Latest / Blockchain Germany – Web3 Startups, Crypto Innovation & Venture Capital by Startuprad.io™ / E 773 — Startup News August 2026: BaFin MiCAR Authorization, Stripe/OpenRouter, and the Concentration Economy

Transcript

E 773 — Startup News August 2026: BaFin MiCAR Authorization, Stripe/OpenRouter, and the Concentration Economy · Blockchain Germany – Web3 Startups, Crypto Innovation & Venture Capital by Startuprad.io™

0:00
This is episode 773 of Startuprad.io:
0:02
Startup News August 2026, recorded for
0:05
publication on Labor Day, Monday, September 7, 2026. Ten,
0:09
minus eleven, twenty-seven — those are three numbers that
0:14
frame the story of the German startup ecosystem this
0:17
month: ten new unicorns in 2026 so far, eleven percent
0:21
fewer financing rounds in the first half of the year,
0:25
and twenty-seven percent of German tech startups
0:29
saying AI caused them to forgo new hires. This is the
0:33
concentration economy, and it changes everything.
0:37
This is Startuprad.io. I'm Joe Menninger,
0:52
joining you from Frankfurt am Main, Germany. This is the
0:55
startup news for Germany, Austria, and Switzerland for
0:59
August 2026. As a tradition established with
1:03
Chris, we publish our summer wrap-ups around Labor Day.
1:07
If you celebrate it, happy Labor Day. This is
1:11
part two of our summer wrap-up, and this is
1:15
episode 773 of the Startuprad.io franchise. This
1:18
month's thesis is what I'm calling the
1:22
concentration economy. For years, we have tracked this
1:26
ecosystem's growth in aggregate: total funding,
1:30
unicorn counts, deal flow. And the
1:34
headline numbers still look strong. Germany has
1:38
produced ten new unicorns in 2026 so far. Amazing.
1:43
But look beneath the surface and you find a very different story.
1:47
EY's data for the first half of 2026 shows about 5.3 billion euros invested in
1:50
German startups, up fourteen percent from the same period last
1:54
year, while the number of financing rounds fell eleven percent to 354. Sixty-seven
1:58
percent of all capital went into rounds above 50 million euros. And an
2:02
August Bitkom survey found that twenty-seven percent of German tech
2:05
startups had forgone new hires because of AI; seven percent had cut jobs because of AI,
2:09
while sixteen percent had added staff because of it. The boom is
2:12
real, but its benefits are concentrating. That is the concentration economy.
2:16
Before we get into the detail, let's check our
2:19
predictions from the July episode. Prediction
2:23
one: Helsing reaches a 25 billion US dollar valuation
2:27
within twelve months. Not yet. The defense AI
2:31
company continues to expand across NATO markets,
2:35
but no new funding round was announced in
2:38
August. We are still watching. Prediction two: two
2:42
or more additional German defense unicorns
2:46
emerge by the end of 2026. This month Cambridge
2:50
Aerospace reached a 3.4 billion US dollar valuation,
2:54
which is impressive, but it is a UK company,
2:57
not a German one. So that prediction remains open.
3:01
We need additional German-headquartered defense startups
3:06
crossing the billion-dollar line. Prediction three: the
3:09
pension-reform pipeline reaches 5 billion by mid-2027. This month
3:13
gave us a useful product signal, but not yet a
3:16
capital-flow confirmation. Scalable Capital is already
3:19
allowing customers to add its planned Altersvorsorgedepot and
3:23
promoting the state-supported retirement product ahead of the new
3:27
framework taking effect in 2027. That is evidence of
3:30
competitive preparation, but we still need to see actual assets
3:34
flow before calling the prediction fulfilled.
3:39
We are also still tracking the June prediction that Isar
3:43
Aerospace reaches orbit by December 31. No decisive update in August.
3:46
The launch target remains
3:50
later this year. Let's talk
3:56
startups. The macro view — maybe you call
4:00
it the unicorn factory. Germany
4:03
has produced ten new unicorns
4:07
in 2026 by late August. Let
4:11
that number sit for a moment. On the
4:15
surface, this is an
4:18
extraordinary success story. But when you
4:22
look at who is crossing the
4:26
billion-dollar line, a pattern emerges.
