The concept of constructing a high-energy proton–proton collider inside CERN's underground infrastructure originated in the early 1980s, while the tunnel was being excavated for the Large Electron–Positron Collider (LEP). Feasibility studies began at the end of 1982, and the formal scientific case was presented at a joint workshop of the European Committee for Future Accelerators and CERN in Lausanne and Geneva in March 1984. Because protons are roughly 2,000 times more massive than electrons, a synchrotron accelerating protons in the same ring loses far less energy to synchrotron radiation than an electron machine, allowing the collider to reach multi-teraelectronvolt energies within the existing 26,659-meter circumference.

Following a decade of magnet research and design work, the CERN Council formally approved the Large Hadron Collider project in December 1994. Initially, council members authorized a two-stage construction plan due to budgetary constraints, before approving a single-stage build to reach full design energy after securing financial and technical contributions from non-member observer states, including the United States, Japan, Russia, India, and Canada. LEP was decommissioned in late 2000, and civil engineering began to dismantle the electron collider and excavate the massive underground caverns required for the new detector complexes, particularly the ATLAS and CMS experimental halls.¹²

Construction spanned from 1998 to 2008 and encountered substantial engineering and financial challenges. A 2001 project review identified significant cost overruns driven by the complexity of manufacturing the superconducting magnets, underground civil engineering difficulties caused by waterlogged ground during the excavation of the CMS cavern, and structural failures during initial testing of magnet supports in 2007. The total cost of constructing the accelerator reached approximately 4.6 billion Swiss francs, with an additional 1.16 billion Swiss francs contributed by CERN toward the construction of the detector experiments, making the facility one of the costliest civilian scientific instruments ever constructed.

The collider occupies a circular tunnel bored at depths ranging from 50 to 175 meters (164 to 574 feet) beneath the surface, crossing the border between the Canton of Geneva in Switzerland and the Ain department in France.¹³ The subterranean placement provides natural radiation shielding and structural stability within the bedrock, while surface buildings house the cryogenics plants, electrical substations, ventilation shafts, and control centers. Because electricity prices in Europe peak during winter months, the facility traditionally concentrates its high-intensity physics operations between spring and autumn, utilizing the winter stop for routine technical maintenance.

Particles do not enter the main ring from rest; instead, they pass through a succession of smaller machines known as the CERN injector complex.¹⁴ Since 2020, proton beams have originated in Linear Accelerator 4 (Linac4), an 86-meter underground facility that accelerates negative hydrogen ions—hydrogen atoms carrying an extra electron—to an energy of 160 megaelectronvolts (MeV).¹⁵ Upon entering the Proton Synchrotron Booster (PSB), the ions pass through a stripping foil that removes both electrons, leaving bare protons that are accelerated to 2 gigaelectronvolts (GeV).¹⁶ The protons then transfer sequentially into the Proton Synchrotron (PS), which raises their energy to 26 GeV, and the Super Proton Synchrotron (SPS), which accelerates them to 450 GeV before injecting them into the main LHC ring in two opposing directions. During heavy-ion runs, vaporized lead or lighter elements such as oxygen and neon are instead fed through Linear Accelerator 3 (Linac3) and the Low Energy Ion Ring (LEIR) before entering the PS and SPS.¹⁷

Inside the main tunnel, the two particle beams travel in separate parallel beam pipes maintained at an ultrahigh vacuum to prevent particles from colliding with stray gas molecules. Superconducting radiofrequency cavities boost the protons from their 450 GeV injection energy to a peak operational energy of 6.8 teraelectronvolts (TeV) per beam, a ramp-up process requiring approximately 20 minutes. At full energy, the particles travel at roughly 99.999999 percent of the speed of light, completing 11,245 laps of the 26.7-kilometer ring every second.¹⁸ Rather than forming a continuous stream, each beam is divided into more than 2,400 discrete packets, known as bunches, with each bunch containing over 100 billion protons.¹⁹

Guiding and focusing these beams around the ring requires thousands of superconducting electromagnets. The primary bending system consists of 1,232 main dipole magnets, each 15 meters long, which utilize niobium-titanium Rutherford cables to generate magnetic fields of up to 8.3 tesla.²⁰ An additional 392 main quadrupole magnets, each measuring 5 to 7 meters in length, keep the beams tightly focused, while specialized insertion magnets squeeze the proton bunches down to a width of less than 10 micrometers just prior to the four interaction points.²¹ To maintain superconductivity without electrical resistance, the magnet coils are cooled to 1.9 kelvins (−271.3 degrees Celsius), a temperature colder than deep space, using a closed-loop cryogenic distribution system filled with superfluid liquid helium.²²

The two counter-rotating beams cross paths at four main interaction points spaced around the circumference of the ring, each housed within an underground cavern containing a large particle detector.²³ Because the heavy subatomic particles produced in high-energy collisions decay almost instantaneously—in the case of the Higgs boson, within roughly $10^{-22}$ seconds—detectors cannot observe them directly.²⁴ Instead, the instruments operate like three-dimensional digital cameras, taking up to 40 million snapshots per second to measure the trajectories, momenta, and energies of the lighter, stable decay products through concentric layers of silicon trackers, electromagnetic and hadronic calorimeters, and muon chambers.²⁵

