This is the post I avoided writing for a year. The financial postmortem at the honest cost of three shutdowns was uncomfortable. This one is harder. Three startup shutdowns over four years did real damage to my mental health, and the founder ecosystem talks about that damage in carefully laundered language that does not match the inside of my head during the worst weeks.

So here is the unlaundered version. What burnout looked like, what depression after MindWave looked like, what worked, what did not, and what I have built into the Gravity work pattern as a result. The first shutdown is in mental-health-platform-postmortem; the framework that came out of all three is in three startups, three shutdowns. This post is the personal layer underneath both.

What burnout actually looked like

Founder burnout in the popular telling is melodramatic. Insomnia at 3am, hands shaking, dramatic crisis. The real version is duller and more dangerous. It looked, for me, like a flatness that crept in over weeks. I would open my laptop and feel a low dread before any specific email. I would draft messages in my head and not send them. I would say I was fine because saying anything else required a kind of emotional bandwidth I no longer had.

The first system that broke was sleep. Around two months before the MindWave shutdown, I started waking at 4:30am and not getting back. I told myself it was the early-morning founder grind. It was not. It was the body's stress system staying alert when the mind had stopped being able to process the load. Appetite went next. By the time I shut MindWave down in late 2022, I had lost roughly six kilograms without trying.

The hardest signal to see was social shrinkage. The friend group that had been six or seven people quietly compressed to two. I was not avoiding anyone; I was just not initiating. The maintenance cost of social ties had become unaffordable on the budget I had left. The warning sign for any founder is "I would still call them, but I cannot bring myself to text first". That is the social half of the depression I did not name at the time.

Shutdown as grief

Shutting down a company you built passes through stages that map closely onto the Kubler-Ross model for grief: denial, bargaining, anger, sadness, and eventually acceptance. The MindWave shutdown took me about four months to clear. Vibe AI took two weeks. The compression is not because the loss got smaller; it is because the recovery loop got faster.

The denial stage looks like "one more pivot will fix this". The bargaining stage looks like "if I can just hit X paying users". Anger is the night you realise the assumptions in the deck were optimistic. Sadness is the week after you tell users. Acceptance is the moment you can talk about the company in past tense without flinching. Skipping any stage does not work. The stage gets you eventually; you only choose whether to face it now or later.

Weeks to recovery, by shutdown MindWave (2022) ~16 weeks (4 months) Super AI (2024) ~9 weeks Vibe AI (2026) ~2 weeks Practice does not erase grief; it shortens the recovery loop. The four practices below are the reason this trends down.
Self-reported time from shutdown decision to working on the next thing without flinching. The line is not linear because life is not.

The four practices that worked

Four practices held up across all three shutdown recoveries. They are not original. They are not glamorous. They are also not interchangeable; each one fixes a thing the other three cannot.

Sleep first

Sleep is the precondition for every other practice working. Eight hours, dark room, phone in another room, no screens after 11pm. Founder culture's "sleep when you're dead" attitude is not strategic; it is a way of pre-paying for collapse. The CDC's sleep-and-mental-health guidance and Walker's "Why We Sleep" both make the point bluntly: sleep deprivation degrades emotional regulation faster than it degrades cognition, which is why burnt-out founders feel before they think they are failing.

Sport second

Sport flushes the stress chemistry that talk-therapy alone cannot reach. For me it is squash, three or four times a week, plus a long walk daily. The point is not the workout; it is the physiological reset. The American Psychological Association's exercise-and-mood evidence is consistent: regular moderate exercise reduces depressive symptoms by a comparable margin to first-line interventions. The dose is real; the floor is "move every day".

Journaling third

Journaling externalises the loop in your head so you can see it. I keep a single text file, undated paragraphs, no audience, no editing. The simple act of writing "I think X is true and here is why" stops X from running on a loop unchallenged. The point is the writing, not the reading; I rarely re-read entries. The friction of having to type the thought out is what kills the rumination.

