Mostly Borrowed Ideas Deep Dives, MBI Deep Dives

MBI Deep Dives

Mostly Borrowed Ideas Deep Dives

MBI is the abbreviation for “Mostly Borrowed Ideas”. Why “Mostly Borrowed Ideas?” As a generalist, I do not have a background in any particular sector. I enjoy navigating across industries, businesses, and countries to learn, understand, and connect the dots. In any case, most ideas and innovations are incremental in nature. Very few of us are smart enough to come up with truly original or groundbreaking ideas. I am Abdullah Al-Rezwan. I was born and brought up in Bangladesh. I went to University of Dhaka for my undergrad and majored in Finance. After working three and half years in Equity Research (Sell-side) covering Financial Sector in Bangladesh, I left Bangladesh to pursue my MBA in the US. I graduated within top 10% of the class from Cornell University's two-year MBA program and worked on the buy-side following my graduation. During my time at Cornell, I won two stock pitch competitions and also became finalists in four other competitions. I thoroughly enjoyed my stint as a generalist in the US large cap team at Madison Investments and covered a wide range of companies from UnitedHealth to Amazon and Boeing before US immigration reality paused my career in investment management. As my work authorization expired, I moved to Canada in January, 2021. Then in early 2022, I was able to return to the US.

Equity Bottom-Up

<ul><li>While Amazon’s 3P business usually grows faster than 1P, both 1P and 3P retail business grew at similar rate in 1Q’25.</li><li>Ads revenue continued its momentum at 19% growth YoY which was higher than both Google and Meta.</li><li>After growing at ~19% YoY for the last three consecutive quarters, AWS growth decelerated this quarter to 17%.</li></ul>

While Amazon’s 3P business usually grows faster than 1P, both 1P and 3P retail business grew at similar rate in 1Q’25.Ads revenue continued its momentum at 19% growth YoY which was higher than both Google and Meta.After growing at ~19% YoY for the last three consecutive quarters, AWS growth decelerated this quarter to 17%.

Managing Director, Zero One, Zero One

Vincent Fernando, CFA

Managing Director, Zero One

Zero One

Founded in 2015, Zero One produces independent investment research and provides corporate finance advisory services with a focus on both listed and private technology companies. Founder Vincent Fernando has over 20 years of experience in US and Asia capital markets.

Independent Investment Research

TSMC (Taiwan Semiconductor Manufacturing) - ADR / Equity Bottom-Up

TSMC (Taiwan Semiconductor Manufacturing) - ADR

<ul>
<li>
TSMC’s Arizona fab hits Taiwan-level yields, easing replication concerns and reinforcing its global leadership across N2 and advanced packaging. </li>
<li>
AI demand offsets smartphone softness in 1Q25; TSMC guides +13% QoQ revenue for 2Q25 as margins hold firm despite tariff and earthquake headwinds. </li>
<li>
Faraday and UMC results ahead this week -- Key readouts on Taiwan’s ASIC, mature node, and design service momentum amid U.S.-China tech decoupling.</li>
</ul>



TSMC’s Arizona fab hits Taiwan-level yields, easing replication concerns and reinforcing its global leadership across N2 and advanced packaging. 

AI demand offsets smartphone softness in 1Q25; TSMC guides +13% QoQ revenue for 2Q25 as margins hold firm despite tariff and earthquake headwinds. 

Faraday and UMC results ahead this week -- Key readouts on Taiwan’s ASIC, mature node, and design service momentum amid U.S.-China tech decoupling.


Taiwan Tech Weekly: TSMC U.S. Bet Pays Off; UMC & Faraday Step Up Next; Latest Mobile Shipments Data

HarmoniQ Insights

William Mann

Independent Insight Provider

At HarmoniQ Insights, we empower investors to enhance their investment strategies by integrating cutting-edge quantitative research processes and technologies. Our solutions provide access to leading quantitative data providers and services, tailored to align with each client’s unique goals. By bridging the gap between traditional investment approaches and advanced data-driven methods, we help clients gain a competitive edge through innovation in both process and technology.

Cross Asset Strategy

<ul>
<li>A fascinating lunch discussion with LLM engineer sparked an investigation into how AI models process information through vector spaces - revealing parallel but mechanistically different biases compared to human analysts </li>
<li>Built comprehensive framework mapping traditional financial biases against their LLM counterparts - showing how things like confirmation bias manifest differently (emotional investment vs statistical pattern reinforcement)</li>
<li>Result: Effective quant research requires understanding both human AND machine biases - can't just apply traditional risk controls to LLMs. Need tailored mitigations for how biases uniquely manifest in #AI.  </li>
</ul>


A fascinating lunch discussion with LLM engineer sparked an investigation into how AI models process information through vector spaces - revealing parallel but mechanistically different biases compared to human analysts 
Built comprehensive framework mapping traditional financial biases against their LLM counterparts - showing how things like confirmation bias manifest differently (emotional investment vs statistical pattern reinforcement)
Result: Effective quant research requires understanding both human AND machine biases - can't just apply traditional risk controls to LLMs. Need tailored mitigations for how biases uniquely manifest in #AI.  


