What holds a hundred billion stars together across a hundred thousand light-years?

A galaxy is a gravitationally bound system — stars, gas, dust, stellar remnants, and enormous quantities of dark matter all orbiting a common centre. Most large galaxies anchor a supermassive black hole at their core. These black holes consume matter so violently that, when active, they outshine their entire stellar population by factors of hundreds.

The word comes from the Greek galaxias — milky. The Greeks named it after spilled divine milk. They were not far wrong. A galaxy is a nursery. Stars are born inside it, burn through their fuel, die in supernovae, and seed the surrounding gas with heavier elements. Carbon. Oxygen. Iron. The materials of planets. Of people.

Galaxies come in three broad shapes. Spiral galaxies — like our Milky Way and Andromeda — rotate, trailing arms of active star formation. Elliptical galaxies are rounder, older, quieter. Their star-forming days are largely over. Irregular galaxies have no clean shape. They are often the aftermath of collisions.

But these categories flatten a stranger reality. There are lenticular galaxies, caught between spiral and elliptical. Dwarf galaxies, small enough to orbit larger hosts — the Milky Way herds dozens of them. Ring galaxies, where a smaller galaxy has punched through a larger one and sent a shockwave of star birth rippling outward. And active galaxies — quasars, blazars, Seyfert galaxies — where the central black hole feeds with such violence that the nucleus alone blazes across the observable universe.

Galaxies are not static. They collide, merge, shred smaller companions, and consume them over timescales that make civilisation look like a spark.

Can you hold two trillion in your mind? Not as a number — as a count?

For most of the twentieth century, the estimate was 200 billion galaxies in the observable universe. That was already beyond imagination. Then in 2016, Conselice's team used deep-field surveys and modelling to revise the figure to roughly two trillion. Most of these galaxies are too small and faint for current instruments to resolve directly. We know they exist because the mathematics of structure demands them.

Each of those two trillion galaxies contains, on average, hundreds of billions of stars. Each star may host planets. The arithmetic of potential life becomes, at this scale, something the body refuses to process.

Our own Milky Way is a barred spiral galaxy roughly 100,000 light-years across. Our Solar System sits in the Orion Arm — a minor arm, in the galactic suburbs — about 26,000 light-years from the centre. We orbit that centre once every 225 to 250 million years. This period is called the cosmic year, or galactic year. In the entire span of complex animal life on Earth, our planet has completed roughly one full lap.

The nearest large galaxy is Andromeda — Messier 31 — 2.537 million light-years away. It is approaching at approximately 110 kilometres per second. In roughly 4.5 billion years, Andromeda and the Milky Way will begin their first pass. Almost no individual stars will collide. Space inside galaxies is too vast for that. The two systems will pass through each other like smoke through smoke, their gravity pulling long tidal streams of stars from each other's bodies, until they merge into a single elliptical galaxy astronomers have already named Milkomeda.

The Sun will likely survive. Nothing about its position in the galaxy will be recognisable.

What did the universe look like before any of this existed?

Go back 13.8 billion years. The universe was a hot, dense plasma — nearly uniform in all directions, nearly featureless. The word "nearly" is everything. Tiny quantum fluctuations, vanishingly small variations in density, were written into that plasma from the first instants. Gravity amplified them. Denser regions pulled in more matter. Less-dense regions emptied out. Over hundreds of millions of years, this process of gravitational collapse and hierarchical assembly produced the first structures.

The first stars — Population III stars — were enormous. Possibly hundreds of times the mass of our Sun. They burned fast, burned hot, and died violently in supernovae that seeded surrounding gas with the first heavy elements. Their deaths may have seeded the first black holes, which grew, pulled in more gas, and began organising matter into the first proto-galaxies.

The Lambda-CDM model — the standard model of cosmology — predicts this happened gradually. Large galaxies assembled slowly over billions of years from smaller pieces. Small structures first, then larger ones, hierarchically. The model has been remarkably successful. It predicted the large-scale structure of the universe — the vast web of filaments, sheets, and voids — with precision confirmed by observation.

Then the James Webb Space Telescope launched on Christmas Day 2021. And the results were not entirely what the model predicted.

What do you do when the data disagrees with the model?

