The LEDs may well last that long. The electronics providing the correct voltage/current to the LEDs, not so much.
On the LEDs, the industry accepted method for measuring lumen maintenance of LEDs sets an "upper reporting bound" on LED life estimates at 6x the length of test data, since we aren't sure whether the prediction model will hold true over the longer term. Either you report L70 as the time when output drops below 70% of original, or you report it like "L70 > 60,000 hours" (if you had 10,000 hours of testing and it was predicted to stay above 70% at 60,000 hours).
Dyson is claiming 154000 hours, implying that they have over 25000 hours of test data. This is very light on supporting data, so I'm naturally skeptical of it. It's not uncommon to see people ignoring the 6x upper reporting bound because they want to publish a big number that may or may not actually mean anything. Dyson also don't mention whether they're talking about L70, L90, or some other criteria. L70 is typical, so I assume it's that.
I'd expect to see a INTMT test and their LM-79 lumen maintenance calculation, based on the temperature data and calculated in EnergyStar's freely available LM-79 spreadsheet. For high-end lighting products, this information gets published.