital media, or online advertising. The people who entered before the rules solidified — before Google defined search, before Facebook defined social, before the MBA playbooks existed — they captured disproportionate value. Not because they were the smartest. Because they were playing while the rules were still being written.

E-commerce, early 2000s. When Alibaba and Amazon were starting, “selling things online” was considered a joke by most serious business people. The rules of retail had been written for a century. The people who ignored those rules and played the unwritten game of digital commerce — they didn’t need capital advantages or network advantages. They needed timing and audacity.

Crypto, 2012-2017. The people who captured the most value from crypto weren’t finance professionals. Many of them were teenagers, hobbyists, outsiders. They had zero conventional edge. But they entered a space where the rules literally did not exist yet. No regulatory framework. No institutional playbook. No “correct” way to evaluate a token. The absence of rules was the edge.

AI, 2023-present. Right now, we’re watching the same pattern unfold. The rules of what AI can and can’t do, what jobs it replaces, what new roles it creates, what business models it enables — none of this is settled. It’s being written in real time, by the people who are playing.

The pattern is unmistakable: in every major wave, the biggest winners weren’t the most talented, the most connected, or the most resourced. They were the ones who entered before the rules were established.

IV. Why Pre-Rule Entry Is the Ultimate Edge

This isn’t just a historical observation. There’s a structural reason why entering before the rules are written creates disproportionate advantage:

1. Established players can’t compete in undefined spaces. Large companies, institutions, and credentialed experts derive their power from mastery of existing rules. When the rules change, their expertise becomes a liability. They’re optimizing for a game that no longer exists.

2. You don’t need resources to explore. In a defined game, resources win. You need capital to buy real estate, credentials to get hired, connections to get funded. But in an undefined game, the currency is curiosity and speed. You need to be willing to look stupid, try things, fail fast, and iterate — skills that are actually easier without the weight of a reputation to protect.

3. The risk-reward asymmetry is inverted. In established games, the downside is real and the upside is capped. You can lose your savings in a market crash but you’ll never earn 1000x in an index fund. In unwritten games, the downside is limited (your time, some failed experiments) but the upside is uncapped — because no one has defined the ceiling yet.

4. First frameworks become the rules. The people who build the first working models in a new space don’t just win the game — they define it. Google didn’t just win at search. It defined what search meant. The early players in any wave don’t find their edge. They create it.

V. “But Most People Who Jump Into Chaos Drown”

Yes. They do.

I’d be lying if I pretended otherwise. For every person who entered crypto in 2013 and built generational wealth, thousands entered in 2021 and got wiped out. For every early internet entrepreneur who became a billionaire, hundreds of dot-com founders went bankrupt. The base rate of failure in unwritten games is brutal.

So the obvious question is: how do you tell the difference between a pre-rule space that’s about to explode and a pre-rule space that’s a dead end?

Here’s the honest answer: you can’t. Not with certainty. Nobody can. And anyone who tells you they can is selling you something.

But there is a skill that separates the people who consistently find the right spaces early from the people who don’t. It’s not intelligence. It’s not connections. It’s not some mystical “business instinct.”

It’s curiosity without bias.

That sounds soft. It’s not. It might be the hardest skill in the world to maintain, because your brain is actively working against you.

Here’s what happens when most people encounter something new: they immediately reach for an old label. The internet in 2000? “It’s just scams and speculation.” Crypto in 2015? “Ponzi scheme. Tulip bubble.” AI tools in 2023? “Gimmick. It’ll plateau.”

The Label Trap feels like wisdom. It feels like pattern recognition. “I’ve seen this before. I know how this ends.” But it’s actually the opposite of pattern recognition. It’s pattern imposition. You’re not seeing what’s in front of you — you’re projecting what you’ve already seen onto something fundamentally different.

The internet in 2000 did have scams. Crypto did have Ponzi schemes. That part was true. But the people who stopped at the label — who filed the entire space under “scam” and moved on — missed the fact that the underlying technology was developing exponentially. They confused the noise of the early stage for the signal of the entire trajectory.

Meanwhile, the people who seemed weirdly “early” to every wave? They weren’t prophets. They weren’t geniuses. They just didn’t label things before they understood them.

That’s it. That’s the entire difference.

The people who look “sensitive” to new opportunities aren’t more perceptive. They’re less prejudiced. They encounter something unfamiliar and their first instinct isn’t to categorize it — it’s to understand it. They sit with the discomfort of not knowing what something is. They ask questions instead of making declarations. They spend time inside the thing before they decide whether it’s real.

