Manufacturing · Time · Field Notes

The Long Now of Manufacturing

Jeff Bezos has pointed out that designing a single aircraft part can take on the order of a decade — from drawing board through field testing and back around the loop. So what else in the physical economy moves on these glacial clocks? We checked 15 domains. A decade, it turns out, is often the fast case.

“An aircraft part… you’re talking about something that takes years — on the order of ten years — from design to field testing and the whole entire loop.” — the premise, paraphrased from Jeff Bezos
15
manufacturing domains
47
named programs analyzed
15–35yrs
typical concept→deployment
93yrs
the longest (Three Gorges)
150
verified sources cited

The time ladder

Concept to deployment, longest first

Each bar measures the longest defensible span — from the first serious concept to the thing actually working in the field. Tap a category to isolate it.

0
10
20
30
40
50
60
70
80
90
years →
Three Gorges Dam
93yrs
Gotthard Base Tunnel
69yrs
SynCardia / Jarvik-7 TAH
36yrs
Gorgon LNG
35yrs
EUV lithography (ASML)
34yrs
Olkiluoto 3 (EPR)
34yrs
Europa Clipper
33yrs
James Webb Space Telescope
32yrs
Cassini–Huygens
29yrs
Pratt & Whitney GTF (PW1000G)
28yrs
HS2 (UK)
27yrs
F-22 Raptor
25yrs
California High-Speed Rail
25yrs
USS Gerald R. Ford (CVN-78)
24yrs
Chūō Shinkansen maglev
24yrs
NuScale SMR
23yrs
F-35 Lightning II
22yrs
Oyu Tolgoi copper
22yrs
Rolls-Royce UltraFan
20yrs
NGAD / F-47
20yrs
Columbia-class SSBN
20yrs
AD700 A-USC steam turbine
20yrs
Vogtle 3 & 4 (AP1000)
18yrs
Flamanville 3 (EPR)
18yrs
Mining: discovery→production
18yrs
NASA SLS
17yrs
AbioCor artificial heart
16yrs
New drug (industry avg)
15yrs
Medtronic DBS (Activa)
15yrs
GE9X engine
14yrs
Boeing 777X
14yrs
Airbus A350 XWB
13yrs
B-21 Raider
13yrs
USS Virginia (SSN-774)
13yrs
Kamoa-Kakula copper
13yrs
Boeing 787 Dreamliner
12yrs
Siemens SGT5-8000H turbine
11yrs
Ariane 6
10yrs
Keytruda (pembrolizumab)
8yrs
Sovaldi (sofosbuvir)
7yrs
SpaceX Falcon Heavy
7yrs
Toyota TNGA
7yrs
GE 9HA.01 turbine
7yrs
VW MEB / ID.3
5yrs
TSMC Arizona Fab 1
4yrs
Typical new car program
4yrs
N700S Shinkansen
4yrs
Where you start the clock changes everything — by decades. The same program is honestly “9 years” or “32 years” depending on the anchor (JWST: 25 vs 32; the geared turbofan: ~8 vs ~28; Europa Clipper: ~9 vs 27). These bars use the concept-to-maturity framing, because that’s the loop Bezos was describing — the full distance from idea to a thing that works.

The full field

Every program, with the receipts

Forty-seven named programs across fifteen domains, each timeline independently fact-checked against primary and authoritative sources. Filter by category, then dig into the phase-by-phase breakdowns.

Filter

Jet engines

Pratt & Whitney PW1000G (PurePower Geared Turbofan, GTF)

Aerospaceadjusted · high
28yrs
1980s concept → 2016 in service

The clock here depends heavily on where you start. The geared-fan concept traces to NASA Advanced Ducted Propulsor (ADP) wind-tunnel research in the late 1980s–1997 and the abandoned PW8000 (~1998). The actual production GTF program launched around 2001 using a clean-sheet core co-developed with MTU. First demonstrator ground run was November 2007 at West Palm Beach; first flight testing on a Boeing 747SP and Airbus A340 testbed followed in 2008. The first variant (PW1500G for the Bombardier CSeries/A220) earned Transport Canada type certification on February 20, 2013, and the PW1100G for the A320neo achieved FAA certification December 19, 2014. Entry into service was January 2016 (Lufthansa A320neo). So program-launch-to-EIS is roughly 15 years; first-run-to-EIS about 8–9 years; first-cert-to-EIS about 3 years.

Why it takes so long

A geared turbofan was a genuinely new architecture (fan drive gear system) requiring long technology-maturation before a product program could even start, which is why the concept research spans the 1980s–2000s. Once launched, the program still needed ~6 years from first ground run to first certification, then multiple variant certifications (PW1500G, PW1100G, PW1200G, etc.) each adding test time. Field maturity took even longer: in-service durability problems (combustor/turbine, and later powder-metal disk issues) meant the engine was not truly mature for years after EIS, illustrating that 'entry into service' is well before 'field maturity.'

  1. late 1980s–~2000 (10+ years of intermittent technology work)Concept / NASA ADP research and PW8000 precursor
  2. ~2001 to Nov 2007 (~6 years)Production GTF program launch to first demonstrator ground run
  3. Nov 2007 to Feb 2013 (~5 years)First ground run to first variant certification (PW1500G)
  4. 2013/2014 to Jan 2016 (~2-3 years)First certification to entry into service (A320neo)
  5. several additional years (durability fixes well past 2016)Entry into service to field maturity
Best-supported Roughly 14-15 years from technology demonstrator (ATFI first run, March 2001) to entry into service (Jan 2016), or ~28-30 years if counted from late-1980s NASA ADP concept research. The true production-program-to-EIS span is shorter (~5-6 years from ~2010 A320neo/PW1100G launch). Best supported: ~15 years demonstrator-to-EIS; ~8-9 years first-engine-run (2007) to EIS; ~3 years first-cert (2013) to EIS.

General Electric GE9X (Boeing 777X engine)

Aerospaceconfirmed · high
14yrs
2012 → ~2026 in service

GE announced studies of a more-efficient GE90 derivative, the GE9X, in February 2012. The first engine to test (FETT) completed its first run in April 2016; the engine first flew on GE's Boeing 747-400 flying testbed on March 13, 2018. Certification flight testing on the 777X aircraft began January 25, 2020. The GE9X received its FAA type certificate on September 25, 2020, after the cert program accumulated just under 5,000 hours and 8,000 cycles. Entry into service, however, is gated by the Boeing 777-9 airframe, which has slipped repeatedly; EIS is now targeted around 2026. That makes concept-to-cert about 8.5 years, but concept-to-EIS roughly 14 years because the engine sat certified for ~6 years waiting on the aircraft. GE has continued running engines to build maturity, accumulating on the order of 17,000+ hours/27,000+ cycles to reach field maturity at service entry.

Why it takes so long

The GE9X is the largest commercial jet engine ever built (~134-inch fan), pushing materials (CMC turbine components, ceramic-matrix composites, composite fan blades) and requiring extensive ground and flight test cycles for certification. Beyond certification, the dominant schedule driver was airframe coupling: the engine cannot enter service until the 777-9 is certified, and Boeing's delays stretched concept-to-EIS far beyond concept-to-cert. GE deliberately kept running test engines for years to ensure the engine is mature (low in-service disruption) at the airplane's eventual service entry.

  1. Feb 2012 to Apr 2016 (~4 years)Concept/announcement to first engine to test (FETT) run
  2. Apr 2016 to Mar 2018 (~2 years)FETT run to first flight on 747 flying testbed
  3. Mar 2018 to Sep 2020 (~2.5 years)Testbed flight + 777X flight test to FAA type certification
  4. Sep 2020 to ~2026 (~6 years of waiting, used for maturity build-up)Certification to entry into service (gated by Boeing 777-9)
Best-supported ~8.5 years concept-to-certification (Feb 2012 to Sept 2020); ~14 years concept-to-entry-into-service (Feb 2012 to Q1 2026)

Rolls-Royce UltraFan (geared turbofan technology demonstrator)

Aerospaceconfirmed · high
20yrs
2014 → 2030s (projected)

UltraFan illustrates the very front end of the clock: a clean-sheet geared, variable-pitch-fan architecture still in the demonstrator phase, not yet a product. Rolls-Royce unveiled the UltraFan concept publicly around 2014. The full-scale technology demonstrator's build was completed in December 2022, and it achieved first run and ran to maximum power (~85,000 lbf thrust class) during 2023, accumulating roughly 70 hours before that test campaign concluded. Rolls-Royce is resuming demonstrator ground testing in early 2026 and is pursuing a narrowbody-sized derivative (UltraFan 30) targeted to run by 2028. There is no committed product program or certified engine yet; entry into service is projected for the 2030s, with flight testing expected before 2030, and timing tied to an airframer launching a new aircraft. This shows that demonstrator-to-product can itself add a decade after a successful technology run.

Why it takes so long

UltraFan is a demonstrator, so its timeline highlights the technology-maturation portion of the clock that precedes any formal product launch. The architecture (power gearbox sized for very high thrust, composite fan system, advanced core) requires proving at full scale before a manufacturer or airframer will commit. Crucially, a new commercial engine cannot reach EIS until an airframer launches a new aircraft to carry it; without a launch platform, even a successful 2023 demonstrator run stretches concept-to-service into the 2030s. This makes it a clear example that 'concept' and 'program launch' are distinct milestones, often years apart.

  1. ~2014 to Dec 2022 (~8 years of technology development)Concept unveiled to demonstrator build complete
  2. Dec 2022 to 2023 (~1 year)Build complete to first run / run to max power
  3. 2023 to early 2026Demonstrator test pause to resumed ground testing
  4. flight test before 2030; EIS projected 2030s (gated by an airframer launching a new aircraft)Demonstrator to product launch / flight test to entry into service
Best-supported Concept-to-first-demonstrator-run: ~9 years (2014 concept to April/Nov 2023 first run / max power). Concept-to-EIS: a projection of roughly 16-20+ years (2014 to 2030s), uncommitted and contingent on an airframer launch; flight testing expected before 2030, with a UltraFan 30 ground-test target of 2028.

Commercial aircraft

Boeing 787 Dreamliner

Aerospaceconfirmed · high
12yrs
2003 → ~2014 mature fleet

Boeing formally launched the 787 program on 26 April 2004 with a 50-aircraft order from ANA, after defining the conventional 7E7 concept in January 2003 (replacing the abandoned Sonic Cruiser). Despite a planned 2008 entry into service, supply-chain and engineering problems delayed first flight to 15 December 2009 and first delivery to ANA to 26 September 2011 -- about three years late. The clock from formal program launch to first delivery was roughly 7.5 years.

Why it takes so long

The 787 was a clean-sheet airplane introducing simultaneous first-of-kind technologies: a one-piece composite barrel fuselage, more-electric architecture, and an unprecedented globally-distributed risk-sharing supply chain. Each novelty needed design, qualification, tooling, and certification, and the integration risk compounded. Outsourced subassemblies arrived incomplete, requiring rework at Everett; FAA certification of new materials/systems is rigorous; and ramping production plus fixing early in-service issues (notably the 2013 lithium-ion battery grounding) added years before the fleet matured.

  1. Jan 2003 - Apr 2004 (~1.3 yr)Concept definition (7E7 announced)
  2. Apr 2004 - Jul 2007 (~3.2 yr)Program launch to rollout
  3. Jul 2007 - Dec 2009 (~2.5 yr)Rollout to first flight (delays)
  4. Dec 2009 - Sep/Oct 2011 (~1.8 yr)First flight to first delivery/EIS
  5. 2011 - ~2014 (~3 yr)EIS to mature fleet (post-battery grounding, ramp)
Best-supported Launch-to-delivery: ~7.4-7.5 years (26 Apr 2004 -> 25 Sep 2011 contractual / 26 Sep 2011 ceremony). Concept-to-delivery: ~8.7 years (29 Jan 2003 7E7 announcement -> Sep 2011). Concept/launch-to-mature-fleet: ~10-12 years (through the 2013 battery grounding and recovery, to ~2014-2015).

