Skip to content

InsightsThe Calaa Global publication

Eight units and one publication. Each company and each division of the group writes about its own field, and here they are all together. The rule is the same for all eight: an analysis that does not bring a data point nobody else has is not published.

  1. 01The foundations that order the group
  2. 02What only appears when the industry's knowledge is cross-referenced
  3. 03Industry intelligence, evaluated against general models
  4. 04The research centre that publishes its tests

The foundations that order the group

The founder's three theses on the automotive industry, with their annual follow-up. Each one is tested against what the industry does, and what holds and what does not is published.

Read the foundations →
Foundation I2026

Fragmentation

The 100,000 million blind spot

A vehicle's history is split among the carmaker, the dealer, the workshop and the insurer, and none of them sees it whole. That fragmentation leaves more than $100 billion a year uncaptured.

Read the foundationDownload the PDF
Foundation II2026

The digital twin

Beyond telematics

Telematics records where a vehicle is. The digital twin records how it has lived, and with that verified history, maintenance, used-car sales, insurance and the workshop all change.

Read the foundationDownload the PDF
Foundation III2026

Data arbitrage

From raw data to residual value

Raw, the data a vehicle generates is noise. Refined and verified, it becomes an asset that raises its resale value, lowers the cost of insurance and shortens repairs.

Read the foundationDownload the PDF

The 2026 review

Last review · November 3, 2026
10 / 22
claims from 2025 already borne out
3
Theses
41
Notes published in 2026
60
Verified sources

The status of the claims

22 claims from the 2025 theses, this year

ConfirmedOn trackQualifiedNot confirmed
IFragmentation9
IIThe digital twin6
IIIData arbitrage7

10 Confirmed5 On track7 Qualified0 Not confirmed

The group's pieces, against the theses

3 in service, 3 with a date

The piece202620272028
CodexIIIIIIThe intelligence layerIn service
GammaIIIIIIThe intelligence layerIn service
Calaa QuantumIIIIIIThe horizonUnder way
Cal-1IIIIIIThe experience layerIn stores on 30 November 2026
KRN-01IIIIIIThe physical layerLaunch in 2027
Cal-1 LinkIIIIIIThe physical layerArrives in 2028

The latest developments

New idea

Calaa Quantum takes the twin into all its futures

Thesis I
The figure, updated

GM pays $220 million for selling driving data

Thesis IIIClaims Journal
What has happened

Chinese brands overtake Japanese brands in Europe

Thesis IDataforce, via To Vima

More than 35 million records and more than 275 million data points

See Codex →
35,182,693
records
277,604,017
data points
7
official registries in the US, the EU and the UK
Finding 0110 Nov 2026Complaints do not announce recalls: they follow them×1.6In the months after a recall campaign, complaints about that part surge: 1.6 times more than for the rest of the car. Before the campaign they move no more than the others.Measured on 10,910 campaigns (2012-2024) and more than 2 million complaints.Finding 0210 Nov 2026No early warning4.9 % vs 6.0 %We tried raising the alarm when complaints about a part doubled. It ended in a campaign 4.9 % of the time; without an alarm, 6.0 %. Complaints do not serve to get ahead: we checked before promising it.Measured on 10,910 campaigns (2012-2024).Finding 0310 Nov 2026Most recalls are not requested by the regulator1 in 18Only 1 in 18 recall campaigns stems from an official investigation. The rest are announced by the manufacturer itself.Measured on 10,910 campaigns (2012-2024).Finding 0410 Nov 2026From investigation to recall110 daysWhen the regulator opens an investigation that ends in a campaign, the campaign arrives after 110 days, at the median. In half of the cases, between 39 and 235 days.Measured on 605 campaigns.Finding 0510 Nov 2026One part, eight manufacturers8 groupsA single supplier appears in recall campaigns from 8 different manufacturer groups in the last year. A fault in one part does not stay in one brand.12 months to September 2026.Finding 0610 Nov 2026Shared suppliers32In the last year, 32 suppliers appeared in campaigns from two or more manufacturer groups, and 14 in three or more. We have measured it for ten years: between 20 and 33 each year.12 months to September 2026; series since 2016.Finding 0710 Nov 2026An ever wider supply chain+56 %The number of distinct suppliers involved in recall campaigns rose from 93 in 2017 to 145 in 2024.12-month windows, 2017 and 2024.Finding 0810 Nov 2026The end of an eraFrom 17 to 2The supplier behind the largest wave of campaigns of the decade went through 17 manufacturer groups. In the last year, it appears in only 2.Campaigns from 2015-2024 against the 12 months to September 2026.Finding 0910 Nov 2026The battery enters the sceneCellsBattery cell manufacturers now appear as suppliers in recall campaigns from several manufacturer groups. The electric car is starting to leave its mark on safety records.12 months to September 2026.Finding 1010 Nov 2026Who makes the part6 in 10In 6 out of 10 recall campaigns, the defective part was made by an outside supplier, not by the carmaker.Measured on campaigns from 2015 to 2026.Finding 1110 Nov 2026One company, fifty namesMore than 50In official reports, the same supplier company is written in more than 50 different ways. Unifying them is what shows who is behind each fault.Campaign reports from 2015 to 2026.Finding 1210 Nov 2026One in five fails21.5 %In the United Kingdom, more than 1 in 5 roadworthiness inspections ends in a fail: 146.7 million out of 681.7 million.Inspections from 2005 to 2023.

