Safran is a French multinational aerospace corporation that helps power more than half of the world's short- and medium-haul commercial flights, through an equal partnership with American aircraft engine supplier GE Aerospace. We bought shares in the spring of 2020, at the very start of our investment journey, when essentially all stocks were on sale amid the pandemic crash.
Why we bought it
We were drawing inspiration at the time from the lists of Obermatt, a Swiss-based analysis firm that simplifies financial metrics into ranks by comparing companies against their peers. Safran seemed to us a relatively safe and easy-to-understand engineering giant. The price paid was 14× trailing earnings, valuing the whole company at €36.5 billion.
There were three revenue streams at Safran in 2020, and they remain the same at the time of writing. These are aerospace propulsion, aircraft equipment (including defense) and aircraft interiors. The first two are key for the company and the third one is less relevant. In our view, it was unnecessary acquisition of a business where Safran lacked expertise. Our analysis focuses on the first two streams only.
In aerospace propulsion realm, new engines are sold at thin margins to win the airframe program. Then money is made over the following two to three decades on spare parts, overhauls, and long-term service contracts. It was and remains roughly half of Safran revenue with operating margin now standing at 23% - above the 20% which was at the time we bought pre-COVID. It appears that older aircraft were retired more slowly ensuring longer revenue streams for Safran. At the same time, new technology which you could argue was still economically unproved in 2020 is now also profitable.
The second sector of aircraft equipment brings in around 40% of revenue. It had and still has tighter margins but was boosted by European defense rearmament even if Safran is not really a defense stock. The company says its activities in defense are about 20% of 2025 sales. We did not and probably could not see in 2020 that Russia will invade Ukraine in two years providing a geopolitical tailwind for Safran. But how could we not see that Safran will keep its dominant position in propulsion?
Why we sold it
"The world has changed." Warren Buffett, on Berkshire Hathaway's exit from its four largest US airline holdings, 2020
As the year of 2020 progressed it became clear that the pandemic was there to stay paralyzing travel and grounding aircraft worldwide. News reached us that Berkshire Hathaway has sold all its shares in four largest US airlines with Warren Buffet himself saying at the time “the world has changed” because of the coronavirus. Work from home proliferated and proved that international meetings can be held online. Our reality was distorted and we investigated Safran financial statements and balance sheets looking for weak points.
Grounded aircraft do not accumulate flight hours. Fewer flights mean less wear and less maintenance. We feared grounding itself would persist longer and full recovery may not even come if “the world has changed.” The year before COVID Safran had 15.5% overall operating margin. Half of operating income was considered by company itself as free cash. We thought this buffer would not be enough to absorb negative changes, and free cash flow might even turn negative. The price of 14 trailing earnings for a company in an industry that might have just peaked suddenly appeared too expensive to us and we sold in November 2020 at roughly the same price we bought. It was a mistake to do so!
What actually happened
Yes, revenue did collapse by 33% during the pandemic. Operating margin dipped to 5%. But Safran stayed profitable. It cut its capital expenditure, suspended dividends and built-up cash reserves. There was no meaningful increase in debt. In 2025 revenue reached €31.3bn, roughly 27% above pre-COVID levels. Its overall operating margin now stands 16.6%, again – above the 2019 level of 15.5%. The stock went up 3.29 times compared to our exit price and did so without extreme valuation changes of its multiples. Seems the world has not changed after all!
| Metric | 2019 (pre-COVID) | 2025 |
|---|---|---|
| Revenue | Baseline | €31.3bn (+27% vs. pre-COVID) |
| Overall operating margin | 15.5% | 16.6% |
| Propulsion operating margin | 20% (at purchase) | 23% |
| Debt | — | No meaningful increase |
| Pandemic trough (2020–21) | — | Revenue −33%, margin as low as 5%, still profitable |
| Stock, vs. our November 2020 exit | — | 3.29× |
Mean reversion
Later in our investment journey we will learn about what Jeremy Grantham called extrapolation. Investors routinely assume that whatever times, margins, or technological trends are happening right now will continue forever into the future. In fact, prices and margins usually return to long-term historical averages. Grantham identified this trend as mean reversion. Investors must identify the mechanism behind it it to make sound decisions.
In Safran case, we could have and should have understood that coronavirus would not change its dominant position in engineering the world's most common narrowbody engine family or its long-term service contracts. The use of Safran services has experienced a brief shock, but it was not a loss of competitive position. It took decades for Safran to learn from its engineering, embed the lessons into processes, tools, suppliers, people and culture, so it can keep building propulsion better, faster and more reliably than others. Even China, which is currently pushing a national strategy of self-reliance, including in manufacturing aircraft, still relies on Western engines and components. Replacing these with fully domestic tech remains a major, long-term challenge. Learning is especially defensible when knowledge is complex, distributed across many people, reinforced by suppliers or customers, and continually advancing as is the case in propulsion.
Meanwhile, air industry itself had no alternative emerging to disrupt it and it was naïve to think the patterns of human behavior can change so quickly in favor of digital environments. We should have kept Safran in the team. With it, we would score 13032 at the end of 2025 rather than 12899 as was the case. We were only protected by relatively small position, because Safran came in very early in our investment journey.
Reviewed 2026, five years after the sale. The next review will come in 2031.