Airbus A350 XWB

Aerospaceconfirmed · high
13yrs
2004 → 2018 full family

The 'A350' label first appeared in 2004 as an A330-derivative concept (shareholder authorisation 10 Dec 2004), but customers rejected it. Airbus redesigned it as the clean-sheet, wider-fuselage A350 XWB, branded at Farnborough in July 2006 and given board industrial launch on 1 December 2006. First flight followed on 14 June 2013, EASA type certification on 30 September 2014, and first delivery to Qatar Airways on 22 December 2014, with entry into service on 15 January 2015. The true clean-sheet clock (XWB launch to delivery) was about 8 years.

Why it takes so long

Even with composite-airframe experience carried over, the XWB was effectively a new airframe family (-800/-900/-1000) with a new wing, new Rolls-Royce Trent XWB engine, and a new wide fuselage cross-section. Designing, building, and certifying a new twin-aisle from scratch -- plus standing up production tooling and a supply chain, then flight-testing across multiple variants -- inherently spans the better part of a decade. Reaching a mature fleet required ramping output and rolling out the later -1000 variant (EIS 2018), extending the effective maturation to ~2018.

  1. Sep 2004 - 2006 (rejected)Original A350 (A330-derivative) concept
  2. 1 Dec 2006XWB clean-sheet launch (board approval)
  3. Dec 2006 - Jun 2013 (~6.5 yr)Launch to first flight
  4. Jun 2013 - Sep 2014 (~1.3 yr)First flight to certification
  5. Sep 2014 - Dec 2014 / Jan 2015 (~3 mo)Certification to first delivery/EIS
  6. 2015 - ~2018 (~3 yr)EIS to mature/full-family fleet (-1000 EIS 2018)
Best-supported ~8 years clean-sheet XWB launch-to-delivery (1 Dec 2006 -> 22 Dec 2014); ~10 years from the original Dec 2004 A330-derivative concept to first delivery; full two-variant family complete ~11 years after launch (A350-1000 delivered 20 Feb 2018), or ~13 years from the 2004 concept.

Boeing 777X (777-9)

Aerospaceconfirmed · high
14yrs
2013 → 2027 (target)

Boeing formally launched the 777X at the Dubai Airshow on 17 November 2013, targeting entry into service around 2020. Rollout occurred 13 March 2019 and first flight on 25 January 2020, but engine (GE9X) issues, a post-737 MAX tightening of certification standards, and a 2024 cracked thrust-link discovery repeatedly pushed the schedule. First delivery is now targeted for 2027 -- roughly a 7-year delay and about $15 billion in program charges -- making the launch-to-delivery span about 13-14 years versus an originally intended ~7.

Why it takes so long

The 777X pairs a derivative fuselage with effectively new major systems: an all-new composite folding wingtip wing and the largest commercial turbofan ever (GE9X), both of which required extensive development and certification. Post-MAX regulatory scrutiny dramatically lengthened FAA certification (system-safety reassessment, scrutiny of crew-alerting and software), and physical findings during flight test (thrust-link cracking grounded the test fleet for months) forced redesign and re-testing. The result is a case study in how new technology plus a stricter post-incident regulatory regime can roughly double the expected timeline.

  1. 17 Nov 2013Program launch (Dubai Airshow)
  2. Nov 2013 - Mar 2019 (~5.3 yr)Launch to rollout
  3. Mar 2019 - Jan 2020 (~0.9 yr)Rollout to first flight
  4. Jan 2020 - 2026/2027 target (~6-7 yr)First flight to certification/delivery (heavily delayed)
  5. Not yet reached (EIS targeted 2027)Delivery to mature fleet
Best-supported ~13.5 years launch-to-delivery: formal launch 17 Nov 2013 -> first delivery targeted 2027 (Boeing-anticipated, ~Q1 2027 to Lufthansa). Roughly 7 years later than the original ~2020 EIS target (a ~6-7 year originally intended span). The claimed "13-14 years" range is accurate.

Fighter & combat aircraft

F-35 Lightning II (Joint Strike Fighter)

Defenseconfirmed · high
22yrs
1994 → 2015 IOC

The JSF effort grew out of 1993 JAST/CALF concept studies; the X-35 vs X-32 Concept Demonstration contracts were awarded Nov 1996; Lockheed won the SDD (full development) contract on 26 Oct 2001; the production-representative F-35A first flew 15 Dec 2006; and the first variant (F-35B, USMC) reached IOC on 31 July 2015, with F-35A (USAF) IOC on 2 Aug 2016 and F-35C (USN) on 28 Feb 2019. Measuring from program launch (SDD award 2001) to first IOC (2015) is ~14 years; from first flight (2006) to IOC is ~9 years.

Why it takes so long

Stealth + sensor-fusion integration, a single airframe forced to meet three very different service requirements (CTOL, STOVL, carrier), and concurrency (building production jets while still flight-testing) drove repeated re-baselines. The program was re-baselined in 2011, slipping IOC from a planned 2010 to 2015. Software (millions of lines of code) and the mission systems were a dominant schedule driver.

  1. 1993-1996 (~3 yrs)Concept studies (JAST/CALF) to Concept Demonstration contract award
  2. 1996-2001 (~5 yrs)Concept Demonstration / X-35 fly-off to SDD (development) contract win
  3. 2001-2006 (~5 yrs)SDD / development to first flight of production-config F-35A
  4. 2006-2015 (~9 yrs)First flight through flight test to first IOC (F-35B)
Best-supported ~14 years from program launch (SDD award Oct 2001) to first-variant IOC (F-35B, Jul 2015); ~19-20 years from the 1996 demonstration-contract award to IOC; or ~21-22 years measured from JAST program inception (1994) to IOC

F-22 Raptor (Advanced Tactical Fighter / ATF)

Defenseconfirmed · high
25yrs
1981 → 2005 IOC

The Advanced Tactical Fighter (ATF) requirement emerged in the early 1980s. A 54-month Dem/Val phase produced the YF-22, which first flew 29 Sep 1990; the Lockheed team won the ATF competition on 23 Apr 1991, launching Engineering & Manufacturing Development (EMD). The EMD/production-representative F-22 first flew 7 Sep 1997, and the F-22A reached IOC on 15 Dec 2005. From program launch (EMD award 1991) to IOC is ~14.5 years; from first EMD-jet flight (1997) to IOC is ~8 years.

Why it takes so long

First-of-its-kind all-aspect stealth combined with supercruise and advanced avionics required maturing entirely new technologies. The 54-month Dem/Val fly-off, a lengthy EMD with extensive flight and operational test, and post-Cold-War budget pressure that stretched out funding all extended the schedule between the 1990 demonstrator flight and 2005 fielding.

  1. early 1980s-1986 (~5+ yrs)ATF concept/requirement studies to Dem/Val (start)
  2. 1986-1991 (~4.5 yrs; 54-month Dem/Val)Demonstration/Validation (YF-22 build & fly-off) to source selection
  3. 1991-1997 (~6 yrs)EMD (development) to first EMD-aircraft flight
  4. 1997-2005 (~8 yrs)First flight through flight/operational test to IOC
Best-supported ~14.5 years from program launch (EMD award, 1991) to IOC (Dec 2005); ~24-25 years from the 1981 ATF concept/requirement to IOC. Both framings are accurate; the 1991-to-2005 EMD-to-service figure is the most defensible single number.

B-21 Raider (Long Range Strike Bomber, LRS-B)

Defenseconfirmed · high
13yrs
2015 → ~2027–28 fielding

The Air Force released the LRS-B RFP in July 2014 and awarded the development (EMD) contract to Northrop Grumman in October 2015 (after a GAO protest was resolved in 2016). The first B-21 flew on 10 November 2023 at Plant 42, Palmdale. A low-rate initial production (LRIP) contract followed in January 2024. First delivery to Ellsworth AFB and entry into operational service are projected for ~2027-2028, implying ~12-13 years from contract award to fielding and ~4-5 years from first flight to operational service. Note: an exact official IOC date has not been publicly declared (the program is heavily classified).

Why it takes so long

A clean-sheet, low-observable strategic bomber with nuclear certification and deep classification. The Air Force deliberately used a mature-technology, digital-engineering approach and combined/accelerated test phases to compress the schedule relative to prior bombers, but stealth integration, nuclear surety, and production ramp still set a multi-year floor between first flight and operational fielding.

  1. 2014-2015 (~1.5 yrs)LRS-B RFP to development (EMD) contract award
  2. 2015-2023 (~8 yrs)EMD / development to first flight
  3. 2023-2024 (~2 months)First flight to LRIP contract
  4. 2023-2027/2028 (~4-5 yrs, projected)Flight test to first delivery / operational service
Best-supported ~12 years from EMD contract award (27 Oct 2015) to projected first fielding/IOC at Ellsworth AFB (2027); up to ~13 years if delivery slips to 2028. First flight (Nov 2023) to fielding is ~3-4 years (or ~5-6 years to full combat capability ~2029).

NGAD / F-47 (Next Generation Air Dominance, sixth-generation fighter)

Defenseconfirmed · high
20yrs
2014 → ~2030s IOC (proj.)

NGAD traces to the DARPA Air Dominance Initiative study (March 2014) and the Aerospace Innovation Initiative (AII) launched ~2015, which flew full-scale technology demonstrator X-planes from ~2020 (three flown by 2023). The EMD (development) contract was awarded to Boeing on 21 March 2025 with the designation F-47. The Air Force targets first flight in 2028 and operational capability around 2029-early 2030s, though senior lawmakers and analysts caution IOC could slip to the mid-2030s. From concept (2014-2015) to projected IOC (~2029-2035) is roughly 15-20 years; from EMD award (2025) to projected IOC is ~4-10 years.

Why it takes so long

A clean-sheet sixth-generation system (stealth, advanced propulsion/adaptive engine, family-of-systems with collaborative combat drones, and new mission systems) requires long upstream prototype/demonstrator maturation before EMD. Even with X-plane demonstrators already flown, EMD plus extensive flight and operational test, plus integration with autonomous wingmen, set a multi-year path to IOC; budget and industrial-base pressures add slip risk.

  1. 2014-2015 (~1 yr)DARPA concept study to AII X-plane demonstrator program launch
  2. 2015-2020 (~5 yrs)Demonstrator development to first demonstrator flight
  3. 2020-2025 (~5 yrs)Demonstrator flights to F-47 EMD contract award (Boeing)
  4. 2025-2029+ (~4-10 yrs, projected)EMD to projected first flight (2028) and IOC (~2029-2030s)
Best-supported ~14-20 years end-to-end: concept origin 2014 (DARPA study) / 2015 (AII) to projected IOC ~2029 (official target) or mid-2030s (realistic, per senior lawmaker/analyst projections). From Boeing EMD award (March 2025) to first flight (target 2028) is ~3 years; to IOC is ~4 years official / up to ~10 years if slipped to mid-2030s.

Semiconductors

EUV lithography development (ASML / industry consortium)

Semiconductorsadjusted · high
34yrs
mid-1980s research → 2019 mass production

Extreme ultraviolet (EUV) lithography went from foundational research to volume chip production over roughly two to two-and-a-half decades. Foundational research began in the mid-1980s when Japanese researcher Hiroo Kinoshita projected the first EUV images (building on earlier Russian multilayer-mirror work). ASML formally launched its EUV program in 1997, shipped the first alpha-demo prototypes in 2006, shipped a pre-production NXE:3100 to Samsung in 2010, delivered its first production-capable systems around 2013, and EUV finally entered high-volume manufacturing in 2019 at the 7nm logic node (the first EUV-made commercial chip shipped in Samsung's Galaxy Note10 that year). ASML invested more than 6 billion euros in EUV R&D over ~17 years and required ~$6 billion in equity from TSMC, Samsung, and Intel in 2012 to finish the program.