4:30
These are largely companies that have achieved
4:34
meaningful revenue, real enterprise traction, or strategic
4:38
relevance in areas such as defense, industrial technology,
4:42
and deep tech. This is not the 2021 vintage, when
4:46
narrative and growth metrics alone could carry more of the
4:50
valuation story. Take Moss as an example. On August 5,
4:54
the corporate-spend platform announced a 35 million
4:57
euro Series C at a 1 billion euro valuation. Its
5:01
product has expanded from corporate cards into
5:05
invoice management, travel and expense workflows,
5:09
and finance automation. Or look at Camunda, whose
5:14
annual recurring revenue is nearing 200 million US dollars
5:18
and is expected to exceed that level before year-end.
5:21
This is real software revenue at scale, not just a
5:25
projection. The unicorn factory is producing a very
5:29
specific kind of company: businesses with stronger operating
5:33
proof than the market demanded a few years ago.
5:37
There's an important structural point here. In the
5:41
2020 and 2021 boom, a billion-dollar valuation
5:44
could be reached with a very different mix of growth,
5:48
narrative, and capital-market expectations. In 2026,
5:52
investors are putting more weight on revenue
5:55
quality, unit economics, strategic
6:00
relevance, and a credible path to durable scale. That filter
6:04
may be healthy in the long run, but it also
6:08
makes the unicorn count less representative of broad
6:12
ecosystem strength. Ten unicorns can sound like a rising
6:16
tide, while much of the market below them faces a tougher
6:19
funding environment. The funding
6:23
paradox is where the concentration thesis really takes shape.
6:27
And the clearest structural data point comes
6:30
from the first half of 2026, not from August alone.
6:34
EY recorded about 5.3 billion euros of venture
6:38
capital invested in German startups in the first half of 2026,
6:41
fourteen percent more than a year earlier. But the number of
6:46
financing rounds fell eleven percent to 354. Seventeen rounds
6:49
exceeded 50 million euros, and those large transactions captured
6:53
67 percent of all invested capital. Let me repeat
6:57
that: roughly two-thirds of the capital went into rounds above
7:01
50 million euros. This is the funding paradox of the
7:05
concentration economy: more capital than in the first half of 2025, but
7:09
spread across fewer financings and concentrated heavily in
7:13
large rounds. That does not mean every early-stage segment
7:17
follows the same pattern, but it clearly raises the importance of
7:21
differentiation, traction, and capital efficiency for founders.
7:24
And to put the shift into a longer perspective,
7:28
compare this with the peak. In the first half
7:32
of 2021, EY counted 588 German startup
7:36
financing rounds and about 7.6 billion euros of
7:40
investment. In the first half of 2026, it counted 354
7:44
rounds and about 5.3 billion euros. That is roughly
7:48
forty percent fewer rounds than at the 2021 peak.
7:51
The ecosystem still attracts substantial capital,
7:55
but participation has narrowed. That
7:59
matters beyond venture funds themselves. Law firms,
8:02
recruiters, coworking providers, accelerators, and
8:06
other service businesses often depend much more on
8:09
the number of active companies and transactions
8:13
than on the size of a handful of mega-rounds.
8:17
The AI labor shock. A Bitkom
8:21
survey released in August found that
8:25
twenty-seven percent of German tech
8:29
startups had forgone new hires during the
8:32
previous twelve months because of artificial
8:36
intelligence. Seven percent said they had
8:40
cut jobs because of AI. But sixteen
8:44
percent reported the opposite: they had hired
8:49
additional staff because of AI. Half said AI
8:53
had no effect on their personnel needs.
8:56
So the signal is not simply 'AI kills
9:00
startup jobs.' It is a bifurcation in what
9:04
kinds of work startups need and how much
9:08
output they expect from each employee.
9:11
On one side, AI can reduce demand for some
9:15
planned administrative, analytical, content,
9:19
support, or software-development work. On the
9:23
other, companies building, deploying,
9:26
integrating, and governing AI systems need
9:30
specialized people. Bitkom also reports that the average
9:34
German startup in its survey employed twelve
9:37
people, down from thirteen a year earlier.
9:41
That connects directly to the concentration thesis.
9:45
If a seed-stage founder can build more with a
9:49
smaller team, the company may need less capital and
9:52
fewer hires to reach the same milestone. That
9:56
can be genuine efficiency, while still reducing
10:00
the number of funding rounds, offices,
10:04
service contracts, and entry-level roles
10:07
generated per startup. The ecosystem can grow in value
10:11
without growing proportionally in participation.
