the wedge in 4 min: what they build, why the lab can't.
cap table in 8: dilution through B + comp set (stainless $300M+ at <10 heads, contextual ai $80M).
then: '18 months. either you're top 5 when the inbound lands, or back at the lab.'
mission opens. cap table closes.
a $28M ARR fintech founder closed an ex-anthropic eng tuesday.
$1.05M TC at the lab. $650k + 1.2% at his shop. 38% cut.
12 minutes of pitch. zero of it was comp.
every founder watches senior engs leave for the labs.
nobody watches the reverse. 6 ex-anthropic and ex-openai engs reached out to peer founders this month wanting back into small co.
labs solved comp. they didn't solve mission.
your hiring pitch is the new TC offer.
the founders who win h2 2026:
structure your cap table so the top 10 stay through earnout.
talk to deal counsel before the inbound.
know which 3 engineers the buyer is actually paying for.
your retention agreements are the new term sheet.
4 frontier labs. 4 deals. 5 days.
none of them were technically acquisitions.
the M&A playbook for our class just rewrote itself in real time.
here's what it actually looks like now:
the founder math just inverted.
your ARR is not the asset anymore.
your top 10 humans are.
if they're not under 18-month retention agreements signed by the right counsel, you don't have a sellable company. you have a payroll.
my daughter turned 5 this month. some math hit me today.
when she has her first interview, openai will be older than i am now.
i wrote four posts on the M&A wave today. all about the next 18 weeks.
her career starts in 2043. our class is solving for the wrong window.
peer founder DM'd me this morning. 9 people. dev tools co. $1.8M ARR.
anthropic emailed him friday. no deck request. just "would you take a call this week."
his message to me was one line: "do i answer or do i wait."
the stainless comp from may 18 was $300M+ for under 10 heads. north of $30M per early engineer.
it's not an ARR multiple. it's neutralize-the-competitor money. 5 buyers are bidding for it right now.
on day 1 of openai's IPO the pool collapses to 1.
karpathy joined anthropic this month. his job: use claude to train claude.
every founder reading this as a talent war is missing the moment.
the labs just turned on recursive self-improvement.
be the dev tools layer they have to acquire. or get left behind.
if you're a senior ai eng reading this:
the highest-EV career bet on the board right now isn't joining a frontier lab.
it's being top-10 at a dev tools company the labs need to buy by end of 2027.
prestige is the booby prize this cycle.
3 senior ai engineers turned down frontier lab offers this month.
each one took a smaller bet.
here's the math they all ran.
and what it says about the next 18 months for our class:
the pattern they're all running:
the labs are mature. high prestige. high comp. dead equity.
the M&A window stays open 12-18 months. dev tools winners get bought at $200M to $500M.
eng #4 at the right shop beats eng #4000 at openai.
3 founders DM'd me this week. same question.
"sell to a frontier lab now or wait?"
openai files S-1. anthropic closes $30B at $900B. stainless went for $300M+.
buyer pool drops from 5 to 1 the day openai prints. h2 2026 is the last great AI M&A window.