Industry intelligence, evaluated against general models

Real automotive questions, the same question to Calaa and to the general models, and the result scored. Correct answers, figures quoted verbatim and invented answers, with their date.

See the evaluation →
Jan 2027first definitive evaluation
EvaluationJan 2027

Calaa against general models in automotive

First publication
Case

The same question, answered by a general model and by Calaa

Pending
Case

A figure quoted verbatim against an invented figure

Pending
Principle

If it does not know, it says so

How Calaa reasons

The research centre that publishes its tests

Experiments, results and what appears when the whole industry is cross-referenced. Every report comes out with its date and its method, and its results can be checked.

Go to Quantum →
Line 01 · The Codex graph

On another machine, the circuit beats chance again.

· ibm_kingston · Heron r2 · Washington DC (us-east)View the researchDownload the PDF
Line 03 · Battery chemistry

Three ZNE measurements, one average.

· ibm_kingston · Heron r2 · Washington DC (us-east)View the researchDownload the PDF
Line 03 · Battery chemistry

The repeat does not confirm the improvement.

· ibm_kingston · Heron r2 · Washington DC (us-east)View the researchDownload the PDF
Line 03 · Battery chemistry

Correcting noise works, and it costs precision.

· IBM Quantum · Heron r2 · Washington DC (us-east)View the researchDownload the PDF
Line 01 · The Codex graph

Today's hardware keeps almost all of the circuit.

· IBM Quantum · Heron r2 · Washington DC (us-east)View the researchDownload the PDF

Research progress

Updated · November 18, 2026

The test calendar

Twelve tests on a real processor, one a day.

From 9 November to 11 December 2026. Each test is first prepared on the simulator and compared with an ordinary computer. What comes out worse is published too.

5 of 12 tests done

Next testNovember 20The hidden network, thirdQAOA · 12 qubits
  1. Nov 7
  2. Nov 93
  3. Nov 101
  4. Nov 122
  5. Nov 16
  6. Nov 174
  7. Nov 185
  8. Nov 19
  9. Nov 206
  10. Nov 237
  11. Nov 248
  12. Nov 259
  13. Nov 2610
  14. Nov 2711
  15. Nov 3012
  16. Dec 2
  17. Dec 6
  18. Dec 10
DoneNextPlannedPreparationPublication

Line 03 · Battery chemistrySimulator

The hydrogen molecule curve

-1.2-1-0.8-0.60.511.522.5Distance between the atoms (Å)Energy (hartree)Minimum · 0.75 ÅExactVQE on a simulatorHartree-FockReal processor · TREXReal processor · ZNEZNE average · 3 tests

The energy of H2 as a function of the distance between its two atoms, at 45 points. The VQE on a simulator lands on the exact solution at every one, with a maximum error of two hundred-millionths of a hartree. Hartree-Fock, the classical starting method, drifts away as the bond stretches. The minimum sits at 0.75 Å, as in the real molecule.

