For four decades, Warren Buffett refused to buy technology stocks. He famously sat out the entire dot-com bubble, was widely mocked for missing Amazon's compounding decades, and consistently described tech as outside his circle of competence. Then, in 2016, he started buying Apple. By 2024 Apple was Berkshire's largest position by a wide margin, and the cumulative gain was over 100 billion dollars. Around the same time, Charlie Munger bought Alibaba (BABA). The lesson was clear: tech is not the opposite of value investing if you know what to look for.
This post unpacks what changed in Buffett's thinking, what the actual value-investing test for a tech stock looks like, and surfaces 5 tech names that pass the invest-like 7-framework consensus screen today.
What changed with Apple
When Buffett explained the Apple purchase in shareholder Q and A, he was specific: he did not view Apple as a tech company in the speculative-tech sense. He viewed Apple as a consumer products company with a brand moat, switching-cost moat, and high-margin recurring services revenue. The business mechanics resembled Coca-Cola more than they resembled a typical tech start-up.
That distinction is the key. The historical Buffett objection to tech was not that tech companies were bad businesses. It was that:
- The competitive landscape changes too fast to project 10 years of cash flow.
- Many tech companies have no durable moat (innovation lap competitors).
- Stock-based compensation hides the true cost of running the business.
- The valuations imply future dominance that may not materialise.
Apple did not meet these objections by being non-tech. Apple met them by being a tech company whose business mechanics resembled the durable consumer franchises Buffett already understood.
The same logic applies to evaluating any tech stock today. The question is not "is this a tech company." The question is "do this company's business mechanics resemble a durable franchise?"
The 5 actual filters for value investing in tech
Translating the Buffett framework specifically for tech requires emphasising certain criteria:
Filter 1: ROIC above 20 percent for 5+ consecutive years
Tech businesses without true moats have ROIC that compresses fast as competitors copy the product. The 20 percent threshold (higher than the 15 percent general threshold) ensures the moat is real, not just a function of being early.
Apple's ROIC has been above 25 percent for over a decade. Microsoft's has been above 25 percent for 8+ years. Google's similarly. These are the tech businesses where the moat is structural.
Filter 2: Stock-based compensation below 8 percent of revenue
The dirtiest secret of tech accounting is that GAAP earnings often dramatically overstate true profitability because stock-based compensation (SBC) is a real expense that GAAP buries. Many otherwise "profitable" tech companies have SBC equal to 15-25 percent of revenue. If you subtract that as a real cash cost, the GAAP earnings shrink dramatically or disappear.
The 8 percent threshold filters out the worst offenders. Apple's SBC is roughly 2 percent of revenue. Microsoft's is roughly 5 percent. Some Silicon Valley software companies run 25 percent SBC; we exclude those.
Filter 3: FCF margin above 25 percent
True software economics produce high FCF margins because the marginal cost of an additional customer is near zero. A "tech" company with 5 percent FCF margins is structurally not running software economics; it is running services or hardware economics, and the multiple it deserves is closer to a services or hardware company.
Filter 4: Predictable recurring revenue or extreme switching costs
The "circle of competence" filter. A tech company is investable in the Buffett sense if you can confidently predict the customer behaviour 10 years out. Recurring subscription revenue with low churn (Adobe, Microsoft 365) meets this test. Customer behaviour that depends on the next product cycle (consumer-app trends, fashion-driven hardware) does not.
Filter 5: Reasonable valuation: owner earnings yield above 3 percent OR PEG below 1.5
Many tech names fail this filter today because they trade at extremely rich multiples. Apple at owner earnings yield 2.9 percent fails this filter (we noted this in the Buffett-Fit analysis of Apple).
The valuation filter is where most tech candidates fall out. The market understands the moats and bids them up. Finding a tech name that passes both quality AND valuation is genuinely rare.
5 tech names that pass the framework consensus today
After running the invest-like universe through the 7-framework screen with the tech filters applied, five tech names pass at the 5+ of 7 level as of May 2026:
1. Microsoft (MSFT)
ROIC above 30 percent for 8+ years. SBC roughly 5 percent of revenue. FCF margin above 35 percent. Recurring revenue from Office 365 and Azure with high switching costs. Valuation stretched (owner earnings yield 2.5 percent) but PEG roughly 2.4 reflects strong growth.
The valuation pillar is the constraint. Microsoft is a 7-of-7 quality compounder at 5-of-7 because of the price. /buffett/msft/. The full deep-dive is at /blog/microsoft-msft-7-framework-analysis/.
2. Alphabet (GOOGL)
ROIC above 20 percent consistently. SBC closer to 7 percent of revenue (higher than Apple or Microsoft but below the 8 percent threshold). FCF margin above 25 percent. Search business has structural advertising network effects.
