What happened in 2024 that caused their operating expenses to increase so much?

Mostly a 356% increase in R&D:

  FY Ended December 31, in millions except percentage

                             | 2023 | 2024 | YoY  
  ---------------------------|------|------|------
  research and development   | $165 | $751 | 356%   
  sales and marketing        | $201 | $472 | 134%   
  general and administrative | $168 | $316 | 88%

And most of that increase came from a one-time charge from allowing employees to sell their RSUs. While not a cash cost for Figma, it was booked as an expense and allocated as follows:

                             | 2024   
  ---------------------------|------
  cost of revenue            | $25    
  research and development   | $463   
  sales and marketing        | $187   
  general and administrative | $184   
  total                      | $858   

If you subtract the one-time charge, you get:

                             | 2023 | 2024 (adj.) | YoY  
  ---------------------------|------|-------------|------
  research and development   | $165 | $288        |  75%  
  sales and marketing        | $201 | $285        |  42%  
  general and administrative | $168 | $132        | (21)% 
  total                      | $534 | $705        |  32%

How'd you manage to summarize these figures so nicely ? Are you good at accounting stuff or are you using some AI tool ?

Manual labor, going back to before LLMs were a thing. I just like reading S-1s.

https://news.ycombinator.com/item?id=27701617#27702669

https://news.ycombinator.com/item?id=26362443#26363190

https://news.ycombinator.com/item?id=25143334#25143800

In May 2024, they removed some of the vesting conditions on RSUs so employees could sell shares in a secondary offering.

From an accounting perspective, it was an $800 million stock-based compensation expense, though it didn't really cost Figma anything.

They received a $1B termination fee in 2023 when their acquisition by Adobe collapsed. They would be losing money in 2023 otherwise

If they were planning to IPO, they will have put in place a strategy to maximise the numbers for the year prior.

So the numbers for the year before IPO may have no relation to future.

Initially Tepid about AI. Didn’t want to upset their base like Adobe was (seemingly) doing. Look at that year’s Figcon for evidence. The keynote led with the new Figma, front loaded anything. He quickly moved past it and spent 90% of the rest of the time on non-consequential features.

Then that AI feature they highlighted was pulled off production because it was cloning iOS.

The AI heavy product dump we just got are lessons from that time.