Why it takes so long

EUV is a fundamentally new physics regime: 13.5nm light is absorbed by virtually all materials and air, forcing the entire system (source, optics, masks) to operate in vacuum with reflective rather than refractive optics. Generating enough EUV power required a novel laser-produced tin-plasma source (firing ~50,000 tin droplets per second hit by a high-power CO2 laser), and the mirrors needed atomic-level flatness with multilayer coatings made by Zeiss. Each subsystem (light source, defect-free masks/pellicles, resists, optics) had to be invented and matured in parallel, requiring a multi-decade international consortium and billions in pooled capital before throughput, uptime, and defectivity reached economically viable levels.

  1. mid-1980s onwardFoundational research (first EUV images, multilayer mirrors)
  2. 1997ASML EUV program launch
  3. ~9 years (1997-2006)Program launch to first alpha-demo prototypes (shipped to imec and SUNY Albany)
  4. ~4 years (2006-2010)Alpha-demo to pre-production NXE:3100 shipped to Samsung
  5. ~3 years (2010-2013)Pre-production to first production systems
  6. ~6 years (2013-2019)Production systems to high-volume manufacturing (7nm node)
Best-supported ~22 years for ASML's program (1997 launch to 2019 high-volume manufacturing); ~33-34 years from mid-1980s foundational research (Kinoshita 1985-86) to 2019 HVM

TSMC Arizona Fab 1 (Phoenix, N4 leading-edge fab)

Semiconductorsconfirmed · high
4yrs
2021 → 2024 production

TSMC's first Arizona fabrication facility (Fab 1) measures the build clock from site groundbreaking to volume output. Construction began in April 2021; the 'first tool-in' (equipment move-in) occurred in December 2022; and the fab reached high-volume production of N4 (4nm-class) chips in Q4 2024, with TSMC stating in January 2025 that yields were comparable to its Taiwan fabs. That is roughly three and a half years from groundbreaking to volume production for a leading-edge fab. The later Arizona fabs show similar spans: Fab 2 (N3) broke ground in 2022/2023 with equipment move-in 2026 and volume production targeted for 2027 (pulled forward a year by AI demand), and Fab 3 (N2/A16) broke ground in April 2025 targeting production by the end of the decade.

Why it takes so long

A leading-edge fab is one of the most complex construction and commissioning projects on earth. The clock spans bulk earthwork and a massive cleanroom shell, then installation and qualification of hundreds of multi-million-dollar tools (EUV scanners, deposition, etch, metrology) each requiring vibration-isolated foundations, ultrapure water, specialty gases, and precise environmental control. After physical construction, tool install, hookup, and yield ramp (achieving acceptable defect density and matching reference-fab yields) consume a large share of the timeline. Overseas builds also faced labor, permitting, and supply-chain frictions that initially pushed TSMC Arizona's first-production target from 2024 across iterations.

  1. April 2021Groundbreaking / construction start
  2. ~1.7 years (April 2021 - December 2022)Construction shell to first tool-in (equipment move-in)
  3. ~2 years (December 2022 - Q4 2024)Tool install and yield qualification ramp
  4. Q4 2024 (~3.5 years total from groundbreaking)High-volume N4 production reached
Best-supported ~3.5 years (construction start April 2021 to high-volume N4 production Q4 2024; First Tool-In equipment move-in December 2022)

Pharmaceuticals

Industry/Regulatory Benchmark — "New Molecular Entity" (PhRMA / FDA / Tufts CSDD composite)

Pharmaconfirmed · high
15yrs
discovery → approval (avg ~12 yrs)

The canonical '10-15 years and ~$2.6 billion' figure is a composite of two distinct authoritative sources, not one. The TIME figure comes from PhRMA and the FDA, who describe the path from initial discovery/target identification through FDA approval as averaging 10-15 years (PhRMA most often says 'at least 10 years on average'). The COST figure ($2.6B, in 2013 dollars) comes from the 2014 Tufts Center for the Study of Drug Development study led by Joseph DiMasi, based on 106 randomly selected drugs from 10 companies first tested in humans 1995-2007. The clock here measures from earliest discovery research to first FDA marketing approval; it does NOT measure a single program's stopwatch but a statistical average across many candidates including the cost of failures.

Why it takes so long

The duration is structural, not bureaucratic delay. (1) High attrition: only ~10-12% of candidates entering Phase I are ever approved, so the '$2.6B per approved drug' figure is dominated by the sunk cost of the ~88-90% of failures — out-of-pocket cost is $1,395M but capitalized cost rises to $2,558M because of the 10+ year time-value/opportunity cost of capital ($1,163M). (2) Sequential gating: each phase must complete and be reviewed before the next can begin (discovery 3-6 yr; preclinical animal/tox; Phase I safety 1-2 yr in 20-100 people; Phase II efficacy/dosing in hundreds; Phase III confirmatory in thousands lasting 1-4 yr; then a ~6-12 month FDA NDA/BLA review). (3) Biology and statistics: long-term safety and efficacy signals require large patient populations followed over years, which cannot be compressed without losing statistical power. The Tufts capitalized cost is roughly double the out-of-pocket cost specifically because money is tied up for over a decade.

  1. 3-6 yearsDrug discovery / target ID & lead optimization
  2. 1-3 years (often counted within early development)Preclinical (in vitro + animal toxicology, IND-enabling)
  3. 1-2 yearsPhase I clinical (safety, dose, PK; 20-100 subjects)
  4. ~2 yearsPhase II clinical (efficacy, side effects; up to several hundred patients)
  5. 1-4 yearsPhase III clinical (confirmatory efficacy/safety; thousands of patients)
  6. ~6 months (priority) to ~12 months (standard)FDA review (NDA/BLA submission to approval)
  7. 6-7 yearsClinical trials combined (Phase I-III)
Best-supported 10-15 years from discovery to first FDA approval (statistical average, including failures); ~12 years is the conventional midpoint, with the measured R&D clock (preclinical + clinical + review) running ~10.5-13.5 years

Keytruda (pembrolizumab) — Merck/Organon anti-PD-1 immunotherapy

Pharmaconfirmed · high
8yrs
2006 → 2014 approval

Pembrolizumab was invented by Organon scientists around 2006 (Organon was absorbed into Schering-Plough, then Merck via the 2009 merger). Clinical development (KEYNOTE-001) began in 2010-2011, it received FDA Breakthrough Therapy designation, and won accelerated approval for advanced melanoma in September 2014. The 'clock' here runs from molecule invention (2006) to first marketing approval (2014). It illustrates a compressed timeline: ~8 years rather than 10-15, achieved via an adaptive trial design and expedited regulatory pathways — a deliberately faster-than-typical case.

Why it takes so long

Even this accelerated case took ~8 years because the molecule sat partly idle after the Organon/Schering-Plough/Merck acquisition reshuffles before Merck prioritized it. The compression below the 10-15 year norm came from (1) FDA Breakthrough Therapy designation and accelerated approval, which let Merck approve on early surrogate endpoints, and (2) an adaptive KEYNOTE-001 trial that combined dose-finding and efficacy cohorts rather than running fully sequential Phase I/II/III studies. Where standard programs gate each phase, this one overlapped them.

  1. ~2006Discovery / invention (at Organon)
  2. 2006-2010Corporate transitions, dormancy, IND prep (Schering-Plough then Merck)
  3. 2010/2011-2014Clinical development KEYNOTE-001 (adaptive Phase I/II/III)
  4. 2013-Sept 2014Breakthrough Therapy designation to accelerated approval
Best-supported ~8 years: molecule invention/discovery 2006 (Organon) to first FDA accelerated approval September 4, 2014 (advanced melanoma). Sub-milestones: IND late 2010, phase 1 KEYNOTE-001 Jan 2011, Breakthrough Therapy 2013.

Sovaldi (sofosbuvir) — Pharmasset/Gilead hepatitis C cure

Pharmaconfirmed · high
7yrs
2007 → 2013 approval

Sofosbuvir was discovered in 2007 by Michael Sofia at Pharmasset. Animal toxicity began May 2008, Phase I in March 2009, Phase II in 2010, and Phase III after Gilead acquired Pharmasset in January 2012 for $11.2 billion. Gilead filed the NDA in April 2013, got Breakthrough Therapy designation in October 2013, and the FDA approved Sovaldi on December 6, 2013. The clock here runs from chemical discovery (2007) to FDA approval (2013) — roughly 6-7 years, one of the fastest modern timelines for a novel small molecule.

Why it takes so long

This is the fast tail of the distribution, and its speed is instructive about why the average is slower. Sofosbuvir was fast because (1) hepatitis C trials read out quickly — the endpoint (sustained virologic response, effectively a cure) is measurable in weeks-to-months rather than the years needed for survival or chronic-disease endpoints; (2) efficacy was so dramatic (near-100% cure) that trials needed fewer patients and shorter follow-up to reach statistical significance; and (3) Breakthrough Therapy designation expedited review. Most drugs lack such a clean, fast, high-effect-size endpoint, which is precisely why the typical program needs the full 10-15 years to accumulate enough events for statistical confidence.

  1. 2007Discovery (Pharmasset, Michael Sofia)
  2. 2008Preclinical / animal toxicology
  3. 2009Phase I
  4. 2010-2011Phase II
  5. 2012-2013Phase III (post-Gilead acquisition Jan 2012)
  6. Apr 2013 - Dec 6, 2013NDA filing to FDA approval (Breakthrough Therapy)
Best-supported ~6.5 years (chemical discovery 2007 to FDA approval December 6, 2013); roughly 2 years from Gilead acquisition close (Jan 2012) to approval

Nuclear power

Vogtle Units 3 & 4 (AP1000, Georgia, USA)

Energyadjusted · high
18yrs
2006 → 2023/24 operation

The first new US nuclear units in decades. Southern Nuclear filed for an Early Site Permit in August 2006 and a Combined Construction and Operating License (COL) in March 2008; the NRC issued the COL in 2012 and safety-related concrete pouring (the formal 'construction start') began in March 2013. Unit 3 connected to the grid on 31 March 2023 and reached commercial operation 31 July 2023; Unit 4 connected 6 March 2024 and entered commercial operation 29 April 2024. From COL application to both units operational was ~16 years; the official 'construction' clock (first nuclear concrete to commercial operation) was ~10-11 years per unit.

Why it takes so long

Two AP1000 first-of-a-kind units suffered supply-chain and quality problems (sub-standard rebar/concrete, design changes during construction), the 2017 bankruptcy of lead contractor Westinghouse, and the loss of its EPC structure. Costs roughly doubled from ~$14B to over $30B and schedule slipped ~7 years versus the original 2016/2017 targets. Licensing, first-of-a-kind engineering, skilled-labor shortages, and the capital intensity of a multi-billion-dollar single asset all compound delay.

  1. ~4-6 yearsEarly site permit / licensing application (ESP 2006, COL 2008) to COL issuance (2012)
  2. ~10 yearsFirst nuclear concrete (Mar 2013) to Unit 3 grid connection (Mar 2023)
  3. ~4 monthsUnit 3 grid connection to commercial operation (Mar-Jul 2023)
  4. ~10.5 yearsUnit 4 first concrete (Nov 2013) to commercial operation (Apr 2024)
Best-supported ~16-18 years overall: ESP application Aug 2006 / COL application March 2008 to both units in commercial operation (Unit 3: 31 July 2023; Unit 4: 29 April 2024). Formal NRC construction clock (first safety-related concrete March 2013 to commercial operation) was ~10.3 years for Unit 3 and ~11.1 years for Unit 4.