Webb was built to see the early universe. It observes in infrared wavelengths — necessary because the expansion of space stretches ancient light toward the red and infrared end of the spectrum. This is cosmological redshift. The higher a galaxy's redshift, the earlier in cosmic time you are seeing it.

Webb's first deep-field images, released in July 2022, were immediately historic. Thousands of galaxies in a patch of sky smaller than a grain of sand held at arm's length. Some of them existing within a few hundred million years of the Big Bang.

The distances alone were unprecedented. But it was the characteristics of these early galaxies that surprised the field. A paper published in Nature in 2023 described six candidate galaxies from Webb observations that appeared to contain as much stellar mass as the present-day Milky Way — yet existed when the universe was less than 700 million years old. According to the standard model, there should not have been enough time to assemble that much stellar mass that quickly.

Several prominent cosmologists have described this as a genuine challenge, not a measurement error to be smoothed over. Photometric redshifts — inferred from colour rather than direct spectroscopy — carry uncertainties, and spectroscopic follow-up will sharpen or revise these estimates. But the general direction holds. The early universe was more active, more structured, and more productive of large galaxies than the Lambda-CDM model predicted.

Something in our models is incomplete. Whether the resolution requires revising dark matter physics, rethinking early star formation, or something more fundamental — that question is currently open. Webb is in the process of forcing us to answer it.

What if the majority of what holds everything together cannot be seen, touched, or identified?

The story begins with Fritz Zwicky, a Swiss astronomer at Caltech in the 1930s. He measured the motions of galaxies within the Coma Cluster and found they were moving far too fast. The visible mass — all those stars, all that gas — was nowhere near sufficient to hold the cluster together gravitationally at those speeds. Zwicky proposed unseen mass. He called it dunkle Materie — dark matter. The scientific community largely ignored him.

The case became undeniable in the 1970s. Astronomer Vera Rubin, working with Kent Ford, measured the rotation curves of individual spiral galaxies. In a simple gravitational system — planets orbiting a star, for instance — objects further from the centre orbit more slowly. Rubin found that stars in the outer regions of galaxies orbit at roughly the same speed as those near the centre, or faster. The only gravitationally consistent explanation was that galaxies are embedded in vast, invisible dark matter halos — roughly spherical distributions of non-luminous matter extending far beyond any visible disc, containing perhaps five to six times as much mass as all the stars and gas combined.

Dark matter does not emit, absorb, or reflect electromagnetic radiation. It interacts with ordinary matter only through gravity. Its physical nature is unknown. Proposed candidates include WIMPs (Weakly Interacting Massive Particles), axions, and sterile neutrinos. Decades of direct detection experiments have returned null results. Some researchers have proposed modifying gravity itself — MOND (Modified Newtonian Dynamics) — to eliminate the need for dark matter. MOND has predictive successes at the scale of individual galaxies but fails at the scale of galaxy clusters.

Then there is dark energy. In 1998, two independent teams measuring distances to Type Ia supernovae discovered that the expansion of the universe is not slowing down. It is accelerating. Galaxies are not just moving apart — they are moving apart faster and faster, driven by an agent we call dark energy. Those teams received the 2011 Nobel Prize in Physics. Dark energy constitutes roughly 68% of the total energy content of the universe. Dark matter accounts for approximately 27%. Ordinary matter — everything visible, everything measured, everything built and eaten and remembered — is roughly 5%.

Did pre-telescopic astronomers know more than we have assumed?

The Milky Way appears in the mythologies of virtually every culture that lived under dark skies. In ancient Egypt it was associated with the goddess Nut — the sky herself, arching her body over the earth. In the Hindu tradition it is Akashaganga, the Ganges of the sky. Many Indigenous Australian traditions contain astronomical knowledge encoded in oral tradition across tens of thousands of years, including detailed attention to dark constellations — shapes defined not by stars but by the dark dust rifts in the Milky Way, silhouettes of animals and ancestors visible only under a truly dark sky. The Inca mapped similar dark-lane figures using the same technique — a sky-watching system entirely distinct from the star-centred approach of European and Middle Eastern traditions.

