The startup funnel

Core claim

What's wrong with Czechs and Slovaks that they do not "start startups"? We have only 308 funded pre-seed companies for 16 million people! If we started them at the Baltic rate, we would have 1,700. Or at the Estonian rate, 3,400. And at the Bay Area rate, 11,000. A startup country can be short of quality or short of quantity. The funnel shows - THERE IS NO PROBLEM WITH THE QUALITY, from those who actually start the funnel on all stages down (or UP if you prefer) to the unicorn is the same as the Baltics. One in three or four reaches Series A, one in fifty becomes a unicorn, on both sides. A Czech or Slovak startup is made of the same thing as an Estonian one. The problem is in Quantity: the Baltics have a third of our people and twice our entrance. Five times per head. Baltics unicorns against CS/SK follow from that alone. So when you chill out on "Early stage events like "Truesdays" you are privileged to be among the most creative, crazy, innovative people. Ths "crows is a rarer crowd than both parliaments (and believe me - with much better vibes :) Another way to imagine it - you can fit ALL of Czech pre-seed founders into one Prague tram. And for Slovakia... there are 1300 bears in Slovakia and only 66 startups on seed stage (and those are type of people that if they meet I would bet on founder, not on the bear). On the diagram there is a nice anomaly - the gold bar with the dollar sign: the Bay Area has more Series A companies than seed. Nothing converts above 100 percent, that is simple - in San Francisco an AI company's first round is already "Series A money". Wordware raised a 30 million dollar "seed". Deepnote, Czech founders, opened with 20 million. Companies like that never enter a pre-seed count. AI era amplified the "jumpers" the founders who can raise that round go to raise it there. The interesting thing that the money is not the constraint. Czech managers raised about 12 billion CZK of new funds in two years; Czech early rounds absorb about a billion a year. But the mandates already point abroad, half foreign by design. Slovakia is adding 90 million euro of state seed money and its fund manager says out loud there are not many startups to reach Series A. Early stage "local" VCs are hunting startups around the world for the same reason. According to investor survey, Czech-only investors went DOWN from 31 percent to 12.5 percent in three years. In my understanding, we don't have quality problem, nor money problem, a quantity problem twice over. A fifth of the Baltic entrance per head, and no AI-sized entrance at all. Every investor here says too few quality startups. The numbers say too few startups! Different problem, different fix, and I do not think a fund is the fix. What can fix it? I don't know honestly, but it's clearly an anomaly, that can be normalized.

LinkedIn post

Key numbers: funded startups by current stage

SINGLE-SOURCE (Dealroom)

Dealroom country and city pages, read 9 Sep 2026; Slovakia read 10 Sep 2026. Stages are Dealroom round-size buckets: pre-seed $100k to 1m, seed $1 to 4m, Series A $4 to 15m, Series B $15 to 40m, Series C $40 to 100m; unicorns $1bn+ valuation. Counts are current stock, not cohorts. Shares and "1 in" ratios are own arithmetic.

Key numbers: per head

SINGLE-SOURCE (Dealroom counts) / populations as noted

Populations used (rounded): Czechia 10.9m, Slovakia 5.4m (Czechia and Slovakia 16.3m), Estonia 1.37m, Latvia 1.87m, Lithuania 2.89m (Baltics 6.1m), Bay Area 7.7m. Geography for the Bay Area column is the nine-county San Francisco Bay Area (Alameda, Contra Costa, Marin, Napa, San Francisco, San Mateo, Santa Clara, Solano, Sonoma). It is not the SF-Oakland-Berkeley MSA (CBSA 41860), which is smaller and would move the per-head rates. Sources: Eurostat tps00001, Population on 1 January 2025 (Czechia 10,909,500; Slovakia 5,419,451; Baltic states from the same table, summed). Bay Area printed 7.7m from Census 2020 PL 94-171 nine-county total 7,765,640 (MTC Bay Area Census Data compilation). ACS series for the same nine counties: Census Bureau table B01003 Total Population, ACS 5-Year Estimates 2019-2023 (ACSDT5Y2023.B01003); county sum not re-pulled on 21 Sep 2026 (Census API key required). Pull date: 21 Sep 2026.

The ladder

What Czechia and Slovakia would have at another region's per-head rate, using 16.3m times the rate per million:

• Baltic rate 103 × 16.3 = 1,678.9, printed 1,700.
• Estonian rate 206 × 16.3 = 3,357.8, printed 3,400.
• Bay Area rate 682 × 16.3 = 11,116.6, printed 11,000.

Rounding note: 626 / 6.1 = 102.6, printed 103. An earlier draft used 102 from a 6.13m population.

