Dimitry Nakhla | Babylon Capital®
In hindsight, perhaps a year from today, the brief moment $MSFT traded at ~26x may look like an obvious opportunity hiding in plain sight. https://t.co/G5EHiyA5AB

A quality valuation analysis on $MSFT 🧘🏽‍♂️

•NTM P/E Ratio: 27.29x
•3-Year Mean: 30.51x

•NTM FCF Yield: 2.09%
•3-Year Mean: 2.46%

As you can see, $MSFT appears to be trading below fair value on a forward earnings basis

Going forward, investors can expect to receive ~12% MORE in earnings per share & ~18% LESS in FCF per share🧠***

Before we get into valuation, let’s take a look at why $MSFT is a super business

BALANCE SHEET✅
•Cash & Equivalents: $102.01B
•Long-Term Debt: $35.38B

$MSFT has an excellent balance sheet, an AAA S&P Credit Rating & 58x FFO Interest Coverage Ratio

RETURN ON CAPITAL✅
•2021: 31.1%
•2022: 34.0%
•2023: 30.9%
•2024: 29.7%
•2025: 28.0%

RETURN ON EQUITY✅
•2021: 47.1%
•2022: 47.2%
•2023: 38.8%
•2024: 37.1%
•2025: 33.3%

$MSFT has strong return metrics, highlighting the financial efficiency of the business

REVENUE✅
•2021: $168.09B
•2026E: $326.83B
•CAGR: 14.22%

FREE CASH FLOW🆗*
•2021: $56.12B
•2026E: $75.05B
•CAGR: 5.98%

*This is largely due to heavy AI-related reinvestment — current 2028 FCF estimate $116.45B — worth noting operating cash flow increases underscore $MSFT efficient AI infrastructure scaling validating high ROI-potential

NORMALIZED EPS✅
•2021: $7.97
•2026E: $16.26
•CAGR: 15.32%

SHARE BUYBACKS✅
•2016 Shares Outstanding: 8.01B
•LTM Shares Outstanding: 7.46B

By reducing its shares outstanding ~7%, $MSFT increased its EPS by ~8% (assuming 0 growth)

MARGINS✅
•LTM Gross Margins: 68.8%
•LTM Operating Margins: 46.3%
•LTM Net Income Margins: 35.7%

PAID DIVIDENDS✅
•2015: $1.24
•2025: $3.32
•CAGR: 10.34%

***NOW TO VALUATION 🧠

As stated above, investors can expect to receive ~12% MORE in EPS & ~18% LESS in FCF per share

Using Benjamin Graham’s 2G rule of thumb, $MSFT has to grow earnings at a 13.65% CAGR over the next several years to justify its valuation

Today, analysts anticipate 2026 - 2028 EPS growth over the next few years to be more than the (13.65%) required growth rate:

2026E: $16.26 (19% YoY) *FY Jun

2027E: $18.75 (15% YoY)
2028E: $22.31 (19% YoY)

$MSFT has an excellent track record of meeting analyst estimates ~2 years out, so let’s assume $MSFT ends 2028 with $22.31 in EPS & see its CAGR potential assuming different multiples

30x P/E: $669💵 … ~17.9% CAGR

29x P/E: $647💵 … ~16.3% CAGR

28x P/E: $625💵 … ~14.7% CAGR

27x P/E: $602💵 … ~13.0% CAGR

26x P/E: $580💵 … ~11.3% CAGR

As you can see, we’d have to assume ~28x multiple for $MSFT to have attractive return potential

At 26x - 27x earnings $MSFT has ok CAGR potential

If $MSFT multiple expands slightly, >15% CAGR

$MSFT is one of the highest quality companies in the world & is firing on all cylinders

Today at $454💵 $MSFT appears to be a strong consideration for investment with a decent margin of safety

$MSFT has large margin of safety at $420💵, where I can reasonably expect ~13% CAGR while assuming a more conservative 25x
___

𝐃𝐈𝐒𝐂𝐋𝐎𝐒𝐔𝐑𝐄‼️

𝐓𝐡𝐢𝐬 𝐜𝐨𝐧𝐭𝐞𝐧𝐭 𝐢𝐬 𝐩𝐫𝐨𝐯𝐢𝐝𝐞𝐝 𝐟𝐨𝐫 𝐢𝐧𝐟𝐨𝐫𝐦𝐚𝐭𝐢𝐨𝐧𝐚𝐥 𝐚𝐧𝐝 𝐞𝐝𝐮𝐜𝐚𝐭𝐢𝐨𝐧𝐚𝐥 𝐩𝐮𝐫𝐩𝐨𝐬𝐞𝐬 𝐨𝐧𝐥𝐲 𝐚𝐧𝐝 𝐝𝐨𝐞𝐬 𝐧𝐨𝐭 𝐜𝐨𝐧𝐬𝐭𝐢𝐭𝐮𝐭𝐞 𝐢𝐧𝐯𝐞𝐬𝐭𝐦𝐞𝐧𝐭 𝐚𝐝𝐯𝐢𝐜𝐞, 𝐚𝐧 𝐨𝐟𝐟𝐞𝐫, 𝐨𝐫 𝐚 𝐬𝐨𝐥𝐢𝐜𝐢𝐭𝐚𝐭𝐢𝐨𝐧 𝐭𝐨 𝐛𝐮𝐲 𝐨𝐫 𝐬𝐞𝐥𝐥 𝐚𝐧𝐲 𝐬𝐞𝐜𝐮𝐫𝐢𝐭𝐲.