The collider first circulated proton beams on 10 September 2008 to international media coverage.³³ Nine days later, on 19 September 2008, during high-current powering tests of the main dipole circuit in Sector 3–4, a faulty soldered electrical bus connection between a dipole and a quadrupole magnet failed.³⁴ The resulting electrical arc punctured the cryogenic helium enclosure, releasing roughly two metric tons of liquid helium into the insulation vacuum.³⁵ The rapid expansion of warming helium triggered a wave of magnet quenches and generated mechanical shockwaves that damaged 53 superconducting magnets and contaminated the vacuum pipes with soot. Repairs, together with the installation of an enhanced quench-protection system and pressure relief valves across the entire ring, delayed operations by 14 months.

Beams returned to the accelerator in November 2009, and the first high-energy physics campaign, designated Run 1, commenced in March 2010 at a center-of-mass collision energy of 7 TeV (3.5 TeV per beam).³⁶ In 2012, the collision energy was raised to 8 TeV (4 TeV per beam), providing the dataset required to isolate the Higgs boson before Run 1 concluded in early 2013.³⁷ The facility then entered Long Shutdown 1 (2013–2015), during which engineers consolidated more than 10,000 superconducting splice interconnects to prepare the magnet rings for operations near their nominal 14 TeV design threshold.³⁸

Run 2 began in June 2015 at a record center-of-mass energy of 13 TeV (6.5 TeV per beam) and continued through late 2018, delivering substantially higher integrated luminosity than the inaugural run.³⁹ From late 2018 to 2022, the facility underwent Long Shutdown 2, a period prolonged by the COVID-19 pandemic. During this hiatus, CERN completed the LHC Injectors Upgrade—most notably connecting Linac4 to the booster complex—while the ALICE and LHCb collaborations replaced major portions of their tracking and readout architectures to handle vastly higher collision rates.

Run 3 began delivering physics collisions on 5 July 2022 at an increased center-of-mass energy of 13.6 TeV (6.8 TeV per beam).⁴⁰ Over four years of operation, including specialized heavy-ion campaigns with lead, oxygen, and neon nuclei in 2025, the collider exceeded its performance targets. Proton beams circulated for the final time in Run 3 on 27 June 2026, and operations formally concluded on 29 June 2026. Across its first three runs, the LHC delivered nearly 550 inverse femtobarns ($\text{fb}^{-1}$) of integrated luminosity—almost double its original baseline design target of 300 $\text{fb}^{-1}$ and equivalent to 50 times the combined dataset collected by all hadron colliders prior to its startup.

The crowning achievement of the Large Hadron Collider is the confirmation of the Brout–Englert–Higgs mechanism, proposed independently in 1964 by theorists Robert Brout, François Englert, Peter Higgs, Gerald Guralnik, Carl Hagen, and Tom Kibble to explain how gauge bosons acquire mass via electroweak symmetry breaking.⁴¹ On 4 July 2012, the ATLAS and CMS collaborations jointly announced the independent observation of a new neutral boson with a mass of approximately 125 GeV (more than 120 times the mass of a proton) at a statistical significance exceeding five standard deviations. Initially detected through its decays into pairs of photons and pairs of Z bosons, the particle was subsequently confirmed to possess the spin, parity, and coupling strengths predicted for the Standard Model Higgs boson, a discovery that led to the 2013 Nobel Prize in Physics for Englert and Higgs.

Beyond the Higgs sector, the collider has reshaped the taxonomy of composite subatomic particles bound by the strong nuclear force.⁴² Through late 2024, LHC experiments had discovered more than 75 previously unobserved hadrons, including over 50 conventional mesons and baryons as well as 23 exotic hadrons made of four or five quarks.⁴³ Beginning with the LHCb collaboration's observation of two pentaquark states in 2015, experiments identified numerous tetraquarks and pentaquarks containing charm quarks, including the first fully charmed tetraquark and double open-charm states, providing empirical benchmarks for quantum chromodynamics. In 2020, the TOTEM collaboration, combining its proton–proton scattering measurements with legacy data from the Tevatron's D0 experiment, also confirmed the existence of the odderon, a quasi-particle compound state consisting of an odd number of gluons.

During Run 2 and Run 3, the collider opened new avenues in neutrino and quantum physics. In March 2023, the FASER collaboration announced the first direct observation of neutrinos produced at a particle collider, identifying high-energy muon neutrino and antineutrino interactions at a significance of 16 standard deviations, a milestone shortly corroborated by the SND at LHC collaboration. In September 2023 and throughout 2024, the ATLAS and CMS collaborations reported the first observation of quantum entanglement between pairs of top and antitop quarks, demonstrating quantum correlations in the heaviest known fundamental particles at energy scales far exceeding prior optical or atomic experiments.⁴⁴ Precision measurements of rare decays, such as the decay of the strange B meson ($B_s^0$) into two muons observed by CMS and LHCb, have simultaneously placed stringent empirical constraints on supersymmetry and other extensions to the Standard Model, where no direct evidence of new fundamental particles has yet emerged.