Valuation slightly more attractive than Microsoft or Apple. Owner earnings yield around 3.2 percent. PEG roughly 1.6. Passes most frameworks. /buffett/googl/. Full analysis at /blog/googl-alphabet-7-framework-analysis/.
3. Fortinet (FTNT)
Cybersecurity hardware and software, founder-led (the Xie family still owns a meaningful stake). ROIC above 30 percent. Recurring subscription revenue dominates. SBC around 6 percent of revenue. FCF margin above 35 percent.
Smaller than the mega-caps, so the moat questions are more nuanced, but the financial profile is among the cleanest in cybersecurity. Passes 6 of 7 frameworks. /buffett/ftnt/.
4. Mastercard (MA) (counted as tech / fintech)
Two-sided payments network with extremely high switching costs and structural network effects. ROIC above 35 percent. SBC roughly 4 percent of revenue. FCF margin above 45 percent. The cleanest "tech-adjacent" name on this list by financial quality.
Valuation stretched (owner earnings yield around 2.5 percent) but PEG and growth profile compensate. Passes 7 of 7 frameworks despite the rich price. /buffett/ma/.
5. Visa (V)
Twin-network to Mastercard with the same structural advantages. ROIC above 25 percent. FCF margin above 50 percent. SBC roughly 3 percent of revenue. Same network-effect moat, slightly larger absolute scale.
Owner earnings yield around 3 percent. Passes 7 of 7 frameworks. /buffett/v/.
What about Apple, Meta, Amazon, Nvidia, Tesla?
The mega-tech names not on the above list:
Apple (AAPL): passes quality at 7 of 7 but valuation is the constraint. At owner earnings yield 2.9 percent, it falls below the 3 percent threshold. Passes 5 of 7 at the consensus level. /blog/apple-aapl-7-framework-analysis/.
Meta (META): passes most quality screens. Switching costs are weaker than Microsoft or Adobe (users can leave for TikTok). SBC is structurally higher than the others on the list. Passes 4-5 of 7 depending on the period. /blog/meta-platforms-7-framework-analysis/.
Amazon (AMZN): AWS is a wonderful business; the retail business is structurally lower margin. Capital intensity is high. ROIC has been below 15 percent in some recent years (heavy reinvestment). Passes 3-4 of 7 currently. /blog/amzn-amazon-7-framework-analysis/.
Nvidia (NVDA): speculative-tech category. Currently extraordinarily profitable but the customer concentration and competitive landscape make 10-year cash flow projection genuinely uncertain. Fails the predictability filter regardless of price. Passes 2-3 of 7. /blog/nvidia-nvda-7-framework-analysis/.
Tesla (TSLA): automotive economics with tech multiples. Heavy capex, narrow margins after price cuts, and competitive landscape now includes Chinese OEMs that match Tesla on cost. Fails on margins, valuation, and predictability. Passes 1-2 of 7. /blog/tesla-tsla-7-framework-analysis/.
The Munger Alibaba example
Charlie Munger publicly disclosed buying Alibaba (BABA) in 2021, calling it a wonderful business at a then-discounted price. The position underperformed (Chinese tech regulatory pressure was much harder than Munger anticipated), and Munger later acknowledged that he had underestimated the regulatory risk in China specifically.
The lesson is honest: even great investors can apply the right framework to the wrong jurisdiction. A wonderful business in a country where the regulatory environment is genuinely adversarial is not a wonderful investment. Munger's experience with Alibaba is a reminder that the "tech is value-investable" thesis applies to specific tech businesses in specific jurisdictions, not as a blanket rule.
What this is not
Value investing in tech is not a free pass to chase any AI-themed stock that has appreciated. It is not an argument that the recent run in mega-tech valuations is permanent. It is not a recommendation to buy a basket of tech names without doing the underlying work.
It is the recognition that the Buffett framework applies to any business with durable economics, regardless of sector classification. Some tech companies (a small minority) have durable economics. Most do not. The framework filters the few from the many.
Where invest-like fits
Every tech stock in the invest-like universe is scored against the same 7 frameworks. The per-stock breakdown is at /buffett/[ticker]/. The tech-specific composite filter is at /best/quality-composite/ filtered by technology sector.
For broader context:
Disclosure
Educational tool. The five tech names above pass our specific screens as of 26 May 2026 and may move between tiers as fundamentals evolve. Tech-sector volatility is higher than the broader market; drawdowns of 30-50 percent in individual tech names are common even in passing names. Past performance does not predict future returns.
Author: Zaid Ghazal, founder of invest-like, Kiel, Germany. Not a registered investment adviser. Full methodology at /methodology/.