Curiosity — genuine, bias-free curiosity — is one of the two or three most important abilities a person can have. Not because it makes you smarter. Because it prevents you from being blind.

The Label Trap is the reason most people miss every wave. Not because they’re stupid. Not because they lack resources. Because they saw the first 5% of a new space, matched it to something they already knew, and stopped looking.

But curiosity alone isn’t enough. Curiosity gets you in the door. Once you’re inside, you need a way to tell a positive-expected-value game from a negative one — before the answer is obvious. Because by the time the answer is obvious, the edge is gone.

VI. The EV Filter

I think about this the way a trader thinks about positions. A good trader doesn’t need to be right most of the time. They need their wins to be much larger than their losses, and they need to cut losers fast. The same logic applies to unwritten games.

You’re not trying to predict which new space will succeed. You can’t. Nobody can. What you can do is evaluate the structure of each bet before you place it. Here are the four variables that matter:

Variable 1: Payoff Asymmetry.

This is the single most important question: what’s the ratio between the worst case and the best case?

Some unwritten games have a structure where the downside is “I spent 3 months learning something that didn’t pan out, but I built skills that transfer elsewhere” and the upside is “I’m now an early expert in a space that’s about to reshape an industry.” That’s asymmetric. The loss is small, the win is enormous.

Other unwritten games have a structure where the downside is “I burned through my savings on equipment/inventory/tokens that went to zero” and the upside is “maybe I make 2-3x.” That’s a terrible trade. The potential win doesn’t justify the potential loss.

Before you commit serious time to any new space, sketch the actual payoff structure. Not the dream scenario — the realistic range. If the worst case is “I learned a lot and can apply it elsewhere,” you have a positive-EV game. If the worst case is “I’m financially wrecked and have nothing to show for it,” walk away. The specific threshold: never risk what you can’t rebuild in 6 months.

Variable 2: Skill Transferability.

This is the variable most people forget. In a positive-EV unwritten game, even your failures compound. The skills, relationships, and pattern recognition you develop transfer to adjacent spaces.

Someone who spent 2017-2019 building in crypto and “failed” — their project didn’t take off, they didn’t get rich — still understood decentralized systems, token economics, and community-driven growth. When DeFi exploded in 2020, they had a massive head start. Their “failure” was actually tuition for the next wave.

In a negative-EV unwritten game, the skills are a dead end. They don’t transfer. If the space collapses, you walk away with nothing applicable. Ask yourself: if this entire space evaporates tomorrow, would the skills I’ve built still be valuable somewhere else? If yes, your downside is much smaller than it looks. If no, you’re making a binary bet — and binary bets in undefined spaces are gambling, not exploring.

Variable 3: Feedback Loop Speed.

How quickly can you tell if your actions are working?

In a positive-EV unwritten game, you can run small experiments and get signal fast. You build something, put it in front of people, and learn within days or weeks whether the space has real demand. AI tools, content creation, software products — the feedback loop is measured in days. You can iterate, adjust, and compound your understanding rapidly.

In a negative-EV unwritten game, the feedback loop is measured in years. You invest time and capital into something and you won’t know if it’s working until long after you can course-correct. Drug development, deep hardware, speculative real estate in unproven markets — the cycle time between action and signal is so long that you’re essentially flying blind. Not all slow-feedback spaces are negative-EV, but the slower the feedback, the higher the bar should be for everything else on this list.

Variable 4: Cost of Exit.

Can you walk away cleanly? Or does leaving cost you almost as much as staying?

The best unwritten games let you fail fast and cheap. You try something for 3 months, it doesn’t work, you pivot. Your sunk cost is time and maybe a few hundred dollars. That’s a well-structured exploration.

The worst unwritten games have lock-in. You sign a lease, take on inventory, commit to a partnership with penalties, or invest capital that’s illiquid for years. Now leaving the game is expensive, which means you stay longer than you should, which means you throw good money (and time) after bad.

Before you enter any unwritten game, know your exit cost. If walking away is cheap, you can afford to be bold with your exploration. If walking away is expensive, you need to be right — and in an unwritten game, you almost never have enough information to know you’re right.

The quick filter: Run any new space through these four variables before committing beyond the initial curiosity phase. High payoff asymmetry + transferable skills + fast feedback + cheap exit = explore aggressively. Low on any two = proceed with extreme caution. Low on three or more = walk away, no matter how exciting the narrative sounds.

The point isn’t to avoid risk. The point is to take well-structured risk. Curiosity tells you where to look. The EV filter tells you where to stay.