Olkiluoto 3 (EPR, Finland)

Energyconfirmed · high
34yrs
1989 design → 2023 operation

Europe's first EPR. The construction license was granted and first concrete poured in 2005, with commercial operation originally promised for ~2009-2010. Instead, OL3 was connected to the grid only on 12 March 2022 and began regular commercial operation on 1 May 2023 — about 18 years after construction start and roughly 13 years late. Construction alone, planned at 56 months, took ~199 months (about 3.5x). The underlying EPR design effort began with Framatome/Siemens forming Nuclear Power International in 1989, so the design-concept-to-first-European-operation arc spans more than three decades.

Why it takes so long

OL3 was a first-of-a-kind fixed-price turnkey EPR with no recent build experience in Europe. It hit regulatory first-of-a-kind reviews, instrumentation-and-control (automation) software disputes with the regulator STUK, quality issues with concrete and forgings, contractor disputes (Areva/Siemens consortium vs TVO ending in multi-billion-euro arbitration), and a long, troubled commissioning phase (turbine and feedwater problems). FOAK engineering plus an unfamiliar supply chain and a strict independent regulator drove the overrun.

  1. ~8 years (design line)EPR conceptual/basic design (NPI founded 1989; basic design report 1997)
  2. ~17 yearsConstruction start / first concrete (2005) to grid connection (Mar 2022)
  3. ~14 monthsGrid connection (Mar 2022) to regular commercial operation (May 2023)
Best-supported Construction: Aug 2005 to commercial operation 1 May 2023 = ~17.7-18 years (grid connection 12 Mar 2022, ~13 years later than the original ~2009/2010 promise; construction planned at 56 months took ~199 months, ~3.5x). Design-to-first-European-operation: NPI formed Sept 1989 to May 2023 = ~34 years (30+).

Flamanville 3 (EPR, France)

Energyconfirmed · high
18yrs
2007 → 2025/26 operation

France's first EPR. Construction (first concrete) began 4 December 2007 with EDF promising commercial operation in 2012 after a planned 54-month build for ~€3.3B. The reactor was connected to the grid only on 21 December 2024, reached full power in December 2025, and entered commercial operation in May 2026 — more than 12 years late, with costs rising more than fourfold to at least €13.2B (and ~€19-24B including financing per the French Court of Auditors). Construction-start to grid connection was ~17 years.

Why it takes so long

Same EPR first-of-a-kind difficulties as Olkiluoto plus France-specific setbacks: anomalies in the reactor pressure vessel head and bottom carbon-segregation (ASN material-qualification dispute), faulty weld quality in the secondary circuit requiring repairs, and a long commissioning campaign under regulator ASN. Loss of nuclear-construction know-how after a long French build gap and the sheer capital scale compounded the schedule.

  1. ~17 yearsFirst concrete (Dec 2007) to grid connection (Dec 2024)
  2. ~12 monthsGrid connection (Dec 2024) to full power (Dec 2025)
  3. ~5 monthsFull power to commercial operation (May 2026)
Best-supported Construction (first concrete) 4 Dec 2007 → grid connection 21 Dec 2024 (~17 years); → full power 14 Dec 2025; → commercial operation 5 May 2026 (~18.4 years from start). So ~17 years to grid, ~18.5 years to commercial operation — about 12+ years behind the original 2012 promise.

NuScale Power Module (US small modular reactor, SMR)

Energyconfirmed · high
23yrs
2000 → 2022 certified (no plant yet)

The leading US SMR design illustrates how long even the design/licensing clock runs for a new reactor type. The technology traces to DOE-funded MASLWR research at Oregon State University starting ~2000; NuScale Power was founded in 2007. NuScale submitted its design certification application to the NRC in 2017, received the Final Safety Evaluation Report and Standard Design Approval in 2020, and the NRC certified the 50 MWe design on 29 July 2022 — the first SMR design certified in the US. An uprated 77 MWe variant received Standard Design Approval in May 2025. As of 2026 no NuScale plant is operating; its flagship deployment (the Carbon Free Power Project / UAMPS) was cancelled in 2023, so the concept-to-grid clock for SMRs remains unfinished and is tracking toward ~25+ years.

Why it takes so long

For a brand-new reactor class the long pole is design maturation plus first-of-a-kind regulatory review: the NRC certification review alone took ~3.5 years (2017-2020/2022) and required millions of pages of analysis for a non-water-margin novel design. On top of design certification, a deployment still needs a site Combined License, supply chain, customer offtake/financing, and construction — none of which NuScale has yet completed. First-of-a-kind regulatory precedent, financing/offtake uncertainty (the cancelled UAMPS project), and an immature manufacturing supply chain extend the timeline well beyond the design phase.

  1. ~2000-2007DOE/OSU MASLWR research (concept) to company founding
  2. ~2007-2017Company founding to NRC design certification application
  3. ~2017-2022 (~5 years)NRC review to design certification (50 MWe)
  4. 2025 (review ~22 months)Uprated 77 MWe Standard Design Approval
  5. not yet achieved as of 2026First commercial unit operational
Best-supported ~2000 (MASLWR concept) to first design certification: July 2022 Commission vote (final rule effective Feb 2023) = ~22-23 years; concept-to-operating-grid remains unfinished and is tracking toward ~25-30 years given the 2023 CFPP cancellation and no unit operating as of 2026.

Naval vessels

USS Gerald R. Ford (CVN-78), lead ship of the Ford-class aircraft carrier program

Defenseconfirmed · high
24yrs
1993 → 2017 commissioned

The Ford-class is the first wholly new U.S. supercarrier class in ~40 years. Concept work traces to mid-1990s 'CVX' studies (kicked off ~1997, with roots back to 1994); the formal design effort began when the CVN(X) Milestone 1 contract was awarded to Northrop Grumman Newport News in 2000, renamed CVN-21 in 2002. Lead ship CVN-78 had a ceremonial steel-cut in Aug 2005, keel laid Nov 2009, christened Nov 2013, delivered May 2017, and was commissioned 22 July 2017. So 'the clock' runs ~17 years from formal program start (2000) to commissioning, or ~23 years if you count earliest concept studies.

Why it takes so long

A clean-sheet class bundled multiple unproven, simultaneously-developed technologies (EMALS electromagnetic catapults, Advanced Arrest Gear, A1B nuclear reactor, Dual-Band Radar, new weapons elevators), each requiring its own R&D and land-based testing before integration. Nuclear-powered, ~100,000-ton lead ships are also one-of-a-kind first articles with no production learning curve, extensive testing/trials, and Congressional authorization and budgeting cycles that stretch construction across many fiscal years.

  1. ~1994-2000 (formal CVX study ~1997)Concept/technology studies (CVX/CVN(X))
  2. 2000-2005Design / Milestone 1 contract (CVN(X), renamed CVN-21 in 2002)
  3. Aug 2005-2007Advance construction / first steel cut
  4. Nov 2009-Nov 2013Main construction: keel to christening
  5. Nov 2013-Jul 2017Outfitting, trials, delivery to commissioning
Best-supported ~17 years from formal program start (2000 Milestone 1 / Northrop Grumman Newport News contract) to commissioning (22 July 2017); ~21-24 years if measured from earliest concept work (1993-1996 CVX studies). The claim's stated ~17-23 years is well-supported.

USS Virginia (SSN-774), lead ship of the Virginia-class fast-attack submarine program

Defenseconfirmed · high
13yrs
1991 → 2004 commissioned

The Virginia class was conceived as an affordable post-Cold-War successor to the costly Seawolf class. The program was initiated around 1991 under the 'Centurion' / New Attack Submarine (NSSN) studies. The lead-ship construction contract went to General Dynamics Electric Boat on 30 Sept 1998, keel was laid 2 Sept 1999, the boat was christened 16 Aug 2003, and USS Virginia was commissioned 23 Oct 2004. 'The clock' is ~13 years from program concept (1991) to commissioning, or ~6 years from the construction contract (1998) to commissioning.

Why it takes so long

Even an explicitly cost-driven design needed years of concept definition and detailed design before steel was cut. Nuclear submarines require reactor integration, extensive land-based prototyping, pressure-hull fabrication, and lengthy sea trials/shakedown. Lead boats absorb first-of-class engineering risk and design changes; the program deliberately spread the early-1990s concept phase over several years to control cost.

  1. ~1991-1998Program initiation / Centurion-NSSN concept studies
  2. Sep 1998-Sep 1999Construction contract award to keel laying
  3. Sep 1999-Aug 2003Construction to christening
  4. Aug 2003-Oct 2004Delivery, trials, and commissioning
Best-supported ~6 years contract-to-commissioning (30 Sept 1998 to 23 Oct 2004); ~13-13.5 years from Feb 1991 Centurion concept study to Oct 2004 commissioning. Both figures are accurate and fairly characterized.

USS District of Columbia (SSBN-826), lead ship of the Columbia-class ballistic-missile submarine program

Defenseconfirmed · high
20yrs
2010 → ~2030 first patrol

The Columbia class replaces the aging Ohio-class SSBNs and is the U.S. Navy's top acquisition priority. Concept/technology development began around 2010-2011 (originally the 'Ohio Replacement Program', renamed Columbia in 2016). The detailed-design phase ran with ~3,000 Electric Boat engineers by late 2016; lead-ship construction began 1 Oct 2020. As of Feb 2026 the lead boat is ~65% complete, with the Navy targeting delivery in 2027-2028 and a first strategic deterrent patrol around 2030-2031. 'The clock' runs roughly 17-21 years from concept (2010-2011) to first operational patrol (~2030-2031).

Why it takes so long

An SSBN is the most demanding ship type to field: it integrates a new Common Missile Compartment (jointly developed with the UK Dreadnought program), a life-of-ship nuclear reactor that never needs refueling, electric-drive propulsion, and acoustic-stealth requirements, all under strict strategic-deterrence reliability standards. The program also rebuilds an atrophied submarine industrial base and supplier network, which has driven schedule slips. Construction of a single ~20,000-ton boat spans most of a decade, layered on top of years of design and Congressional funding milestones.

  1. ~2010-2016Concept/technology development (Ohio Replacement Program)
  2. ~2016-2020Detailed design phase
  3. Oct 2020Lead-ship construction start
  4. 2020-~2027/2028Construction to projected delivery
  5. ~2028-2030/2031Trials to first strategic deterrent patrol
Best-supported ~20 years: formal concept/Milestone A (Dec 2010 - Jan 2011) to first strategic deterrent patrol (FY2030, possibly 2031); lead-ship construction 2020-2028 (delivery), with ~65% complete as of Feb 2026. Range 19-21 years.

Spacecraft & probes

James Webb Space Telescope (JWST)

Spaceconfirmed · high
32yrs
1989 → 2021 launch

Discussions of a Hubble successor began around 1989, but the formal Next Generation Space Telescope (NGST) concept and design study started in 1996. It was renamed the James Webb Space Telescope in 2002 when contractor teams were selected, the prime contract went to TRW (later Northrop Grumman) in 2003, construction proceeded through 2016, and it finally launched on an Ariane 5 on December 25, 2021. First science images were released July 11, 2022. The 'clock' here is best measured from the 1996 formal concept/design study to the 2021 launch (25 years), with operations beginning in 2022.

Why it takes so long

JWST stretched out because it pushed multiple technologies to the bleeding edge simultaneously (a 6.5 m segmented gold-coated beryllium mirror that had to unfold in space, a tennis-court-sized five-layer sunshield, cryogenic detectors operating near 40 K) with essentially zero tolerance for failure at a Sun-Earth L2 orbit where no servicing is possible. Repeated cost overruns and a major 2005 redesign triggered budget rebaselining and a near-cancellation by Congress in 2011, forcing schedule slips. Each novel deployable mechanism required extensive ground testing, and integration/test of a non-serviceable observatory dominated the back half of the program.