Conclusions

Each conclusion, with how firm it is

Line 03 · Nov 7, 2026Firm

On a real processor today, noise still dominates.

Line 03 · Nov 6, 2026Firm

The method reproduces the exact energy of hydrogen.

Line 03 · Nov 6, 2026Firm

With today's noise, the result is not yet useful for chemistry.

The car's environment, observed from space.

KRN studies the road, water, heat and air where cars live, and crosses them with Codex. It starts with the public satellites of Copernicus and NASA. It will go on with its own.

View KRN →
Copernicus
01 · Floods and the cars they damage

In the Valencia DANA, six in ten insurance claims were for vehicles

57.9 %of insurance claims were for vehiclesCopernicus Emergency Management Service (© 2024 European Union), [EMSR773] Horta Sud (AOI03): Delineation Map · Copernicus Emergency Management Service (© 2024 European Union), [EMSR773] Valencia province (AOI01): Delineation Map · © colaboradores de OpenStreetMap, ODbL
02 · Floods and the cars they damage

In the East Midlands, the 2023 storms flooded 5,851 hectares and only 6 km of main roads

14.8 %of claims were for vehicles, against 57.9% in ValenciaEnvironment Agency, Recorded Flood Outlines. Contains public sector information licensed under the Open Government Licence v3.0 · © colaboradores de OpenStreetMap, ODbL
03 · Floods and the cars they damage

More than half of the 2024 DANA floodwater fell outside the official flood zone

56.5 %of the 2024 footprint fell outside the official flood zoneCopernicus Emergency Management Service (© 2024 European Union), [EMSR773] Horta Sud (AOI03): Delineation Map · Copernicus Emergency Management Service (© 2024 European Union), [EMSR773] Valencia province (AOI01): Delineation Map · © Ministerio para la Transición Ecológica y el Reto Demográfico, Zonas inundables con periodo de retorno de 100 años (SNCZI), servicio WMS INSPIRE de peligrosidad por inundación (IDEE) · © colaboradores de OpenStreetMap, ODbL
04 · Floods and the cars they damage

Three weeks after DANA «Adam», radar still saw water on 188 hectares

188hectares were still under water 22 days after DANA «Adam»Copernicus Emergency Management Service (© 2026 European Union), [EMSR935] Moraira (AOI01): Delineation Map · Copernicus Emergency Management Service (© 2026 European Union), [EMSR935] Gandia (AOI02): Grading Map · Contains modified Copernicus Sentinel data 2026 (Sentinel-1 RTC, Microsoft Planetary Computer) · © colaboradores de OpenStreetMap, ODbL
05 · Industry from above

SQM sold 31% more lithium without its Atacama ponds growing

30.7 %more lithium sold by SQM in 2024 than in 2022, with the same pond surfaceContains modified Copernicus Sentinel data 2016–2026 (Sentinel-2 L2A, Microsoft Planetary Computer)
06 · Industry from above

Sagunto already has 36 hectares of buildings; Figueruelas, 4.2; Navalmoral, still bare earth

36hectares of new roof at PowerCo's site in Sagunto, October 2026Contains modified Copernicus Sentinel data 2021–2026 (Sentinel-2 L2A, Microsoft Planetary Computer)
07 · Industry from above

In April 2020 almost no cars left the Port of Barcelona, and its yards were fuller than ever

10,832cars passed through the Port of Barcelona in April 2020, and its yards were fuller than ever© colaboradores de OpenStreetMap, ODbL
08 · The air

From space, Madrid's NO₂ fell 22% in 2020, just as its 24 stations measured

-22 %Madrid's NO₂ from space in 2020, just as its 24 stations measuredCopyright © KNMI/ESA, TEMIS, «Monthly mean tropospheric NO2 from TROPOMI», version 1.0. Contains modified Copernicus Sentinel-5P data · © colaboradores de OpenStreetMap, ODbL