Orchid
reads the
whole book.

Five cells, a courier, and a verdict that arrives by email.

On a Tuesday morning in a fertility clinic in California, an embryologist plucks five cells from a five-day-old embryo no larger than a poppy seed. The cells are dropped into a buffer, sealed, and handed to a courier. Forty-eight hours and a few thousand miles later, the sequence of nearly the entire human genome - all six billion base pairs, give or take - lands on a server in San Francisco. A few days after that, prospective parents log in to a portal and read something no parent has ever been able to read before: a near-complete genetic forecast of a person who does not yet exist.

That is the workday at Orchid. It is also the thing Orchid is selling. The company is small - around 58 people - and the price is, by most measures, expensive: $2,500 per embryo, layered on top of an IVF cycle that already runs north of $20,000. But Orchid has done what a long line of larger, better-funded labs could not, which is take preimplantation whole-genome sequencing out of the conference-poster phase and into a clinic you can actually phone.

The IVF gap.

Here is the awkward fact about modern reproductive medicine: the standard genetic test offered during IVF, called PGT-A, reads roughly 0.25% of an embryo's DNA. It counts chromosomes the way a postman counts envelopes - not what's inside, just how many. For a couple paying $20,000 to $30,000 a cycle, that is a remarkably thin file.

Carrier screening fills part of the gap by flagging diseases the parents already know to look for. But carrier screens cannot catch de novo mutations - the brand-new genetic typos that appear for the first time in the child and account for a meaningful share of severe pediatric disease. They also cannot weigh polygenic risk, the way thousands of small variants add up to a real-world tendency toward, say, schizophrenia or early-onset Alzheimer's.

Orchid's pitch is that this gap is not a small inconvenience. It is a clinical and moral void, sitting inside a $25 billion industry that has spent the last decade optimising everything except the actual data its patients want most.

Noor Siddiqui, and the unfashionable opinion.

Noor Siddiqui founded Orchid in 2019. She was a Thiel Fellow, a former machine-learning researcher at Stanford, and the daughter of a mother who lost much of her vision to retinitis pigmentosa - a genetic condition Siddiqui has spoken about often, and which informs the company more than any deck ever will.

Her unfashionable opinion, circa 2019: ordinary couples would pay real money to know what was in an embryo's DNA, and the technology to tell them was finally close. The fashionable opinion - held by most of bioethics, much of the genetics establishment, and a few op-ed sections - was that this kind of testing was either impossible, premature, or unwise. All three, in some readings.

Investors did not, on the whole, find this argument unwelcome. Refactor Capital and Village Global led a $4.5M seed in 2021. In late 2023, a $12M round closed with a cap table that reads like a strange dinner party: Anne Wojcicki of 23andMe, Dylan Field of Figma, Vitalik Buterin of Ethereum, the Harvard geneticist George Church, Instacart's Fidji Simo, Counsyl's Balaji Srinivasan. Total raised to date: $16.5 million.

Six years, in compressed form.

What "99 percent" actually means.

The number "99%" gets thrown around a lot in startup marketing. Orchid is one of the few companies where it is a literal measurement, not a metaphor. The Whole Genome Embryo Report sequences more than 99% of the embryo's genome from a standard five-cell biopsy. It reports on three things at once: monogenic disease risk for over 1,200 conditions, chromosomal abnormalities including microdeletions and microduplications too small for the legacy chip, and polygenic risk scores for a smaller, carefully bounded list of common diseases.

The thing on the left is what Orchid sells. The thing on the right is what most IVF clinics currently treat as standard care. Reasonable people may have feelings about this.

By the numbers, sort of.

The customers, as of writing, are largely U.S. IVF patients who find Orchid through their fertility clinic or directly online and ship cells in via the clinic's standard biopsy workflow. The company has not published patient counts, and we are not going to invent any. What is verifiable: the test exists, it is sold, and it is being shipped against in commercial fertility cycles - a sentence that was not true at the start of 2023.

Information, not eugenics.

The hardest part of writing about Orchid is resisting the gravitational pull of the word "designer babies." Orchid does not edit DNA. It does not select traits. It does not promise a smarter, taller, blonder anything. What it does is read the genome of embryos that already exist in an IVF cycle and report what is there. The decision-making sits, as it always has in IVF, with the parents and their physician.

That distinction will not satisfy everyone, and Orchid knows it. The company has leaned into genetic counselling for this reason; every report is delivered alongside conversations with board-certified counsellors. The model is closer to a radiology read than a recommendation engine, and the company seems intent on keeping it that way.

Orchid's stated mission is to give prospective parents the most complete genetic information possible before pregnancy begins - so that preventable disease is the result of an informed choice, not an accidental one. The vision: a world in which the worst monogenic diseases no longer arrive by lottery.

The price comes down. Then the question gets harder.

Sequencing cost has fallen roughly a million-fold in twenty years. Orchid's $2,500 price tag is high today, deliberately positioned at the premium end of an already expensive procedure. Siddiqui has said publicly that the goal is to make the test cheap enough that it stops being a luxury. If she succeeds, the awkward conversation will not be about Orchid. It will be about everyone who chose not to offer this information to their patients.

That is the bet, narrowed all the way down: that prospective parents, given the choice, will not want less information about the genome of their future child. They will want more. And that the standard of care will, eventually, catch up to what is possible.

Back to the Tuesday morning clinic in California. Five cells in a buffer. A courier. A portal. The embryo has not been implanted. The pregnancy has not begun. Somewhere on a server in San Francisco, the report is finishing. The parents will read it before they decide. That used to be science fiction. That is now a Tuesday.