  1. 1996-2002 (6 yrs)Concept / pre-formulation (NGST design study, naming)
  2. 2002-2007 (5 yrs)Design & contractor selection (TRW/Northrop Grumman, 2005 redesign)
  3. 2007-2021 (14 yrs)Construction, integration & test
  4. Dec 2021 launch, July 2022 first imagesLaunch to first science / operations
Best-supported 25 years (1996 formal NGST concept/design study to 2021 launch); up to ~32 years if measured from the 1989 "next Hubble" workshop. Operations/first science began 2022.

Cassini-Huygens (Saturn orbiter / Titan probe)

Spaceadjusted · high
29yrs
1988 → 2017 end of mission

ESA selected Cassini-Huygens as a major mission in late 1988 and the US program received major funding in 1989. It passed critical design review in December 1992, launched on a Titan IVB/Centaur on October 15, 1997, took a nearly 7-year gravity-assist cruise (Venus, Earth, Jupiter flybys), entered Saturn orbit on July 1, 2004, and operated until the deliberate atmospheric 'Grand Finale' plunge on September 15, 2017. The 'clock' spans 1988/89 program start to 1997 launch (~9 years), to 2004 operational arrival (~16 years), to 2017 end of mission (~29 years).

Why it takes so long

Cassini was a large international flagship (NASA, ESA, ASI) carrying a sophisticated orbiter plus the Huygens Titan-descent probe, requiring complex multi-agency integration. Beyond the development period, the physics of reaching Saturn imposed a built-in multi-year delay: the spacecraft was too heavy to fly direct, so it used a Venus-Venus-Earth-Jupiter gravity-assist trajectory that added nearly 7 years of cruise before any science at Saturn. Planetary-mission timelines are therefore dominated not just by build time but by orbital mechanics and long interplanetary transit.

  1. 1988-1992 (4 yrs)Selection & concept (ESA selection, US funding)
  2. 1992-1997 (5 yrs)Design & build (CDR 1992 to launch)
  3. 1997-2004 (~7 yrs)Interplanetary cruise (launch to Saturn orbit insertion)
  4. 2004-2017 (13 yrs)Operations at Saturn to Grand Finale
Best-supported ~28-29 years end-to-end (late-1988 ESA selection / 1989 US funding to 15 Sep 2017 end of mission); ~16 years program-start to 1 Jul 2004 Saturn arrival; ~9 years program-start to 15 Oct 1997 launch; ~7 years (not 9) from launch to operational Saturn arrival

Europa Clipper (Jupiter/Europa flyby orbiter)

Spaceconfirmed · high
33yrs
1997 → 2030 arrival

A Europa Orbiter mission was first proposed in 1997, followed by the cancelled Jupiter Icy Moons Orbiter (JIMO) concept in the early 2000s. Funding for the multiple-flyby 'Clipper' concept ramped up from 2013-2014, NASA formally approved the mission concept in June 2015, it moved through preliminary design (Phase B, 2017-2019), final design/fabrication (Phase C, 2019), and assembly/test (Phase D, 2022). It launched on a SpaceX Falcon Heavy on October 14, 2024 and is due to arrive at Jupiter in April 2030 for ~50 Europa flybys. The 'clock' is best read from the formal 2015 concept approval to the 2024 launch (~9 years), or from the original 1997 concept (~27 years to launch).

Why it takes so long

Europa Clipper's long arc reflects how long mission concepts can sit in study, repackaging, and funding limbo before formal start: the science goal (a Europa orbiter/flyby) existed from 1997 but cycled through cancelled designs (JIMO) for over a decade before the radiation-tolerant multi-flyby architecture was adopted and funded. Once formally approved in 2015, development still ran ~9 years because of Jupiter's extreme radiation environment (requiring a heavily shielded vault and radiation-hardened electronics), large solar arrays, and a complex instrument suite. As with Cassini, orbital mechanics then add roughly 5.5 years of transit before operations begin in 2030.

  1. 1997-2012 (~15 yrs in study/limbo)Early concept studies (Europa Orbiter, JIMO)
  2. 2013-2019 (~6 yrs)Formulation & formal approval (Phase A/B)
  3. 2019-2024 (5 yrs)Final design, fabrication, assembly & test (Phase C/D)
  4. Oct 2024 launch, ~2030 arrival, ~4 yrs scienceCruise to Jupiter / operations
Best-supported 27 years from the original 1997 Europa Orbiter proposal to the October 14, 2024 launch (33 years to April 2030 Jupiter arrival); however the Clipper-specific timeline is best read as ~9 years from June 2015 mission-concept approval to launch (or ~11 years from 2013 funding start)

Rockets & launch vehicles

Ariane 6 (ESA / ArianeGroup)

Spaceadjusted · high
10yrs
2014 → 2024 first flight

The European Space Agency selected and funded the Ariane 6 design (Ariane 62 and Ariane 64 variants) in December 2014, with an original debut target of 2020. After repeated slips driven by technical issues and the COVID-19 pandemic, the maiden flight (VA262) launched on 9 July 2024, and the first operational/commercial mission (VA263, carrying the CSO-3 reconnaissance satellite) flew on 6 March 2025.

Why it takes so long

Ariane 6 is a clean-sheet expendable vehicle developed by a multinational consortium under ESA governance, requiring intergovernmental funding agreements and work-share negotiations across member states. New engines (Vulcain 2.1, Vinci with re-ignition), new solid boosters (P120C shared with Vega-C), a new upper stage, and a new launch pad in French Guiana all had to be qualified together. Coordinating distributed manufacturing, certifying human-rated-class reliability for commercial payloads, and absorbing COVID-era disruptions stretched the schedule roughly four years beyond the original 2020 plan.

  1. Dec 2014Program selection and funding approval (concept to authorized development)
  2. 2015-2024 (~9.5 years; original target was 2019-2020 completion)Development, hardware qualification, and ground/pad testing
  3. 9 July 2024 (~10 years after program start)Maiden/inaugural flight (VA262)
  4. 6 March 2025 (~11 years after program start)First operational / commercial flight (VA263, CSO-3)
Best-supported ~9.6 years from program start (2 Dec 2014) to maiden flight (9 Jul 2024); ~10.25 years to first operational/commercial flight (6 Mar 2025). Best phrased as ~10 years to first flight and ~10 years to operational, not ~11.

NASA Space Launch System (SLS)

Spaceconfirmed · high
17yrs
2005 heritage → 2022 first flight

SLS was mandated by the NASA Authorization Act of 2010 and the design was formally announced on 14 September 2011, as a replacement for the retired Space Shuttle and the cancelled Constellation program's Ares I/Ares V rockets. Its first flight was the uncrewed Artemis I mission, which launched on 16 November 2022, roughly 11-12 years after the program clock started. The project drew heavy criticism for cost overruns and delays, costing about USD 31.6 billion as of 2025.

Why it takes so long

SLS is a super-heavy-lift, human-rated vehicle built largely from heritage Shuttle components (RS-25 engines, five-segment solid rocket boosters, the core stage derived from the External Tank), but adapting that hardware into a new architecture proved harder than reuse implied. As a congressionally directed program subject to annual appropriations, a distributed contractor base, and human-rating safety requirements, it accumulated repeated schedule slips. The original notional 2017 first flight slipped to late 2022 due to core-stage manufacturing/welding challenges, the Green Run static-fire campaign, and integration/launch-pad issues.

  1. Oct 2010Congressional authorization (NASA Authorization Act of 2010)
  2. 14 Sep 2011Design announced / formal program start
  3. 2011-2022 (~11 years; original estimate targeted ~2017)Development, core-stage build, Green Run testing, integration
  4. 16 Nov 2022 (~11 years after 2011 design announcement; ~12 years after 2010 authorization)First flight (Artemis I, uncrewed lunar test)
Best-supported ~11-12 years (Oct 2010 NASA Authorization Act / Sept 2011 design announcement to Nov 2022 Artemis I first flight); up to ~17 years if Constellation/Ares heritage from 2005 is counted

SpaceX Falcon Heavy

Spaceconfirmed · high
7yrs
2011 → 2018 first flight

SpaceX publicly unveiled the Falcon Heavy at an April 2011 Washington, D.C. news conference, projecting an initial test flight in 2013. The maiden flight slipped about five years and finally lifted off on 6 February 2018, carrying Elon Musk's Tesla Roadster as a dummy payload. The delay illustrates that even a derivative vehicle built from existing Falcon 9 cores took roughly seven years from announcement to first successful flight.

Why it takes so long

Falcon Heavy was conceived as three strapped-together Falcon 9 first stages, which the company expected would be straightforward, but combining three cores introduced unanticipated structural loads, separation dynamics, and integration challenges that required substantial redesign. The schedule was further disrupted because two Falcon 9 anomalies (the June 2015 CRS-7 in-flight failure and the September 2016 Amos-6 pad explosion) diverted SpaceX engineering resources to failure investigations and halted operations for months. The constantly evolving Falcon 9 baseline also meant the Heavy design had to keep re-tracking the parent rocket.

  1. Apr 2011Public announcement (concept unveiled, 2013 first-flight target)
  2. 2011-2018 (~7 years; original target was 2013)Development, core integration redesign, delays from Falcon 9 anomaly investigations
  3. 6 Feb 2018 (~7 years after announcement)First flight (demonstration mission, Tesla Roadster payload)
Best-supported April 2011 to February 2018 (~6.8 years, i.e. ~7 years announcement-to-first-flight)

Automotive

Volkswagen MEB platform & VW ID.3 (first MEB vehicle)

Transportconfirmed · high
5yrs
2015 → 2020 deliveries

MEB (Modularer E-Antriebs-Baukasten) was Volkswagen's first dedicated, clean-sheet battery-electric architecture. Development began around 2015. The architecture was first previewed as the I.D. concept car at the 2016 Paris Motor Show, the production ID.3 was unveiled in September 2019, start of production (SOP) ran at the Zwickau plant at the end of 2019, and retail customer deliveries began in September 2020. So roughly four years from development kickoff to SOP, with a fifth year to volume customer deliveries.

Why it takes so long

MEB was a clean-sheet electric architecture, not a derivative, so VW had to design the battery skateboard, e-powertrain, high-voltage electrical architecture and software stack from scratch, then revalidate the whole vehicle. Software in particular caused well-documented delays (reported 'massive software problems'). On top of engineering, a clean-sheet platform requires new tooling, a converted/dedicated plant (Zwickau), supplier contracting for new components (cells, modules, inverters), and full safety/crash and homologation validation before any unit can be sold.

  1. 2015Platform development start (MEB concept/architecture work begins)
  2. 2016Concept reveal (I.D. concept car, Paris Motor Show)
  3. 2016-2019 (~3 years)Engineering, prototype, validation and software integration
  4. Sep-Nov 2019Production reveal (ID.3, Frankfurt) and start of production (Zwickau)
  5. by Sep 2020Ramp-up to retail customer deliveries
Best-supported ~4 years development start (Oct 2015) to SOP (Nov 2019); ~5 years to first customer deliveries (Sept 2020)

Toyota New Global Architecture (TNGA) platform + TNGA powertrains

Transportadjusted · high
7yrs
2008 → 2015 first product

TNGA is Toyota's clean-sheet modular architecture program that replaced roughly 100 legacy platform variants and was paired with an all-new family of TNGA powertrains (engines, transmissions, hybrid systems). The strategy was conceived in the aftermath of the 2008 financial crisis and the first TNGA product, the fourth-generation Prius, reached market at the end of 2015 — a multi-year clean-sheet effort that simultaneously redeveloped the vehicle architecture and the powertrain line.

Why it takes so long

TNGA was not a single car but a ground-up redesign of the company's entire platform-and-powertrain strategy: new standardized seat heights, shared modules (steering, pedals, seat frames, airbags), plus a completely new engine/transmission/hybrid family that each carries its own multi-year design, dyno testing, durability and emissions/safety validation cycle. Powertrain programs in particular are built on ~10-year business cases spanning multiple vehicle programs, so the architecture had to be validated to serve dozens of future models (now 40+) rather than one.

