EGNT

Europe is ageing fast.
AI can help it age better.

EGNT builds and backs AI-native products.
Current focus: an MVP for Europe's ageing population. The investment thesis follows.

01Europe 65+ at a glance

81 million Europeans are 65+, 96 million by 2036.

Euro area (EA-21), individuals (m)
AGE GROUP
20262036
10 YRS GROWTH
85+
12.4
15.2
+22%
80–84
11.1
14.2
+28%
75–79
16.3
18.9
+16%
70–74
19.0
23.0
+21%
65–69
22.3
24.4
+9%
Share of total pop.
22%→26%
Connected, 65–74
81%→93%
Digital skills, 65–74
37%→53%
Yearly spending
€1.4trillion
Household wealth
€185kmedian
Home ownership
72%

40m in Italy, France & Spain: ~50% of the euro-area total. By 2035, Italians +65 will be 30% of total population.

36m (45% of total) are active. Travel, attend cultural events, read and socialise regularly. Half take at least one leisure trip a year.

Governments pay the bill. Age-related public spending is 27% of GDP in Italy, 30% in France; peaks in 2045 as 1960s cohorts turn 80.

Source: Eurostat demo_pjangroup 2025 & EUROPOP2023 (euro area, EA-21); Eurostat isoc_ci_ifp_iu, isoc_sk_dskl_i21 (2025, 65–74, euro area); EU-SILC; ECB HFCS; EC Ageing Report 2024. Connected = last 3 months; digital skills = at least basic. 2036 connected and skills = today's 55–64 cohort (calc).

02Active ageing is a must

Active ageing improves lives, saves money and optimizes public resources.

€1.3tn / year
public spend on health and long-term care
€185bn / year
cost of dementia in Europe (7.5m cases > 12m in 2050)
26m live alone
isolation → dementia risk +57%, mortality +29%

Decline starts early. Europeans live to 85 but stay healthy only to 74. The 11 dependent years carry the €1.3tn health and care bill.

Onset can be delayed. 45% of dementia cases trace to modifiable factors. A 2-year delay saves ≈€55bn a year by 2050.

Families absorb the gap. 58% of the dementia bill, over €100bn a year, falls on households as unpaid care by children aged 45-65.

Source: EC Ageing Report 2024; Alzheimer Europe / OHE; Lancet Commission 2024; Kuiper 2015; Alzheimer's Association 2025; Eurostat EU-SILC. "Europe" = euro area (EA-21): EU-27 rates applied to EA-21 population and GDP (calc).

03AI drives active ageing

AI serves body, brain and bonds. Starting with effortless digital access.

Fill the skill gap. In Italy, 73% of 65–74s are connected and 27% are skilled. France and Spain show gaps of 42 and 51 points. AI closes it with zero-skills UX.

Convergence is fast: voice, vision and agents were lab demos in 2023. Frontier accuracy on Humanity's Last Exam went 9% → 50% in 15 months.

Free time is opportunity. 65+ watch TV ~2.8 hours a day, 40% of their free time. AI brings active contents, stimulation, memory and family contact.

Source: Eurostat isoc_ci_ifp_iu, isoc_sk_dskl_i21 (2025, 65–74); Stanford AI Index 2025/2026 (Humanity's Last Exam, Jan 2025 – Apr 2026); Barlovento 2025; ISTAT 2024.

04Robust AI ecosystem is ready

Trillions invested in AI make intelligence cheap, abundant and constantly improving.

Top-5 hyperscaler capex, $bn
230
2024
380
2025
675
2026E
Amazon · Alphabet · Microsoft · Meta · Oracle
$1,300bn
invested in 3 years by 5 companies alone
−99%
cost per task in 2 years, at constant quality

Model quality keeps improving. Suppliers compete, and every upgrade is available on day one.

Unit economics work. Generating 1m words costs <€1 on fast-tier models, trending down.

Access is open. Frontier models are available by API to any team; open models run on-device.

Source: company guidance (Futurum compilation, 2026E $660–690bn); Stanford AI Index 2025 (inference cost, GPT-3.5-level output, Nov 2022 – Oct 2024); Epoch AI. 2024 figure from company reports, approximate.

05Compounding moats

Relationship, orchestration and model economics compound into long-term value.

LayerMoat
Relationship
Memory, tone, pace, consent
MoatCompounding context dataEach conversation adds memory; switching cost grows with time, for the user and for consented partners.
Orchestration
Best model per task, EU rules built in
MoatBest quality at the lowest costTasks routed to frontier, general or on-device models (cost index 100 / 10 / 1); consent and EU data residency built in.
Models
Interchangeable suppliers
MoatHealthy marginsSuppliers compete for the app layer. Model cost is a minor share of the price a user pays, and the margin widens over time.

Context is the key moat. Switching cost grows with every interaction as data compounds. A competitor cannot match the experience on day one.

Compliance lives in the router. Consent, EU data residency and logging are enforced once, for every model behind it. New rules land in one place.

No supplier lock-in. Models are interchangeable and a swap takes days. A supplier outage triggers automatic fallback. The user sees only the improvement.

Source: public API price lists (Sep 2026), cost index indicative; GDPR; EU AI Act (Reg. 2024/1689).

Progress

Concept defined. CEO identified. Team taking shape.
Now focus on target, retention and MVP for validation.

Happy to compare notes with founders and builders in consumer AI, operators who already serve the over-65s, and investors who follow the space. Write to me

Edoardo Giorgetti

egiorgetti@egnt.ai
Edoardo Giorgetti

Spent 30+ years building digital-first products — FinecoBank, ePRICE, Nexi YAP — and scaling them into businesses that served millions of customers and reached hundreds of millions in annual revenues. More on LinkedIn

Each of them brought a bank, a store or a payment card to customers who had never used one online. The over-65s are the same opportunity.

Now working alongside founders building AI-native products, combining their vision, skills and speed with my experience, hands-on work and capital. Write to me

EGNT is my trademark, not yet a company.