Your brain is about 2 percent of your body mass and takes 20 percent of the energy your body produces. At rest it draws more than 60 percent of the glucose in your blood.1
That is the bill when nothing is wrong.
Achim Peters, Bruce McEwen and Karl Friston spent a 2017 review in Progress in Neurobiology arguing that the bill is what stress is. Their definition is narrow and worth holding: stress is uncertainty. Hours and workload enter only through it.
Uncertainty is expensive because resolving it takes cerebral energy, and the brain has first claim on the body’s supply.
Uncertainty is expensive because resolving it takes cerebral energy, and the brain has first claim on the body’s supply.
When the question stays open, the draw continues. Peters and his co-authors call what follows a persistent cerebral energy crisis, and the load it leaves surfaces as impaired memory, atherogenesis, diabetes, and the cardiac and cerebrovascular events that come after.
Now look at what you are holding.
Southeast Asia raised US$7.4 billion in the first half of 2026, more than double the same period last year. One Singapore data-centre operator, DayOne, took US$4.5 billion of it across two Series C rounds. Strip that company out and the region raised US$2.9 billion, below where it stood a year ago.
Fintech took US$685 million. Round counts fell to 127 from 153.
The headline says the money came back. Your cap table says something else, and the distance between those two facts is a question you have no date for.
That gap is the exposure. It is also the part the research grades honestly, which is where most coverage of this stops being useful.
Mika Kivimäki and Ichiro Kawachi put perceived job insecurity at a relative risk of 1.19 for coronary heart disease, with a confidence interval running from 1.00 to 1.42. Across work stressors generally, the excess risk sits between 10 and 40 percent. They also note the association is partly explained by poorer circumstances and worse risk-factor profiles among the insecure, which trims it further.
A lower bound touching 1.00 is not a five-alarm number, and you should not treat it as one.
What the mechanism does say is specific. The cost lands on the unresolved question, which makes the date the lever rather than the outcome.
A runway you have modelled to a week in March is a bounded problem the brain can stop re-pricing. The same runway left as a feeling is one it re-prices every night.
The adjunct with real evidence behind it is slow-paced breathing at resonance frequency, roughly six breaths a minute, usually run as heart-rate-variability biofeedback.
Vera Goessl and colleagues pooled 24 studies and 484 participants and found a between-groups Hedges’ g of 0.83 against control. Paul Lehrer’s group screened 1,868 papers, kept 58, and landed on a small-to-moderate effect, largest for anxiety, depression and anger, smallest for PTSD, sleep and quality of life.
Two things in that record deserve your attention before your Saturday does.
The outcome is self-report, so you are buying a change in how the week feels rather than a demonstrated change in coronary risk. And neither review found efficacy moderated by the number of sessions or weeks.
A flat dose-response is unusual, and it cuts in your favour. Twenty minutes appears to buy what an hour buys, which makes this one of the few interventions in this section that costs almost nothing to run.
Most of the founders who came through 2022 got a date before they got an answer.
A dated bad number is cheaper to hold than an undated one.
A dated bad number is cheaper to hold than an undated one, and your body has been quietly covering the difference all year.
Footnotes
-
The figures come from Peters, McEwen and Friston’s 2017 review, which builds them into the ‘selfish brain’ account: in scarcity, the brain supplies itself first and the body settles up afterwards. ↩