  1. ~2009 onwardStrategy/concept (post-Lehman rethink of platform strategy)
  2. several yearsArchitecture + new powertrain family development and validation
  3. late 2015First TNGA product (4th-gen Prius) start of production / launch
  4. 2015 onwardRoll-out across 40+ models (Corolla, Camry, RAV4, etc.)
Best-supported Platform program: ~7 years (2008 post-crisis inspiration → late-2015 4th-gen Prius), or ~3 years from public strategy articulation (2013) / ~7 months from formal presentation (Mar 2015) to launch. Powertrain (Dynamic Force) line: a separate, partly-overlapping effort that did not reach market until 2017 (announced Dec 2016), rolling out through ~2021 — NOT delivered with the 2015 first product.

Industry-typical new-vehicle program (legacy OEM vs Chinese EV benchmark)

Transportconfirmed · high
4yrs
~40–53 months legacy / ~24 months China EV

Across the industry, the 'clock' for a new program is most often measured from program approval/concept to start of production (SOP). McKinsey and the Center for Automotive Research put a typical legacy-OEM clean-sheet program at roughly 40-50 months (sometimes longer), while best-in-class programs run ~24-36 months. Leading Chinese EV makers have compressed concept-to-launch to roughly 24 months by using carryover modular architectures, over-the-air software updates, and parallelized validation.

Why it takes so long

A clean-sheet program must serialize and validate many interdependent systems: body/chassis architecture, powertrain (engine/transmission or e-drive and battery), electrical/software stack, plus full safety validation via physical crash testing to NHTSA/Euro NCAP standards and emissions/ECU calibration. Beyond engineering, automakers need long lead times for tooling, plant preparation, supplier contracting, and government homologation — which is why SOP, not the auto-show concept, is the binding milestone. Software-heavy EVs and the option to fix issues via OTA after launch are the main levers shrinking the pre-SOP window.

  1. early phase of the V-modelConcept / program approval (target setting, business case)
  2. bulk of the timelineDesign and engineering development (down the left side of the V)
  3. back half of the programPrototype build, integration, and validation (right side of the V: powertrain + crash/safety + emissions)
  4. final monthsPre-production / tooling / ramp to Start of Production (SOP)
  5. ~24 months (best/China) to 40-50 months (typical legacy)Total concept-to-SOP
Best-supported Legacy/incumbent OEM clean-sheet program: ~40-53 months concept/design-to-SOP (McKinsey: mass-market 45, sports 48, premium 53; "best-in-class" ~24-36 mo achievable). Leading Chinese / new EV makers: ~18-24 months (McKinsey new-EV-OEM 24 mo = 10+5+9; BYD cited ~18 mo).

High-speed rail

California High-Speed Rail (Phase 1, San Francisco–Los Angeles/Anaheim)

Transportconfirmed · high
25yrs
2008 → 2033+ (counting)

California's HSR is the most prominent U.S. example. The clock here is best measured from the 2008 voter-approved bond (Proposition 1A, the program launch / funding authorization) to revenue service. Construction in the Central Valley began in January 2015. As of 2026 no segment is yet operational; the Merced–Bakersfield Initial Operating Segment is the first to carry passengers, with station construction running mid-2026 to mid-2029 and service expected in the early 2030s. The original 2020 full-Phase-1 target has been abandoned, illustrating the gap between program launch and operational reality.

Why it takes so long

HSR megaprojects combine multi-decade structural drivers: securing political consensus and funding (the 2008 bond covered only a fraction of cost), environmental review and litigation under CEQA/NEPA, right-of-way acquisition across hundreds of parcels, utility and rail-corridor relocations, and phased construction of dedicated grade-separated track. Cost escalation forces re-scoping and re-sequencing (the project pivoted to a Central Valley 'starter' line), and funding arrives incrementally rather than upfront, stretching schedules. None of these can be compressed below roughly a decade even before construction.

  1. pre-2008 to Nov 2008Planning, studies, and bond authorization (Prop 1A program launch)
  2. 2008–2014 (~6 years)Environmental review, design, funding assembly, pre-construction
  3. Jan 2015 onwardConstruction of Central Valley civil works begins
  4. station build mid-2026 to mid-2029; service early 2030sInitial Operating Segment (Merced–Bakersfield) stations and systems to revenue service
Best-supported 2008 (Prop 1A authorization) to early 2030s for first passenger service on the Merced–Bakersfield IOS (~2031–2033 per current Authority/Inspector General projections), i.e. ~23–25 years just to the first operating segment; full Phase 1 SF–LA/Anaheim remains unfunded with no committed date, so total elapsed time will exceed 25 years and is "still counting." The claimed "2008–2033+ / ~25+ years" range is accurate and, if anything, conservative.

HS2 (High Speed 2), Phase 1 London–Birmingham, United Kingdom

Transportconfirmed · high
27yrs
2009 → 2036+ (counting)

HS2's clock runs from the January 2009 creation of HS2 Ltd (program launch) to revenue service. The government confirmed the project would proceed in January 2012; Phase 1 received Royal Assent (the hybrid bill / legal authorization) in February 2017; construction began in 2020. As of 2026, Phase 1 passenger services (Old Oak Common–Birmingham Curzon Street) are not expected before 2036, with London Euston later still — a dramatic slip from the December 2026 date envisioned when proposed in 2010.

Why it takes so long

HS2 shows the same structural drivers as California plus a parliamentary 'hybrid bill' process that itself took years to grant construction powers. Compulsory land purchase, extensive tunneling and environmental mitigation, contractor mobilization, and repeated political re-scoping (the northern legs to Manchester were cancelled, and Euston was paused) all extended the schedule. Cost overruns triggered formal resets that pushed completion estimates back multiple times, so the interval from program launch to first revenue service exceeds a quarter century.

  1. Jan 2009HS2 Ltd founded (program launch)
  2. 2012–2016Government go-ahead and route consultation/design
  3. Feb 2017Phase 1 hybrid bill Royal Assent (construction powers)
  4. 2020 onwardMain civil construction
  5. now estimated 2036–2039Phase 1 revenue service (Old Oak Common–Birmingham)
Best-supported 2009-2036+ (HS2 Ltd created Jan 2009; Phase 1 Old Oak Common-Birmingham revenue service now officially scheduled May 2036-Oct 2039, London Euston/full scheme 2040-2043). Original 2010 concept targeted a December 2026 Phase 1 opening, so the slip is ~10-13+ years versus that target and the elapsed program runs ~27+ years from 2009.

Chūō Shinkansen maglev (Tokyo–Nagoya), JR Central, Japan

Transportconfirmed · high
24yrs
2011 → 2035+ (slipped)

The Chuo Shinkansen is the leading maglev megaproject. Its deepest roots go back to the 1973 Basic Plan designation, but the actionable clock starts with government construction approval on 27 May 2011, followed by groundbreaking on 17 December 2014. Commercial Tokyo–Nagoya service was originally targeted for 2027; it has been repeatedly delayed and is now expected no earlier than 2035, primarily because Shizuoka Prefecture withheld permission for tunneling work over environmental (water) concerns. The Nagoya–Osaka extension is targeted as early as 2037.

Why it takes so long

Beyond the standard HSR drivers, the Chuo Shinkansen is roughly 86% tunnel through mountainous terrain, including the deep Southern Alps tunnel — excavation alone was expected to take at least a decade. A single prefecture's refusal to authorize tunneling (over groundwater impact on the Oi River) stalled a critical section for years, showing how local environmental permitting can dominate the timeline. Building first-of-kind superconducting maglev infrastructure at commercial scale adds engineering and approval complexity absent from conventional steel-wheel HSR.

  1. 1973 onward (test track running since the 1990s)Basic Plan designation / decades of maglev R&D
  2. 27 May 2011Government construction approval (program go-ahead)
  3. 17 Dec 2014Construction start (groundbreaking)
  4. 2014–2030s (delayed by Shizuoka permitting from 2020)Tunneling and infrastructure build
  5. now no earlier than 2035Tokyo–Nagoya commercial service (original 2027 target)
Best-supported 2011–2035+ for Tokyo–Nagoya (authorization 27 May 2011; construction start 17 Dec 2014; original target 2027; now no earlier than 2035). Nagoya–Osaka extension targeted as early as 2037. Deepest planning roots: 1973 Basic Plan.

N700S Series Shinkansen (new rolling stock design), JR Central, Japan

Transportconfirmed · high
4yrs
2016 → 2020 service

The N700S ('S' for Supreme) is a clean example of how fast a new high-speed train design — as opposed to a whole rail line — can move when built on an established platform and proven manufacturing base. JR Central announced the new trainset concept in June 2016, unveiled the J0 prototype on 10 March 2018 (testing from 20 March 2018), ran 360 km/h high-speed trials in 2019, and entered revenue service on the Tokaido Shinkansen on 1 July 2020. The clock here measures concept announcement to first passenger service.

Why it takes so long

New rolling stock is far quicker than new infrastructure because there is no land acquisition, environmental review, or civil construction. The dominant time drivers are engineering design, prototype fabrication, and a rigorous testing/validation program (high-speed running trials, safety and earthquake-resilience verification) before type approval and series production. The N700S was the first Shinkansen designed solely by JR Central since the 300 Series, yet it still reused the mature N700 platform and Nippon Sharyo production line, keeping the cycle to about four years.

  1. June 2016Concept announcement / start of design
  2. to March 2018Prototype (J0) build and unveiling
  3. 2018–2019Testing and high-speed trials (up to 363 km/h)
  4. first set delivered April 2020; service 1 July 2020Series production and first revenue service
Best-supported June 2016 – July 2020 (~4 years from concept announcement to revenue service)

Power turbines

Siemens SGT5-8000H heavy-duty gas turbine (Irsching 4)

Energyconfirmed · high
11yrs
2000 → 2011 commercial

Siemens began the internal product-development/strategic-planning process for its H-class SGT5-8000H in 2000, publicly announced the machine in 2005, and completed the prototype in Berlin in April 2007. The prototype was shipped to the Irsching 4 site, first-fired in December 2007, synchronized to the grid in March 2008, and ran an ~16-18 month in-service validation (simple-cycle) campaign through 2009. After conversion to combined cycle, Irsching 4 set a 60.75% net efficiency world record in a May 2011 test run and was handed over to E.ON for commercial operation in July 2011.

Why it takes so long

The clock is dominated by an extended single-engine validation loop on a first-of-a-kind machine: firing temperatures and component stresses push materials and cooling to limits that only physical testing can confirm, so the prototype must accumulate real grid-synchronized operating hours, instrumented measurement campaigns, and an endurance test before commercial release. A purpose-built test plant (Irsching 4) had to be constructed first, and the simple-cycle validation engine had to be physically reconfigured into combined-cycle afterward. Design, manufacture (large single-piece forgings/castings), site construction, multi-season testing, and uprate iterations each add years.

  1. 2000-2005Concept / strategic product planning / design start
  2. 2005-April 2007Detailed design and prototype manufacture (Berlin)
  3. 2007-2008Transport, installation, first fire (Dec 2007), grid sync (Mar 2008)
  4. 2008-2009 (~16-18 months)Simple-cycle in-service validation and endurance testing at Irsching 4
  5. 2009-2011Conversion to combined cycle; 60.75% efficiency record (May 2011); commercial handover to E.ON (July 2011)
Best-supported 2000 (development program start) to July 2011 (commercial handover to E.ON) — approximately 11 years concept-to-commercial. Key intermediate dates: public announcement 2005; prototype completed April 2007 (Berlin); first fire December 20, 2007; grid synchronization March 7, 2008; ~16-month simple-cycle validation concluding August 2009; 60.75% net combined-cycle efficiency record May 2011. Design-start to prototype first-fire is ~7-8 years (2000 to Dec 2007).