Understanding LLM Biases in Quantitative Research: From Lunch Conversation to Framework

<ul>
<li>A Comprehensive Guide to Building AI-Powered Investment Systems: The application of Large Language Models (LLMs) in quantitative finance, focusing on the current state of LLM technology in financial applications </li>
<li>I also cover: 1. Comparative analysis of leading models                2. Implementation frameworks for production systems 3. Risk management and quality control considerations </li>
<li>Key Findings: 1. Multi-model approaches outperform single-model solutions 2. Production implementation requires robust quality controls 3. Model selection should be task-specific within the investment process </li>
</ul>


A Comprehensive Guide to Building AI-Powered Investment Systems: The application of Large Language Models (LLMs) in quantitative finance, focusing on the current state of LLM technology in financial applications 
I also cover: 1. Comparative analysis of leading models                2. Implementation frameworks for production systems 3. Risk management and quality control considerations 
Key Findings: 1. Multi-model approaches outperform single-model solutions 2. Production implementation requires robust quality controls 3. Model selection should be task-specific within the investment process 


The LLM Quant Revolution: From ChatGPT to Wall Street

Head of Research, IPOs & Placements, Aequitas Research

Sumeet Singh

Head of Research, IPOs & Placements

IPOs and Placements coverage across Asia Pacific

Aequitas Research

We cover all listing across the Asia Pacific along with follow-up placements. As of Dec 2020, our team had covered 539 IPOs with a hit rate of 73.5% and 685 placements with a hit rate of 66.8%.

We combine fundamental bottom-up views with our proprietary quantitative framework to provide a holistic coverage across new listings and placements. Our coverage includes pre-IPO notes before the deal is launched, follow-up analysis once the deal is live and post-listing trading analysis as well.

Equity Capital Markets

<ul>
<li>
<a href="https://www.smartkarma.com/entities/xiaomi-corp" target="_blank" data-smartkarma-citation="entities:24409:1" rel="noopener">Xiaomi Corp (1810 HK)</a> is looking to raise around US$5.3bn via selling 3% additional shares.</li>
<li>The shares have done exceedingly well this year and are now trading at their all time highs.</li>
<li>In this note, we will talk about the placement and run the deal through our ECM framework.</li>
</ul>



Xiaomi Corp (1810 HK) is looking to raise around US$5.3bn via selling 3% additional shares.
The shares have done exceedingly well this year and are now trading at their all time highs.
In this note, we will talk about the placement and run the deal through our ECM framework.


Xiaomi US$5.3bn Placement - Relatively Small, Strong Momentum but Is Expensive

United States

Alphabet, Inc. operates as a holding company. The Company, through its subsidiaries, provides web-based search, advertisements, maps, software applications, mobile operating systems, consumer content, enterprise solutions, commerce, and hardware products.

Price Gauge
52-week Range	0.00 - 0.00
Market Cap (USDm)	0.00
P/E	0.00
Dividend (Yield)	0.00

Alphabet (GOOG US)

Amazon 1Q'25 Update

While Amazon’s 3P business usually grows faster than 1P, both 1P and 3P retail business grew at similar rate in 1Q’25. Ads revenue continued its...

MBI Deep Dives

Taiwan Tech Weekly: TSMC U.S. Bet Pays Off; UMC & Faraday Step Up Next; Latest Mobile Shipments Data

TSMC's Arizona fab achieves Taiwan-level yields, reinforcing global leadership in N2 and advanced packaging, while AI demand offsets smartphone...

Vincent Fernando, CFA

Understanding LLM Biases in Quantitative Research: From Lunch Conversation to Framework

We are going to provide a full-blown update on LLMs through Q1 2025. In the meantime, we are still finding issues with 🤖 “Hallucinating on Drugs” –...

William Mann

The LLM Quant Revolution: From ChatGPT to Wall Street

Please see the AI Agent Podcast explanation of this paper here: https://youtu.be/11A2tremqm0?si=VUMm6WhWGEQRRkNU If you enjoy this podcast, please...

William Mann

Xiaomi US$5.3bn Placement - Relatively Small, Strong Momentum but Is Expensive

Xiaomi Corp (1810 HK) is looking to raise around US$5.3bn via selling 3% additional shares. In this note, we run the deal through our ECM framework.

Sumeet Singh

Price Gauge

52-week Range

Market Cap (USDm)

P/E

Dividend (Yield)

Amazon 1Q'25 Update

While Amazon’s 3P business usually grows faster than 1P, both 1P and 3P retail business grew at similar rate in 1Q’25. Ads revenue continued its...

MBI Deep Dives

Taiwan Tech Weekly: TSMC U.S. Bet Pays Off; UMC & Faraday Step Up Next; Latest Mobile Shipments Data

​TSMC's Arizona fab achieves Taiwan-level yields, reinforcing global leadership in N2 and advanced packaging, while AI demand offsets smartphone...

Vincent Fernando, CFA

Understanding LLM Biases in Quantitative Research: From Lunch Conversation to Framework

We are going to provide a full-blown update on LLMs through Q1 2025. In the meantime, we are still finding issues with 🤖 “Hallucinating on Drugs” –...

William Mann

The LLM Quant Revolution: From ChatGPT to Wall Street

Please see the AI Agent Podcast explanation of this paper here: https://youtu.be/11A2tremqm0?si=VUMm6WhWGEQRRkNU If you enjoy this podcast, please...

William Mann

Xiaomi US$5.3bn Placement - Relatively Small, Strong Momentum but Is Expensive

Xiaomi Corp (1810 HK) is looking to raise around US$5.3bn via selling 3% additional shares. In this note, we run the deal through our ECM framework.

Sumeet Singh

TSMC's Arizona fab achieves Taiwan-level yields, reinforcing global leadership in N2 and advanced packaging, while AI demand offsets smartphone...