VII. How to Actually Play the Unwritten Game

So what does this look like day to day? Here’s the system:

Step 1: Allocate exploration time that’s non-negotiable.

Block 5 hours per week — not for “learning” in the passive sense, but for active exploration of spaces you don’t understand yet. This isn’t reading articles about AI. It’s using AI tools to build something, however small. It’s joining a crypto community and watching how they think, not reading a Bloomberg take on crypto. It’s making a short-form video, not watching a course about content creation. The distinction matters: observers label, participants understand.

Step 2: Use the 3-Month Immersion Test.

When you find a space that interests you, don’t dabble. Spend 3 months going deep. Build something. Ship something. Talk to the people who are already in the space — not the commentators, the players. After 3 months, you’ll know more about the real dynamics of that space than 95% of the people who’ve been “following” it from the outside for years. If it’s a dead end, you’ll know. If it’s real, you’ll have a 3-month head start on everyone who’s still reading articles about it.

Step 3: Be cheap with money, reckless with energy.

This is where most people get the equation backwards. They either play it safe on every dimension — cautious with money and cautious with effort, which means they never learn anything — or they go all in on every dimension at once, quit their job, drain their savings, and flame out in two months. The correct split is asymmetric: protect your financial downside while being unreasonable with your time and intensity.Your job, your savings, your basic stability — those aren’t what you sacrifice. They’re what fund the exploration. Bet only what you can afford to lose financially. But within the game itself? Build the most. Ship the fastest. Ask the best questions. Connect the most dots. Be the person who’s unreasonably engaged while everyone else is still reading articles about whether the space is legitimate. The early winners of every wave didn’t outspend the latecomers. They out-playedthem — from positions that were often modest, sometimes broke, but never so fragile that one bad month ended the game.

VIII.

Everyone says choosing matters more than effort.

They’re right. But they leave out the hard part: choosing is harder than effort.

Effort is simple. Show up. Grind. Follow the checklist. There’s a strange comfort in exhaustion — it feels like progress even when you’re running on a treadmill. Choosing, on the other hand, requires you to confront the possibility that the game you’ve been grinding at is the wrong game entirely. That the years you invested were spent on someone else’s scoreboard.

Most people avoid that confrontation. They’d rather work harder inside a broken game than face the terror of choosing a new one. Because choosing means admitting uncertainty. And uncertainty feels like failure before it feels like freedom.

But here’s what I’ve come to believe:

The only path out of mediocrity is finding the best game you can play — and playing it to the absolute extreme.

Not the game your parents chose for you. Not the game that looks impressive on LinkedIn. Not the game where the rules are clear and the rewards are predictable. The game where your specific combination of curiosity, obsession, and tolerance for chaos gives you an advantage that no amount of someone else’s capital or credentials can replicate.

The Edge Paradox dissolves the moment you stop looking for advantage inside finished games and start looking for the game that fits you like a weapon.

Every person who changed their trajectory — not incrementally, but categorically — did two things: they found the right game, and they played it with an intensity that made everyone else uncomfortable.

Right now, the rules of work, value creation, and wealth-building are being rewritten at a speed we’ve never seen. The old game — study, credential, climb, retire — is a Squid Game with a predetermined outcome. But inside that chaos, new games are being born every day. Games that haven’t been named. Games with no established leaderboard. Games where showing up early and playing relentlessly is the edge.

Choosing is terrifying. But effort without choosing is just expensive procrastination.

Find the unwritten game that was made for the way your brain works. Then play it like your life depends on it.

Because it does.

If you follow semiconductor and AI infrastructure research, you know the feeling.

You open SemiAnalysis. There’s a 12,000-word deep dive on AMD’s MI450X rack-scale architecture. You skim it. You highlight a few lines. You tell yourself you’ll come back to it.

You won’t.

The next day, there’s a new piece on TSMC’s N2 process node. Then one on DeepSeek’s true training costs. Then NVIDIA’s GTC announcements. Then export controls. Then datacenter cooling systems. Then AWS Trainium3. Then tariff impacts on GPU supply chains.

Each article is dense, technical, and genuinely important. Each one contains data points, forward-looking predictions, and investment implications you can’t find anywhere else. And each one disappears into the scroll the moment the next one lands.

I spent months in this cycle. I had a SemiAnalysis subscription. I read the articles. I even took notes sometimes. But when someone asked me, “What’s the current state of AMD’s software ecosystem relative to NVIDIA?” I’d fumble. I knew I’d read about it. I knew there were specific numbers. I just couldn’t retrieve them.