GE 9HA.01 heavy-duty gas turbine (EDF Bouchain)

Energyconfirmed · high
7yrs
2009/11 → 2016 in operation

GE's HA/9HA program was backed by a roughly $1 billion R&D investment under its 'FlexEfficiency' high-firing-temperature (>2600F) portfolio. Full-load validation trials of the 9HA.01 test unit were completed in January 2015 at GE's variable-speed, variable-load test stand in Greenville, South Carolina. The first production unit was completed at Belfort, France in May 2015, shipped to EDF's Bouchain combined-cycle plant in June 2015, achieved 'first fire' on December 17, 2015, and reached commercial operation date (COD) in July 2016 - the plant later earning a Guinness World Record for 62.22% combined-cycle efficiency.

Why it takes so long

GE shortened the loop relative to a pure prototype-in-the-field approach by building a dedicated full-speed, full-load (FSFL) test facility in Greenville that can run the engine at variable load and validate aero/thermal/mechanical performance before a unit is ever shipped, but the underlying multi-year R&D, materials qualification, and component testing still consume most of the clock. After factory validation the unit still needs manufacturing of very large single castings/forgings, transport, site commissioning, first fire, and a tuning/commissioning ramp to reach COD - and the program then iterates into uprated variants (9HA.02).

  1. ~2009-2014R&D / FlexEfficiency HA development (~$1B investment)
  2. through January 2015Full-speed full-load validation of 9HA.01 test unit (Greenville, SC) completed
  3. May 2015First production unit manufactured at Belfort, France
  4. 2015Shipment to Bouchain (June 2015) and first fire (Dec 17, 2015)
  5. Dec 2015 - July 2016Commissioning to commercial operation date (COD)
Best-supported FlexEfficiency program announced 2011; HA turbine line introduced 2014; 9HA.01 validation completed Jan 2015 (final thermal mapping Mar 2015); first production unit completed at Belfort May 28, 2015; first fire Dec 17, 2015; grid sync Jan 21, 2016; combined-cycle Mar 1, 2016; COD July 26, 2016. Best-supported R&D-to-COD span: ~2011-2016 (~5 years from FlexEfficiency program; up to ~7 years if counting earlier high-firing-temperature R&D lineage).

AD700 / Advanced Ultra-Supercritical (700C) steam turbine and plant program (Europe; parallel US DOE/OCDO A-USC)

Energyadjusted · high
20yrs
1998 → 2018 planned program

The European AD700 Advanced (700C) Pulverised-Fuel Power Plant program targeted a coal-fired steam plant with 700C turbine inlet steam (vs. ~600C for conventional ultra-supercritical), requiring entirely new nickel-based superalloys for the steam turbine, boiler, and high-temperature piping. It was structured as four phases over roughly 20 years (1998-2018): Phase 1 Part A finished December 2001, Part B in 2004, Phase 2 (2002-2007), with the central AD700 effort running 1998-2011. The parallel US DOE/Ohio Coal Development Office Advanced Materials for Ultrasupercritical Coal-Fired Boilers project was an ~$50 million multi-year R&D program building the same materials base.

Why it takes so long

Steam-turbine timelines are dominated by materials qualification rather than aerodynamics: pushing inlet steam to 700-760C exceeds the creep/oxidation limits of ferritic and austenitic steels, forcing development and long-duration creep-rupture testing (often 50,000-100,000+ hours) of new nickel-based superalloys for rotors, casings, valves, and thick-section piping. Large forgings in these alloys must be manufacturable at full size, weldable, and code-qualified, and component/loop test rigs must run for years before a demonstration plant can be committed. This validation loop, plus the cost and risk of first-of-a-kind demonstration, is why A-USC steam plants have taken two decades and several have stalled at the demonstration stage.

  1. 1998-Dec 2001Phase 1 Part A (feasibility, initial materials screening)
  2. 2001-2004Phase 1 Part B
  3. 2002-2007Phase 2 (component development / materials qualification)
  4. 2007-2011 (program horizon to 2018)Core AD700 R&D through completion; demonstration phases (planned to 2018)
Best-supported Program launched January 1998 (contract 1997). Planned 4-phase, ~20-year program with a 1998-2018 horizon (per IEA CCC/229 Gantt). Phase 1 1998-2004 (Part A conceptual feasibility ~1998-2001, Part B materials properties to 2004); Phase 2 ~2002-2006 (initiated 2001) [claim's 2007 is ~1 year long]; Phase 3 (COMTES 700 component test) ~2004-2009/2011; Phase 4 full-scale demo planned but not built on schedule (demo target 2014). Core AD700 R&D effectively ran 1998 to ~2011.

Medical devices

AbioCor Total Artificial Heart (AbioMed)

Medicalconfirmed · high
16yrs
1990s → 2006 (lineage 25–40 yrs)

AbioMed began developing the AbioCor — the first fully self-contained, implantable total artificial heart with no external tubes — in the early-to-mid 1990s, ran animal studies starting 1998, received an FDA Investigational Device Exemption (IDE) and performed its first human implant in July 2001, and ran a 14–15 patient clinical trial through 2004. After an initial 2005 FDA advisory-panel rejection, the device received FDA Humanitarian Device Exemption (HDE) approval on September 5, 2006. Note: the device's underlying research lineage stretches back roughly 30 years through earlier artificial-heart work.

Why it takes so long

Class III life-sustaining implants require extensive bench and animal testing before any human use, then an IDE before a clinical trial can even begin. The implantable population is tiny and very sick, so enrollment is slow; mortality and adverse events trigger panel scrutiny (the HDE was initially rejected). Each iteration of an implantable pump, battery, and transcutaneous energy transfer system must be validated for biocompatibility, durability, and infection control before regulators will approve even humanitarian-use marketing.

  1. early–mid 1990sConcept / company development begins
  2. 1998–2000Animal studies
  3. Jan–Jul 2001IDE granted + first human implant
  4. 2001–2004Clinical trial (14–15 patients)
  5. 2005FDA panel review (initial rejection 2005)
  6. Sep 2006FDA HDE approval (to market)
Best-supported First-in-human (July 2001) to FDA HDE approval (Sept 5, 2006): ~5 years. AbioCor-specific development (early-to-mid 1990s) to market (2006): ~11–16 years. Broader program lineage to market: ~25 years from Abiomed's 1981 founding, or ~30+ years if counting the NHLBI Artificial Heart Program begun in 1964 (the figure cited as "research lineage").

SynCardia / CardioWest Total Artificial Heart (Jarvik-7 lineage)

Medicalconfirmed · high
36yrs
late-1960s → 2004 approval

The device descends from the Jarvik-7, designed by Willem Kolff and Robert Jarvik at the University of Utah in the late 1960s and first implanted in Barney Clark in December 1982. After the FDA withdrew the IDE and shut down the manufacturer (Symbion) in 1990, the technology was transferred to CardioWest in 1991 and re-developed. A multi-year pivotal study of 81 implants underpinned FDA Premarket Approval (PMA) as a bridge-to-transplant device on October 15, 2004 — described by SynCardia as following roughly a 10-year pivotal clinical study.

Why it takes so long

Beyond the standard Class III bench/animal/IDE/clinical-trial sequence, this program absorbed a complete regulatory reset: the original IDE was withdrawn and the manufacturer closed in 1990, forcing a technology transfer and effectively restarting the path to market. The pivotal bridge-to-transplant trial spanned roughly a decade because eligible patients are rare and critically ill, and durability/infection data must accrue over long implant durations before a PMA can be granted.

  1. late 1960s–1970sOriginal design (Jarvik-7, Univ. of Utah)
  2. Dec 1982First permanent human implant (Barney Clark)
  3. 1990Regulatory reset: IDE withdrawn, Symbion closed
  4. 1991Technology transfer / rebrand to CardioWest
  5. ~1993–2002Pivotal clinical study (~81 implants, ~10 yrs)
  6. Oct 2004FDA PMA approval (bridge to transplant)
Best-supported Lineage from late-1960s Kolff/Jarvik Utah program → Jarvik-7 first human implant Dec 2, 1982 → technology to CardioWest ~1991 → pivotal 81-patient study spanning roughly 1993–2002 → FDA PMA approval October 2004 (commonly cited Oct 15; some sources Oct 18–19). Modern CardioWest program to approval: ~10–13 years.

Medtronic Deep Brain Stimulation (Activa system)

Medicalconfirmed · high
15yrs
1987 → 2002 Parkinson's approval

Modern DBS traces to Alim-Louis Benabid's 1987 finding that high-frequency thalamic stimulation suppresses tremor. Medtronic developed an implantable pulse generator and lead system (the Activa platform) through the 1990s and earned FDA Premarket Approval (PMA) for thalamic (VIM) DBS to treat essential tremor and Parkinsonian tremor in 1997. A larger multi-site pivotal trial supported FDA PMA approval of STN/GPi DBS for the motor symptoms of advanced Parkinson's disease in 2002. Subsequent indications followed (dystonia HDE 2003, OCD HDE 2009, epilepsy 2018), each requiring its own trial and review.

Why it takes so long

DBS is a Class III implant requiring PMA, the most rigorous FDA pathway. The path from a basic-science observation (1987) to a marketed system demanded designing implantable, biocompatible leads and a long-life pulse generator, then proving safety and efficacy in controlled multi-center trials in a delicate brain-surgery population. Each new clinical indication restarts much of the trial-and-review cycle rather than reusing the original approval, which is why the Parkinson's indication trailed the tremor approval by five years.

  1. 1987Foundational research (Benabid thalamic stimulation)
  2. late 1980s–1990sDevice development + clinical trials (Activa platform)
  3. 1997FDA PMA: essential tremor / Parkinsonian tremor
  4. late 1990s–2002Multi-site pivotal trial for Parkinson's motor symptoms
  5. 2002FDA PMA: STN/GPi DBS for advanced Parkinson's
Best-supported 1987 foundational research (Benabid, thalamic HFS) → July 1997 FDA PMA P960009 for thalamic/VIM DBS (essential tremor and Parkinsonian tremor) → 2002 FDA PMA for STN/GPi DBS in advanced Parkinson's disease. Approximately 10 years from foundational research to first FDA approval, and approximately 15 years to the major Parkinson's (STN/GPi) indication. Later indications: dystonia HDE 2003, OCD HDE 2009, epilepsy PMA 2018.

Mining

Industry-wide benchmark (S&P Global Market Intelligence study of 127–268 mines; IEA Critical Minerals report)

Miningadjusted · high
18yrs
~16 yr avg, rising toward ~18

The widely repeated '15–20 years from discovery to production' figure is well supported. S&P Global Market Intelligence found an average of 15.7 years across 127 mines, and a follow-on analysis (gold, copper, nickel, lithium mines online since 2000) found the average lead time has risen from 12.7 years for mines started in 2005–09 to 17.9 years for mines started in 2020–23. The IEA's 2021 Role of Critical Minerals report cites ~16 years on average. Copper specifically averages 17.9 years (S&P, Dec 2024). The figure is rising over time, and the US (~29 years) and Zambia (~34 years) are far above average.

Why it takes so long

The 'clock' measured here is discovery of the deposit to first commercial production — NOT just construction. Construction is only the final ~2–3 years. The bulk of the time is everything before ground is broken: delineating and proving up the resource through years of drilling, completing pre-feasibility and bankable feasibility studies, securing environmental and operating permits (often the longest single bottleneck, especially in the US and developed jurisdictions), arranging multi-billion-dollar project financing, and negotiating fiscal/ownership terms with host governments and local/Indigenous communities. S&P's phase breakdown: ~12.5 years discovery→feasibility, ~1.8 years construction planning, ~2.6 years construction→production.