𝐁𝐚𝐛𝐲𝐥𝐨𝐧 𝐂𝐚𝐩𝐢𝐭𝐚𝐥® 𝐚𝐧𝐝 𝐢𝐭𝐬 𝐫𝐞𝐩𝐫𝐞𝐬𝐞𝐧𝐭𝐚𝐭𝐢𝐯𝐞𝐬 𝐦𝐚𝐲 𝐡𝐨𝐥𝐝 𝐩𝐨𝐬𝐢𝐭𝐢𝐨𝐧𝐬 𝐢𝐧 𝐭𝐡𝐞 𝐬𝐞𝐜𝐮𝐫𝐢𝐭𝐢𝐞𝐬 𝐝𝐢𝐬𝐜𝐮𝐬𝐬𝐞𝐝. 𝐀𝐧𝐲 𝐨𝐩𝐢𝐧𝐢𝐨𝐧𝐬 𝐞𝐱𝐩𝐫𝐞𝐬𝐬𝐞𝐝 𝐚𝐫𝐞 𝐚𝐬 𝐨𝐟 𝐭𝐡𝐞 𝐝𝐚𝐭𝐞 𝐨𝐟 𝐩𝐮𝐛𝐥𝐢𝐜𝐚𝐭𝐢𝐨𝐧 𝐚𝐧𝐝 𝐬𝐮𝐛𝐣𝐞𝐜𝐭 𝐭𝐨 𝐜𝐡𝐚𝐧𝐠𝐞 𝐰𝐢𝐭𝐡𝐨𝐮𝐭 𝐧𝐨𝐭𝐢𝐜𝐞.

𝐈𝐧𝐟𝐨𝐫𝐦𝐚𝐭𝐢𝐨𝐧 𝐡𝐚𝐬 𝐛𝐞𝐞𝐧 𝐨𝐛𝐭𝐚𝐢𝐧𝐞𝐝 𝐟𝐫𝐨𝐦 𝐬𝐨𝐮𝐫𝐜𝐞𝐬 𝐛𝐞𝐥𝐢𝐞𝐯𝐞𝐝 𝐭𝐨 𝐛𝐞 𝐫𝐞𝐥𝐢𝐚𝐛𝐥𝐞 𝐛𝐮𝐭 𝐢𝐬 𝐧𝐨�[...]

� 𝐠𝐮𝐚𝐫𝐚𝐧𝐭𝐞𝐞𝐝 𝐚𝐬 𝐭𝐨 𝐚𝐜𝐜𝐮𝐫𝐚𝐜𝐲 𝐨𝐫 𝐜𝐨𝐦𝐩𝐥𝐞𝐭𝐞𝐧𝐞𝐬𝐬. 𝐏𝐚𝐬𝐭 𝐩𝐞𝐫𝐟𝐨𝐫𝐦𝐚𝐧𝐜𝐞 𝐝𝐨𝐞𝐬 𝐧𝐨𝐭 𝐠𝐮𝐚𝐫𝐚𝐧𝐭𝐞𝐞 𝐟𝐮𝐭𝐮𝐫𝐞 𝐫𝐞𝐬𝐮𝐥𝐭𝐬. - Dimitry Nakhla | Babylon Capital® tweet

Quiver Quantitative
Someone on Polymarket has bet almost $100K that the US will strike Iran by the end of the month.

Insider or gambler? https://t.co/ObLZHiUTo7
tweet

App Economy Insights
Two signals in Big Tech this week:

☁️ SaaSpocalypse Now $WCLD
🏭 Intel’s Supply Squeeze $INTC

Full story with visuals 👇
https://t.co/6vfDyTOC1i
tweet

memenodes
One day i will go away from everyone to the mountains like this https://t.co/OlzsDEVkKa
tweet

Fiscal.ai
RT @StockMKTNewz: Netflix $NFLX brought in $5.3 Billion from the United States 🇺🇸 and Canada 🇨🇦 last quarter up from $3.3B in Q4 2021 https://t.co/ro6g2sG1Pf
tweet

God of Prompt
RT @ytscribeai: life hack for content creators:

1/ n8n watches youtube channels via RSS (no API key)
2/ https://t.co/sWBIcxGUoD extracts transcripts ($0.0035 each)
3/ AI turns them into newsletters, blogs, tweets

one workflow. infinite content engine.

https://t.co/sWBIcxGUoD https://t.co/cRrSCMABas
tweet

God of Prompt
RT @alex_prompter: This paper from Google DeepMind, Meta, Amazon, and Yale University quietly explains why most “AI agents” feel smart in demos and dumb in real work.

The core idea is simple but uncomfortable: today’s LLMs don’t reason, they react. They generate fluent answers token by token, but they don’t explicitly plan, reflect, or decide when to stop and rethink. This paper argues that real progress comes from turning LLMs into agentic reasoners systems that can set goals, break them into subgoals, choose actions, evaluate outcomes, and revise their strategy mid-flight.

The authors formalize agentic reasoning as a loop, not a prompt:

observe → plan → act → reflect → update state → repeat.

Instead of one long chain-of-thought, the model maintains an internal task state. It decides what to think about next, not just how to finish the sentence.

This is why classic tricks like longer CoT plateau. You get more words, not better decisions.

One of the most important insights: reasoning quality collapses when control and reasoning are mixed. When the same prompt tries to plan, execute, critique, and finalize, errors compound silently. Agentic setups separate these roles.

Planning is explicit. Execution is scoped. Reflection is delayed and structured.

The paper shows that even strong frontier models improve dramatically when given:

• explicit intermediate goals
• checkpoints for self-evaluation
• the ability to abandon bad paths
• memory of past attempts

No new weights. No bigger models. Just better control over when and why the model reasons.

The takeaway is brutal for the industry: scaling tokens and parameters won’t give us reliable agents. Architecture will. Agentic reasoning isn’t a feature it’s the missing operating system for LLMs.

Most “autonomous agents” today are just fast typists with tools.

This paper explains what it actually takes to build thinkers.
tweet