Because the collider produces up to one billion proton–proton interactions per second inside its main detectors, recording every collision event is technically impossible, as raw uncompressed data rates exceed one petabyte per second.⁴⁵ The experiments rely on multi-tiered hardware and software "trigger" systems that evaluate collision geometries in real time within microseconds, discarding over 99.9 percent of routine background collisions and retaining only the small fraction containing signatures of interest. Even after this filtering, the detectors generate tens of petabytes of archival data annually.⁴⁶

To store, distribute, and analyze this output without overwhelming local infrastructure at Geneva, CERN established the Worldwide LHC Computing Grid (WLCG), launched in the early 2000s.⁴⁷ Organized in a hierarchical tier architecture connected by dedicated optical-fiber networks, the grid links the Tier-0 data center at CERN with 11 major Tier-1 national computing laboratories across Europe, North America, and Asia—such as Fermilab and Brookhaven National Laboratory in the United States—and more than 140 Tier-2 university and research facilities across over 33 countries. By pooling more than one million processor cores worldwide, the grid allows physicists in any participating institute to run distributed analysis jobs on global collision datasets.

Prior to its 2008 commissioning, the collider became the subject of public controversy and media speculation regarding the safety of high-energy particle collisions.⁴⁸ Critics raised concerns that concentrating multi-teraelectronvolt energies into subatomic volumes could generate stable microscopic black holes capable of accreting surrounding matter, hypothetical negatively charged clumps of strange quark matter known as strangelets, or vacuum phase transitions.⁴⁹⁵⁰ In March 2008, retired radiation safety officer Walter Wagner and Spanish writer Luis Sancho filed a lawsuit in the United States District Court for the District of Hawaii against CERN, the U.S. Department of Energy, Fermilab, and the National Science Foundation, seeking an injunction to halt the collider's startup.⁵¹⁵²⁵³ The federal court dismissed the suit in September 2008 for lack of jurisdiction, a ruling upheld on appeal in 2010 on the grounds that the plaintiffs lacked legal standing and failed to establish a credible threat of harm.⁵⁴

To address public concerns scientifically, CERN commissioned two independent safety assessments: an initial report in 2003 and an updated review in 2008 by the LHC Safety Assessment Group, whose findings were endorsed by an external committee of 20 scientists and the Division of Particles and Fields of the American Physical Society before publication in the Journal of Physics G. The reports concluded that the collider posed no conceivable danger, noting that nature routinely conducts equivalent or far higher-energy collisions when ultra-high-energy cosmic rays strike the atmospheres and surfaces of Earth, the Moon, and dense astronomical bodies such as white dwarfs and neutron stars, all of which have endured for billions of years without destruction.

Following the end of Run 3 at the close of June 2026, the Large Hadron Collider entered Long Shutdown 3 (LS3), a 47-month engineering and installation phase scheduled to run until 2030. The primary objective of this shutdown is the installation of the High-Luminosity Large Hadron Collider (HL-LHC, or HiLumi LHC), a major hardware reconfiguration designed to increase the machine's instantaneous luminosity by a factor of five beyond its original design value and raise its lifetime integrated luminosity by an order of magnitude, aiming for a total dataset of roughly 3,000 $\text{fb}^{-1}$ over the following decade.⁵⁵⁵⁶

The upgrade requires dismantling and replacing more than 1.2 kilometers of accelerator hardware on either side of the ATLAS (Point 1) and CMS (Point 5) interaction regions. Central to the project is the first large-scale deployment in a particle collider of inner-triplet quadrupole magnets wound from brittle niobium-tin ($\text{Nb}_3\text{Sn}$) superconducting coils rather than conventional niobium-titanium. These new quadrupoles generate peak magnetic fields of up to 12 tesla—roughly 50 percent higher than existing LHC magnets—allowing operators to focus the colliding proton beams into substantially smaller cross-sectional areas. To maximize the geometric overlap of these tightly focused bunches as they cross at an angle, 16 superconducting radiofrequency "crab cavities" constructed from high-purity niobium are being installed adjacent to ATLAS and CMS to apply a transverse electromagnetic kick that tilts the particle bunches immediately before and after collision.

Simultaneously, the ATLAS and CMS detectors are replacing their inner silicon tracking systems, upgrading their calorimeters, and installing high-granularity timing layers to distinguish up to 200 simultaneous proton–proton interactions occurring during every bunch crossing (a condition known as pile-up). While CERN's injector chain and fixed-target facilities are scheduled to resume operations gradually from mid-2028, the upgraded High-Luminosity LHC is slated to recommence physics collisions for Run 4 in 2030, extending the operational lifespan of the facility into the early 2040s.