Key numbers: Czech venture funds, 2024 to 2026

CONFIRMED per row by the source named; sum is own arithmetic at 25 CZK/EUR

Slovak state money

Label: state money only, a floor; Slovak private funds not counted

SIH tender for three seed funds, 90m EUR total, 30m each, 2026 to 2030, tickets 50k to 1m EUR (oPeniazoch.sk, 10 Mar 2026; also SIH). Venture to Future Fund raised to 25m EUR by end 2026, tickets 400k to 7m EUR (Forbes.sk). The post's "its fund manager says out loud" refers to the VFF investment-manager line on scarcity of Series A or B startups; confirm the exact quote line against the Forbes.sk piece if challenged.

Deployment ("about a billion a year")

Estimate, arithmetic shown. 2025 about 100 VC deals (Ambit VC Deals Report 2025 via BusinessInfo) or 79 (Rolu via Lupa, 6 Jan 2026), over 80 percent early stage, early-stage median 500k EUR. Own floor estimate: 80 rounds × 500k EUR = 40m EUR, about 1bn CZK.

DEPO Ventures investor survey

DEPO Ventures is the author's own fund; this page says so in the same sentence. Czech-only investors 31 percent (2023 edition) to 25 percent (2024) to 12.5 percent (2026). Sources: 2024 edition reported via BusinessInfo (25 percent Czech-only); 2026 edition via BusinessInfo (12.5 percent).

The winks

Parliaments

CONFIRMED (constitutions)

Czech Chamber of Deputies 200, Senate 81 (Constitution of the Czech Republic, art. 16). Slovak National Council 150 (Constitution of the Slovak Republic, art. 73). 431 seats in the two countries against 308 funded pre-seed startups. Unit note: startups are companies, not founders; at two founders per company the people count passes the seat count.

The tram

CONFIRMED (Škoda Group and Wikipedia)

A Prague Škoda 15T ForCity carries about 300 passengers (Škoda Group ForCity Alfa Prague: up to 180 passengers at 4 per m², 61 seated; page also lists total 284 under parameters, so the ~300 figure is approximate; Wikipedia gives the model's range across configurations as 43 to 72 seated, 150 to 539 standing; both read 10 Sep 2026). 242 Czech pre-seed startups at one founder each fit in one tram; at two founders each they need two. The Bay Area's 5,252 need 18 trams at one founder each (5,252 / 300 ≈ 17.5, printed 18).

The bears

CONFIRMED for the Slovak government figure (two press reports) plus the 2013 scientific range

As posted: "1300 bears in Slovakia and only 66 startups on seed stage." Beside that line, Dealroom Slovakia: 66 is pre-seed, not seed; seed is 29; funded startups at any stage are 130. The quote is not rewritten.

Slovakia bear count: about 1,300 brown bears, the 2025 government figure reported when the cull of 350 was approved (The Week, 7 Apr 2025; CNN, 3 Apr 2025). Scientific estimates 800 to 1,100 for the mountain population (Kaczensky et al. 2013) and 600 to 900 (European Commission, 2012), via carnivores.eu.

Czechia for contrast: 1 to 3 bears (Kaczensky 2013) or "at least five" (Radio Prague, 22 Jun 2006), against 242 pre-seed startups. The bear line holds for Slovakia and not for Czechia.

The AI effect (the gold bar)

Dealroom buckets by round size, so a first round above 4m USD lands in Series A and never in the pre-seed count.

• Wordware, 30m USD seed, 2024 (Vestbee).
• Deepnote, posted as opening with 20m; corrected by the author (see corrections log): 3.8m seed in 2020 (Index, Accel, YC), then 20m Series A in 2022 (TechCrunch, 31 Jan 2022). Both led from abroad.
• ElevenLabs, 424m EUR, Feb 2026 (EU-Startups). Not in the posted text; on this page.

Method

Dealroom "pool of funded startups" pages, one read per country on the dates given; stage counts copied as shown; shares and "1 in" ratios are own arithmetic on stock, not cohort tracking; per-head figures and the ladder use the populations above; the Bay Area is a metro against countries. "Quality" in the post means the conversion ratios in table 1, the author's definition, and nothing else. Same author, same method as cars per employee.

Critical context

• Single source for all startup counts.
• Stock not cohorts, so "one in fifty becomes a unicorn" is a ratio of two stocks.
• The ladder assumes a linear per-head rate and is a comparison, not a forecast.
• The Bay Area Series A count above its seed count is partly definitional (round size) and partly residence and inflow, unmeasured.
• "Funded" excludes unfunded startups everywhere equally.
• Fund sizes mix targets, first closes and announcements.
• VC-invested headlines on the Dealroom pages have no stated period and are not used.
• Estonia's unicorn count differs within its own page.
• Slovak money is state money only.
• Bear counts differ by a factor of two depending on who counts.
• Tram capacity depends on the standing density assumed.

Falsifier

The posted text ends without a falsifier (owner's choice). Stated on this page: two Dealroom reads from now, a Czech and Slovak pre-seed count growing faster than the Baltic one with the shape unchanged; or a CEE-based company opening with a Series A sized round and staying. Next read: 10 Dec 2026 (first read). Call 9 resolves on the second read, around Sep 2027. When /calls/ exists, that is the home for call 9.