The problem wasn’t that I lacked information. I was drowning in it. What I lacked was a system to *extract the intelligence from the information* and make it retrievable on demand.

So I built one. In a single afternoon. Using AI.

II. Why Your Notes App Won’t Save You

Before I explain the system, let me address the false solutions -- because I tried all of them.

Highlighting and bookmarking. This is the default behavior. You highlight passages, bookmark the article, and feel productive. But highlighted text without structure is just colored noise. Six months later, you have 200 highlights across 50 articles and no way to connect them.

Manual note-taking. Better, but not scalable. A single SemiAnalysis article on AMD’s MI450X architecture runs 600+ lines with dozens of data points spanning GPU specs, software ecosystem gaps, competitive dynamics, and acquisition strategy. Manually distilling that into a useful note takes 45 minutes *per article*. Multiply by 50 articles and you’ve lost a week of your life.

AI summaries (ChatGPT, “summarize this article”). This is the most tempting and most dangerous approach. A generic AI summary strips away precisely the information that makes SemiAnalysis valuable -- the specific numbers, the non-obvious analytical insights, the investment implications. You get a paragraph that tells you “AMD is making progress on software but still lags NVIDIA” -- something you already knew. The analytical signal gets crushed into a flat summary.

What I needed wasn’t summarization. It was structured intelligence extraction -- a system that could take a 600-line technical deep-dive and produce a research note that preserves the data points, isolates the non-obvious insights, and surfaces the investment implications, all in a format that connects to every other note in my knowledge base.

III. The Intelligence Pipeline -- A System for Turning Articles into Alpha

I call the system The Intelligence Pipeline. It has three components, and the entire thing runs on two tools: Obsidian (a note-taking app that stores everything as plain markdown files) and Claude Code (Anthropic’s AI coding agent that can read files, process them, and write structured output).

Here’s the architecture:

RAW INPUT          PROCESSING ENGINE       STRUCTURED OUTPUT
─────────          ─────────────────       ─────────────────
Obsidian           Claude Code AI          Research Notes
Web Clipper  ───>  + Processing Prompt ──> in Obsidian
(50 articles)      (format spec)           (50 notes)

Layer 1: Capture. Obsidian Web Clipper saves full SemiAnalysis articles as markdown files into a “Clippings/” folder. No formatting loss. No paywall issues. Raw, complete articles preserved as plain text.

Layer 2: Process. Claude Code reads every clipping and transforms it through a structured processing prompt -- a precise specification that tells the AI exactly what to extract and how to format it.

Layer 3: Connect. The output goes into a “reports/” folder as structured Obsidian notes with YAML frontmatter, wikilinks, and standardized sections. Every company, concept, and person becomes a clickable link in your knowledge graph.

The magic is in Layer 2 -- the processing prompt. That’s where the intelligence extraction happens. Everything else is just plumbing.

IV. The Processing Prompt -- Where Intelligence Extraction Happens

The processing prompt is a one-page specification that tells the AI exactly what to extract from each article. It’s the most important component of the system, and getting it right took several iterations.

Here’s what it produces for every article:

1. YAML Frontmatter (The Metadata Layer)

Every note starts with structured metadata: title, source URL, author, publication date, sector, and -- critically -- three linked fields:

Companies: Every company mentioned, as a wikilink. "[[AMD]]", "[[NVIDIA]]", "[[TSMC]]", "[[DeepSeek]]".
Concepts: Every technical concept, as a wikilink. "[[ROCm]]", "[[CUDA]]", "[[Jevons Paradox]]", "[[HBM]]", "[[Rack-Scale AI]]".
Tags: Searchable labels. "gpu-competition", "export-controls", "inference-pricing".

This metadata layer is what makes the system compound over time. After processing 50 articles, I don’t just have 50 notes -- I have a *graph* where clicking “[[NVIDIA]]” shows me every article that discusses NVIDIA, clicking “[[KV Cache]]” shows me every article where KV cache optimizations matter, and clicking “[[Dylan Patel]]” shows me every analysis by tat author.

2. Thesis (The 30-Second Brief)

Three sentences maximum. What is the core claim, what makes it unique, what are the implications. This is the section you read when someone asks “What did SemiAnalysis say about DeepSeek?” and you need an answer in 30 seconds.