  1. ~12.5 yearsDiscovery and exploration through completed feasibility study (resource delineation, PEA/PFS/bankable FS)
  2. ~1.8 yearsPost-feasibility: permitting, financing, construction planning (offtake, government terms, capital raise)
  3. ~2.6 yearsConstruction to first/commercial production
Best-supported ~16 years on average (S&P 15.7 across 127 mines; IEA 16.5), rising from ~12.7 years for mines started 2005-09 to ~17.9 years for mines started 2020-23. The commonly cited "15-20 years discovery-to-production" range is accurate. Copper specifically is ~14 years (faster than average), NOT 17.9 years.

Kamoa-Kakula Copper Complex (DRC) — Ivanhoe Mines / Zijin Mining

Miningconfirmed · high
13yrs
2008 → 2021 production

One of the largest and highest-grade copper discoveries of the modern era. The Kamoa deposit was discovered by Ivanhoe geologists in 2008 (the adjacent Kakula deposit was discovered in 2016); first copper concentrate was produced in May 2021 with commercial production declared July 1, 2021. From the original Kamoa discovery that gave the project its life, the clock ran about 13 years — fast for a project of this scale, but still consistent with the lower end of the 15–20 year range despite a very favorable, high-grade orebody.

Why it takes so long

The clock here measures original deposit discovery (2008) to first commercial production (2021). Even with an exceptional tier-one orebody, the years went to additional drilling and resource expansion, multiple feasibility studies, negotiating DRC government ownership (20%) and fiscal terms, project financing of a multi-billion-dollar build, and ~2–3 years of physical construction of the concentrator and underground mine. Kakula (2016) was the body that was actually mined first, illustrating that the productive deposit is often found years after the project's initial discovery.

  1. 2008 onwardKamoa discovery and early resource definition
  2. 2008–2018Resource expansion, Kakula discovery (2016), feasibility studies, government/financing
  3. ~2019–May 2021Construction to first copper concentrate
Best-supported ~13 years from original Kamoa discovery (2008) to commercial production (July 1, 2021); ~5 years from the Kakula discovery (2016) that defined the high-grade starter deposit

Oyu Tolgoi (Mongolia) — Ivanhoe Mines / Turquoise Hill / Rio Tinto

Miningconfirmed · high
22yrs
2001 → 2023 underground

A giant copper-gold porphyry system. Ivanhoe Mines estimated the gold-copper resource in 2001; construction began around 2010; first concentrate was produced in January 2013 and the first batch of copper shipped July 9, 2013 — roughly 12 years from discovery to production for the open-pit phase. The far larger underground block-cave phase took until 2023 to reach production, showing how single deposits can have multi-decade development arcs.

Why it takes so long

Discovery-to-production clock (2001 → 2013) was dominated not by construction (~2010–2013) but by resource delineation of a deep porphyry system, feasibility work, and an unusually protracted investment-agreement negotiation with the Government of Mongolia (34% state stake) plus assembling one of the largest project financings in mining history (capital costs ballooned from ~$4.6B to ~$10B). Permitting, financing, and host-government terms — not engineering — were the long poles.

  1. 2001 onwardDiscovery / resource estimate and exploration drilling
  2. ~2001–2010Feasibility, Mongolia investment agreement, financing
  3. ~2010–July 2013Construction to first copper shipment (open pit)
Best-supported Open-pit phase: ~12 years (2001 resource estimate → first copper shipped 9 July 2013). Full deposit-to-underground-production arc: ~22 years (2001 → Q1/March 2023 underground block-cave production).

Civil megaprojects

Gotthard Base Tunnel (Switzerland) — world's longest rail tunnel, 57 km

Megaprojectsconfirmed · high
69yrs
1947 → 2016 opened

Swiss engineer Carl Eduard Gruner first proposed a Gotthard base tunnel concept in 1947. After decades of study, Swiss voters/authorities approved the NEAT (New Rail Link through the Alps) program in 1992, which funded the tunnel. Construction by AlpTransit Gotthard AG began in 1999, the final breakthrough came in 2011, and the tunnel was officially inaugurated on 1 June 2016 with commercial service starting December 2016.

Why it takes so long

The clock depends heavily on what is measured. The idea existed for decades before political/financing approval (the 1992 NEAT vote and a 1998 funding referendum were prerequisites). Once launched, the physical work itself took 17 years because of the extreme geology: 57 km of tunneling under up to 2,300 m of overburden, multiple tunnel boring machines, ~28 million cubic metres of excavated rock, and years of commissioning/testing before revenue service. Cross-border alpine rail megaprojects also require sustained multi-decade public funding commitments and democratic ratification.

  1. 1947Initial concept (Gruner proposal)
  2. 1960s–1970Government study committees / recommendation
  3. 1992NEAT program approval (program launch)
  4. 1999–2016 (17 years)Construction
  5. 2011Final breakthrough
  6. 2015–2016Testing/commissioning
  7. June 2016 (commercial Dec 2016)Official opening / operational
Best-supported Concept-to-operational ~69 years (1947→2016); program-launch (NEAT/NRLA approval)-to-operational ~24 years (1992→2016); construction-to-operational 17 years (Nov 1999→June 2016). All three confirmed.

Three Gorges Dam (China) — world's largest hydroelectric / gravity dam

Megaprojectsconfirmed · high
93yrs
1919 → 2012 full operation

Sun Yat-sen first envisioned a great Yangtze dam in 1919. After decades of intermittent study, China's National People's Congress formally approved the project in 1992. Construction began on 14 December 1994; the main dam structure (185 m high, ~2,309 m wide) was completed in 2006; the full power plant reached full operation in 2012 when the last of the 32 main turbines began producing.

Why it takes so long

The earliest 'clock' (1919 vision) reflects nearly a century of political instability, war, and repeated feasibility debate before approval. Even measured from the 1992 approval, full operation took 20 years because the project is staged: damming and structure first, then progressive turbine commissioning and a large underground power plant added after the visible dam wall was finished. Massive resettlement (over a million people), reservoir filling in stages, and sequential generator installation stretch the gap between 'structure complete' (2006) and 'fully operational' (2012).

  1. 1919Initial vision (Sun Yat-sen)
  2. 1992NPC approval (program launch)
  3. Dec 1994Construction start
  4. 2006Main dam structure complete
  5. 2012Full power plant operational (last turbine online)
Best-supported Concept-to-full-operation ~93 years (1919 vision -> 2012); approval-to-full-operation 20 years (1992 -> 2012); construction-to-full-operation 18 years (Dec 1994 -> July 2012), with dam structure complete in ~12 years (1994 -> 2006).

Gorgon LNG Project (Barrow Island, Western Australia) — ~$54 bn, 15.6 Mtpa LNG

Megaprojectsconfirmed · high
35yrs
1980/81 → 2016 first cargo

The Gorgon gas field was discovered by WAPET in 1980/1981. After long development planning and environmental approval for the sensitive Barrow Island nature reserve, Chevron and partners (Shell, ExxonMobil) took final investment decision in September 2009 and broke ground in December 2009. The first LNG cargo shipped in March 2016, with the project ultimately costing roughly US$54 billion (up from an initial ~US$37 bn estimate).

Why it takes so long

The clock that matters most for an LNG plant is FID-to-first-cargo (~7 years here), but the field was known for nearly three decades before development was sanctioned, because monetizing remote/deep offshore gas required prior LNG-market conditions, long-term sales contracts, and a construction-ready cost case. Construction itself ran ~6 years and over budget due to building a giant LNG train complex plus the world's largest CO2 injection/sequestration scheme on a Class-A nature reserve (Barrow Island), with strict quarantine/environmental controls, remote-site logistics, and a major scope including subsea gas-gathering infrastructure.

  1. 1980/1981Gas field discovery
  2. 2000sDevelopment planning / environmental approval
  3. Sept 2009Final Investment Decision (FID)
  4. Dec 2009Construction start (ground-breaking)
  5. March 2016First LNG cargo
Best-supported Discovery (1980/81) to first LNG (March 2016): ~35 years. FID (14 Sept 2009) to first LNG (21 March 2016): ~6.5 years (claim rounds to ~7). Construction (1 Dec 2009) to first LNG: ~6.3 years.

The mechanism

Why manufacturing loops are so long

Across every domain, the same structural forces keep reappearing. None of them are about laziness; most are about physics, consequence, and money.

01 Physics of testing can’t be rushed

Some clocks are set by nature. A gas turbine needs years of grid-synchronized running (the SGT5-8000H ran a ~16-month validation campaign; the GE9X logged 17,000+ hours). Drug trials follow thousands of patients for years to reach statistical power. New superalloys need 50,000–100,000-hour creep tests. Probes lose 5–7 years to orbital mechanics alone.

02 Certification as a multi-year gate

FDA discovery-to-approval averages 10–15 years. Class III implants face the rigorous PMA pathway and can be rejected outright. Nuclear designs undergo first-of-a-kind regulatory review (Olkiluoto’s automation-software disputes; NuScale’s multi-year NRC review). The post-737-MAX tightening visibly lengthened the 777X.

03 Capital intensity & financing cycles

Multi-billion-dollar single assets get funded incrementally, not upfront — which stretches schedules by design. Vogtle’s cost roughly doubled past $30bn; Flamanville rose more than fourfold; EUV needed pooled customer equity; California HSR’s bond covered a fraction of the cost.

04 Supply chains & an atrophied base

The 787’s globally distributed supply chain delivered incomplete subassemblies needing rework. SMRs and Columbia-class submarines must rebuild atrophied supplier networks. Even a ~3.5-year fab build (TSMC Arizona) hit labor and supply-chain friction.

05 Irreversibility sets a hard floor

Where failure is catastrophic or unfixable, exhaustive pre-deployment proving is mandatory. JWST cannot be serviced at L2. Carriers and ballistic-missile subs carry nuclear-surety standards. Dams and tunnels involve mass resettlement and life-safety. None of it compresses.

06 Tooling & the first-article penalty

Clean-sheet products need new tooling, dedicated plants, and have no production learning curve. VW’s MEB needed a converted plant; lead ships and first-of-kind fabs and turbines absorb all the first-article engineering risk that later units never face.

The surprises

Seven counterintuitive findings

01

“It flew / powered up / opened” ≠ “it works in the field.”

The GE9X was certified in 2020 but waits ~5.5 years for its airframe. The geared turbofan “entered service” in 2016 yet needed durability fixes for years. The Three Gorges dam wall finished in 2006 but full operation came in 2012. Entry-into-service is consistently well before field maturity.

02

Construction is the short part.

For mines, S&P finds ~16 years average from discovery to production — of which construction is only ~2.6 years. The rest is resource delineation, permitting, financing, and government and community terms.

03

The trend is getting worse.

Mining lead times rose from ~12.7 years for mines started 2005–09 to ~17.9 years for those started 2020–23. The clocks are lengthening, not shrinking.

04

The “fast” cases prove why the average is slow.

Sovaldi (~6.5 yrs) and Keytruda (~8 yrs) were fast only because hepatitis-C and melanoma offered clean, high-effect-size endpoints plus expedited pathways — precisely the conditions most programs lack.

05

Speed is possible — when the architecture already exists.

The N700S Shinkansen went concept-to-revenue-service in ~4 years, and leading Chinese EVs in ~24 months — both by building on mature platforms with software absorbing post-launch fixes. New infrastructure is the long pole; new rolling stock on an existing base is fast.

06

One local actor can dominate a national-scale clock.

Japan’s Chūō Shinkansen maglev slipped from 2027 to 2035+ largely because a single prefecture (Shizuoka) withheld tunneling permission over groundwater concerns.

07

Some designs outlive their designers’ careers.

The Three Gorges Dam took ~93 years from Sun Yat-sen’s 1919 vision to full operation; the Gotthard Base Tunnel ~69 years from a 1947 sketch; the SynCardia artificial heart lineage ~36 years. The idea can predate the engineering by half a century.