Friends fourth

Friends are the load-bearing wall when the rest of the building shifts. Two or three people who know what you are working on, who do not need a status update, and who will sit with you when you cannot articulate what is wrong. Founder communities are useful, but they are not a substitute for the friend who knew you before you had a startup. Maintain those relationships actively, especially when you have no bandwidth, especially when the work is going badly.

On therapy

I started therapy six months after the MindWave shutdown, after a friend more senior in the founder world told me bluntly that founders are running mental loads that are not normal and the maintenance cost is non-negotiable. That conversation changed how I think about it. Therapy is not a luxury, not a weakness, and not a sign that anything is broken; it is a professional consultation in the same category as your accountant or your lawyer.

The WHO estimates that depression affects roughly five percent of adults globally; for founders the rate is materially higher across multiple academic surveys. If your accountant tells you a quarter is bad, you call them more often, not less. Mental health works the same way. Find a therapist whose mode you respect, see them when things are good so you have a relationship before you need one, and treat the cost as a line item. Indian founders should know that decent therapy in Bangalore is in the 1500-3000 INR per session range; the cost is small relative to the next bad decision you avoid.

The signals to take a break

Three signals to know when to step away. First, sleep degraded for more than ten consecutive nights. Second, small decisions feel impossibly large. Third, you start drafting Slack messages and not sending them. Any one of those is a yellow flag. Two together is red. Three is non-negotiable; take three days fully offline before the loop tightens further.

The break does not need to be a holiday. It needs to be three days where the laptop does not open, the social media apps are off the home screen, and the schedule is replaced with sleep, food, and movement. Most founder burnout escalations I have seen, in myself and in friends, came from skipping three-day micro-breaks until the body forced a thirty-day break.

What Gravity does differently

Gravity is the first build where I have written the four practices into the work pattern itself. Sleep is non-negotiable; squash three times a week is on the calendar; journaling happens daily; friends get a phone call weekly. The economics post at economics of bootstrapped AI agents explains why a bootstrapped solo founder cannot afford a burnout cycle: the cash runway is short, and the only renewable resource is the founder.

If you are a founder reading this in the middle of your own version of the MindWave four-month recovery, the only useful thing I can say is that it ends. The grief is finite. The four practices are not optional. The therapy line item is worth it. And nobody is keeping score on whether you took the week off; the only person who notices is the one you have to keep alive to ship the next thing.

If you want to compare notes, my email is at the top of /contact. I read everything that lands there.

Frequently asked questions

What does founder burnout actually look like?

For me, it looked like flat affect, a permanent dread when I opened my laptop, and a slow shrinking of the social world to two or three people. Sleep degraded first, appetite second, exercise third. The work output stayed roughly normal until it cliffed. Burnout is not loud; it is the slow erosion of every system that previously made you function.

Is shutting down a startup like grief?

Yes, structurally. Shutting down something you built passes through identifiable stages, similar to the Kubler-Ross model for loss: denial, bargaining, anger, sadness, and eventually acceptance. The first shutdown took me about four months to clear. The third took two weeks. Practice does not eliminate the grief; it shortens the recovery.

What four practices actually worked?

Sleep first, sport second, journaling third, friends fourth. Sleep is the precondition for everything else. Sport flushes the stress chemistry that talk-therapy alone cannot reach. Journaling externalises the noise so you can see it. Friends are the load-bearing wall when the building shifts. Skip any one of the four and the other three start failing.

Should founders see therapists?

Yes if you can afford it, and most founders can. Therapy is not a luxury or a sign of weakness; it is a professional consultation in the same category as your accountant or your lawyer. The MindWave shutdown made me a believer. Founders are running mental loads that are not normal, and the maintenance cost of that load is real.

How do you know when to take a break?

Three signals: when sleep degrades for more than ten consecutive nights, when small decisions feel impossibly large, and when you start drafting Slack messages you do not send. Any one of those is a yellow flag. Two together is a red flag. Three is non-negotiable. Take three days fully offline before the loop tightens further.

Three takeaways before you close this tab

Sources