The Dogon people of Mali have attracted particular attention for their apparent pre-telescopic knowledge of the Sirius star system — including Sirius B, a white dwarf invisible to the naked eye, which they appear to have encoded in cosmological traditions. Legitimate scholarly debate persists about whether this knowledge predates Western contact. The case is not settled. But it points toward something worth holding: that pre-telescopic astronomical attention may have been more precise, and more widely distributed, than we have been willing to assume.

Every tradition that looked up intuited structure, relationship, and meaning in the arrangement of what it saw. The modern picture — two trillion galaxies, dark matter halos, accelerating expansion, supermassive black holes — does not erase that intuition. It expands its object beyond any prior imagination. The structure is real. The mystery is real. The only question is what relationship human consciousness bears to either.

Is a galaxy an object, or a process?

A galaxy behaves less like a fixed structure and more like a self-regulating system — one with something resembling metabolism.

Star formation is the galaxy's primary creative act. Giant molecular clouds — dense, cold regions of interstellar gas and dust — become gravitationally unstable and collapse. New stars ignite inside them. The Milky Way today forms roughly one to two new stars per year. In the early universe, starburst galaxies were forming stars at rates thousands of times higher, blazing with the light of newborn suns across distances Webb can now see.

But star formation is self-limiting. Massive stars die in supernovae that blast energy back into surrounding gas, heating it and suppressing further collapse. The supermassive black hole at a galaxy's centre plays the same regulatory role when active. As an AGN (Active Galactic Nucleus) or quasar, it drives jets and winds capable of heating and ejecting gas across tens of thousands of light-years — shutting off its own fuel supply in the process.

The result is galactic metabolism: gas cools, collapses, forms stars, energetic feedback heats the gas, star formation slows, gas eventually cools again. Large elliptical galaxies — red, old, largely inert in terms of star formation — are thought to have had this cycle quenched by AGN feedback early in cosmic history.

Galaxy mergers are another evolutionary engine. When two galaxies interact gravitationally, tidal streams of stars stretch for hundreds of thousands of light-years. Gas clouds compress and ignite into bursts of star formation. Eventually the central black holes of both galaxies sink toward each other and merge — an event that releases gravitational waves detectable across the observable universe.

The Milky Way has eaten smaller galaxies throughout its history. The Sagittarius Dwarf Galaxy is being disrupted and absorbed right now, its stars spread in long streams around our own disc. The chemical and kinematic structure of the Milky Way's stellar populations carries the signatures of multiple past mergers. The galaxy we live in is made, in part, of galaxies that no longer exist as such.

Carl Sagan observed that we are made of star stuff — the nuclear ash of ancient supernovae, assembled by a galaxy's gravitational ecology into temporary structures capable of wonder. That is not metaphor. It is nuclear physics. The iron in blood and the calcium in bone were forged in stellar cores and scattered by their deaths. A galaxy did not produce us incidentally. We are, in the most literal chemical sense, how the galaxy recombines itself.

Here is the only honest response to scale this extreme: stop waiting for permission to think at this size.

The data from Webb is already forcing cosmologists to revise models that held for decades. The revision is happening in real time. What JWST is demonstrating is that observation changes the frame — and frames that seemed permanent are not. The standard model of cosmology will be updated. Perhaps significantly. The people updating it are alive now.

Dark matter remains undetected in any laboratory. Dark energy remains unexplained. Two trillion galaxies remain mostly uninventoried. The Fermi Paradox — the silence of a universe statistically packed with potential civilisations — remains without resolution. These are not solved problems waiting for a textbook summary. They are live, contested, unfinished questions, and they belong to anyone willing to engage them seriously.

The ancient cultures that read dark constellations in the Milky Way were not waiting for telescopes. They were working with the attention they had, the skies they had, the traditions they had — and building cosmologies precise enough to last tens of thousands of years. Precision matters. Attention matters. Community matters. None of those require institutional authority.

Andromeda is approaching at 110 kilometres per second. Milkomeda will form in 4.5 billion years. The Sun will survive. In some deep and non-trivial sense, the outcome of that collision is partly ours — what kind of civilisation, what kind of awareness, what kind of relationship to the cosmos, will exist in the galaxy that emerges. That is either a paralyzing thought or a generative one.

Make it generative. Build the communities, the practices, the knowledge structures, and the relationships that can hold questions this large. Not because an institution told you to. Because the sky demands it.

Self-governance is the only answer. Build now.