Example from my DeepSeek note:

DeepSeek is a genuinely impressive fast-follower lab whose true training costs are far higher than the viral $6M figure, with total hardware spend exceeding $500M and server CapEx around $1.6B. The panic narrative that efficiency gains eliminate demand for compute is contradicted by Jevons Paradox dynamics already visible in soaring Nvidia H100 spot prices, and the real structural impact is margin compression on closed-model providers like OpenAI, not reduced GPU demand. Export Controls did not fail but rather came too late to prevent stockpiling, and future restrictions on the H20 will increasingly constrain DeepSeek’s scaling trajectory relative to US labs.

Three sentences. The entire analytical signal of a dense technical article.

3. Key Data Points (The Lookup Table)

Bulleted. Specific numbers, metrics, forecasts -- things you’d look up again. The kind of information that wins arguments and anchors investment theses.

Examples:

• DeepSeek GPU fleet: ~50,000 Hopper GPUs total; includes ~10,000 H800s, ~10,000 H100s

• True hardware spend. >$500M over company history; $6M covers only pre-training GPU cost for V3

• MLA KV Cache reduction: ~93.3% reduction per query versus standard attention

• Bank of China AI subsidy: $140B (1 Trillion Yuan) over 5 years

These are the numbers that vanish from your memory within a week of reading the original article. Now they’re permanently indexed and retrievable.

4. Key Insights (The Non-Obvious Analysis)

This is the highest-value section. Not summaries -- *analytical reasoning* that the AI extracts from the article’s deeper arguments. Each insight gets a bold title and a 2-3 sentence explanation.

Example:

The $6M figure is a misleading BOM line item, not total cost. The viral $6M training cost attributed to DeepSeek V3 covers only the GPU hours for the pre-training run itself. It excludes R&D, architecture experimentation (e.g., Multi-Head Latent Attention took months of dedicated team effort), ablation studies, data curation, synthetic data generation for R1, and total cost of ownership. The same logic would make Anthropic’s Claude 3.5 Sonnet look like it cost only tens of millions when the company has raised tens of billions.

This is intelligence you can deploy in a conversation. It’s the kind of insight that separates someone who “read about DeepSeek” from someone who *understands* DeepSeek.

5. Forward-Looking Predictions (The Crystal Ball)

Bulleted. Specific claims about the future with explicit timeframes. These are the testable predictions that make a research note appreciating over time rather than depreciating.

6. Investment Implications (The Action Layer)

Organized under three subheadings: Bullish, Bearish, Watch. Each entry: bold company wikilink + 1-2 sentence reasoning.

This section is what makes the system useful beyond pure intellectual exercise. Every note ends with actionable investment implications, tagged to specific companies, with clear reasoning.

V. The Transformation -- Before and After

Here’s what the system actually does to a real article.

Before (Raw Clipping): A 617-line SemiAnalysis article on AMD’s MI450X architecture. Dense paragraphs mixing executive summary points, Lisa Su meeting details, compensation analysis, Python DSL comparisons, rack-scale hardware specs, and forward-looking product roadmaps. Valuable, but impossible to scan.

After (Processed Research Note): A ~130-line structured note with:

• 17 companies wikilinked (AMD, NVIDIA, Microsoft, OpenAI, Oracle, Google, Meta, xAI, Tesla, Broadcom, Astera Labs, ZT Systems, Quanta, Foxconn, Celestica, Lenovo, Huawei)

• 19 concepts wikilinked (ROCm, CUDA, NCCL, NVLink, PyTorch, Triton, UALink, HBM, Rack-Scale AI, Python DSL...)

• 21 specific data points (GPU cluster sizes, compensation gaps, bandwidth specs, TCO comparisons)

• 7 non-obvious insights (compensation benchmarking as the real bottleneck, Python ecosystem as NVIDIA’s true new moat, UALink effectively dead for H2 2026)

• Timeline-specific predictions

• Bullish/Bearish/Watch investment implications for individual companies

A 617-line wall of text became a structured intelligence asset. And it took the AI about 60 seconds.

VI. The Build Protocol -- How to Set This Up

Here’s exactly what I did, step by step.

Step 1: Set Up the Capture Layer

Install Obsidian. Install the Obsidian Web Clipper browser extension. Configure it to save clippings as markdown into a “Clippings/” folder inside your vault. Clip your articles. I clipped 50 SemiAnalysis articles spanning January 2025 to February 2026 -- covering export controls, chip architecture, datacenter infrastructure, AI training economics, and GPU competition.

Step 2: Write the Processing Prompt

This is the critical step. Create a file called “_processing-prompt.md” in your “reports/” folder. This file is the instruction set for the AI. It specifies:

• The exact YAML frontmatter structure (with rules for wikilink quoting, block list format, no inline arrays)

• The six sections: Thesis, Key Data Points, Key Insights, Forward-Looking Predictions, Investment Implications

• Length constraints (150-250 lines per note)

• Quality rules (extract analytical signal, don’t paraphrase, use wikilinks for all entities)

The processing prompt is your editorial voice encoded as a specification. It determines what the AI pays attention to and what it ignores. Get this right and the output is consistently excellent. Get it wrong and you get generic summaries.

Step 3: Run the Processing Engine

Open Claude Code in your vault directory. Point it at the processing prompt and the clippings folder. Let it read each article, apply the specification, and generate the structured notes.

For 50 articles, the entire processing run takes an afternoon. Each note takes about 60-90 seconds. The AI reads the full article, identifies companies and concepts, extracts data points, isolates non-obvious insights, generates predictions, and formats investment implications -- all following the processing prompt specification.

Step 4: Verify and Refine

Spot-check the output. The first few notes may reveal gaps in your processing prompt -- maybe it’s missing a section you care about, or the thesis length isn’t right, or the data points aren’t specific enough. Iterate on the prompt. Re-run the notes that need improvement.

After one round of refinement, I had 50 structured research notes covering the entire SemiAnalysis output from the past year. Every note interlinked. Every company and concept clickable. Every data point indexed.

VII. What Changes When You Have a Second Brain

The transformation isn’t just organizational. It changes how you think about the sector.

Retrieval becomes instant. Someone mentions TSMC N2 in conversation. You open Obsidian, click [[TSMC]], and see every article that discusses TSMC -- with thesis statements, data points, and investment implications already extracted. You’re not searching your memory. You’re querying a knowledge graph.

Connections become visible. When you click [[KV Cache]], you see it mentioned in the DeepSeek article (MLA’s 93.3% reduction), the AMD article (MI450X’s 4-tier memory hierarchy for cache tiering), and the inference benchmark article. Patterns that were invisible when the information was scattered across 50 separate readings become obvious when they’re linked in a graph.

Investment thinking compounds. Every note ends with Bullish/Bearish/Watch implications. After 50 notes, you don’t just have 50 separate investment views -- you have a multi-source, cross-referenced investment thesis that strengthens or weakens as new data arrives. When a new SemiAnalysis article drops, you process it, and its implications automatically connect to everything you’ve already built.

Your reading becomes non-linear. Instead of reading articles sequentially as they publish, you can now navigate by company, by concept, by time period, or by investment implication. Want to understand AMD’s trajectory? Follow the [[AMD]] link through every article. Want to understand the inference cost curve? Follow [[Inference Cost Deflation]] across multiple analyses.

VIII. The System Behind the System

There’s a meta-insight here that goes beyond semiconductor research.

The traditional approach to knowledge work is: read, highlight, forget, re-read. The AI-augmented approach is: capture, process, structure, connect. The difference is that the second approach produces a compounding asset rather than a depreciating memory.

Every note I add makes every other note more valuable -- because it creates new connections in the graph. The 50th note is more valuable than the 1st note, not because the article was better, but because it connects to 49 other structured analyses.

This is what I call The Intelligence Compound Effect: the value of your knowledge base grows non-linearly with every new input, because each input creates connections to everything that came before.

And here’s the part that matters most: the bottleneck is no longer processing capacity. AI handles that. The bottleneck is now editorial judgment -- choosing what to read, how to structure the extraction, and what connections to pursue. The processing prompt is where your expertise lives. The AI is the engine. You’re the architect.

IX. Start Here

If you’ve read this far, you’re probably thinking about what domain you’d build this for. Here’s the minimum viable version:

1. Pick your 10 highest-signal sources. Newsletters, research reports, earnings transcripts -- whatever contains the raw intelligence for your domain.

2. Write a one-page processing prompt. Define the sections. Define what “insight” means in your context. Define the metadata structure that will power your graph.

3. Process 10 articles. Not 50. Ten. See what the output looks like. Refine the prompt.

4. Use the output for one real decision. Answer a question, write an analysis, make an investment call. The system proves its value the first time you retrieve something you would have otherwise forgotten.

The people who will have an information advantage in the next few years aren’t the ones who read the most. They’re the ones who built systems to extract, structure, and compound the intelligence from what they read.

The tools exist. The articles exist. The only question is whether you’ll build the pipeline -- or keep drowning in the scroll.