10:15
The infrastructure layer. Three
10:18
developments deserve attention. First, High-Tech
10:22
Gründerfonds announced fundraising for HTGF V, its
10:26
fifth seed fund, on August 26. The important
10:29
signal is that HTGF is positioning itself as a
10:33
multi-stage platform that can connect seed investing
10:37
with later-stage support, while continuing to
10:41
combine public anchor investors with private
10:45
capital. Second, Scalable Capital is already
10:49
marketing and allowing customers to add its planned
10:53
Altersvorsorgedepot ahead of
10:57
the new state-supported retirement framework
11:01
taking effect in 2027. That matters because a
11:05
successful capital-market pension wrapper could
11:09
eventually create a much broader distribution
11:13
channel for long-term investment products.
11:16
If Scalable or its competitors eventually
11:19
build venture-adjacent exposure into these pension
11:23
products, the potential distribution channel reaches
11:27
millions of German savers. That would be a structural
11:30
change in how long-term capital reaches the innovation
11:34
economy, although that step has not happened yet.
11:37
Third, Bavaria announced roughly 30 million
11:41
euros of support for WERK1 in Munich for the
11:44
period from 2027 to 2032. That is not a new 30 million
11:48
euro venture fund; it is infrastructure funding for one
11:52
of Bavaria's central startup hubs. The
11:55
broader signal is still relevant: regions are competing
11:59
for startup activity not only with branding, but with
12:03
long-term ecosystem infrastructure. Top signal
12:07
segments. Segment one: Moss and the fintech pulse.
12:10
Moss crossed the 1 billion euro valuation
12:14
mark on August 5, and it is worth
12:18
understanding why this particular unicorn
12:21
matters for the broader fintech thesis. Moss
12:25
started with corporate cards and has
12:29
expanded into a broader finance platform,
12:33
including invoice management, travel and
12:36
expense workflows, and AI-supported finance
12:41
operations. The strategic lesson is
12:45
cross-sell: win a finance team with one product,
12:49
build trust inside the workflow,
12:52
then add adjacent products on top of the
12:56
existing relationship. Meanwhile,
13:00
Berlin-based Pliant has surpassed 100 million US
13:04
dollars in annual recurring revenue and
13:07
has expanded into the United States. That
13:12
creates a serious competitive dynamic in
13:16
European B2B payments and spend management.
13:20
The fintech sector is moving beyond an era
13:24
of many small players jockeying for position.
13:28
The stronger companies are increasingly
13:32
competing through platform breadth,
13:37
infrastructure, cross-sell, and international
13:41
scale. Segment two: the robotics surge.
13:45
If there is one DACH sector that moved
13:48
from interesting to unmissable this
13:52
month, it is robotics and physical AI. NEURA
13:56
Robotics expanded its platform through
14:00
two concrete acquisitions: ACTIVE Shuttle from
14:04
Bosch Rexroth and 100 percent of ADLATUS
14:08
Robotics. The strategy is clear — bring
14:11
different robotic applications onto a
14:15
shared physical-AI infrastructure rather than
14:18
scaling only one machine or one use case.
14:22
Gravis Robotics added another major
14:26
signal. The Swiss ETH spin-off announced a 200
14:30
million US dollar Series A on August 17 to
14:33
scale autonomous heavy machinery, and
14:37
ETH reported that the round put the company
14:41
at unicorn status. So this is not only a
14:44
German robotics story; it is a DACH
14:48
physical-AI story. Capital is backing systems
14:52
that combine software, autonomy, sensors,
14:56
and heavy industrial hardware.
14:59
Segment three: AI infrastructure
15:03
consolidation. By late August, reports said NVIDIA was
15:07
in talks to acquire Hugging Face. That
15:10
transaction was formally announced on
15:14
September 3, after the August cut-off, at
15:18
roughly 13 billion US dollars. So for
15:21
August, the correct signal is
15:25
acquisition talks, not a completed funding round.
15:28
There were other major infrastructure deals
15:32
around the same time. Stripe was reported
15:35
to have finalized a deal to acquire
15:39
OpenRouter for more than 7 billion US
15:42
dollars, although Stripe itself had not
15:45
publicly confirmed the transaction. Lovable raised
15:49
400 million US dollars in Series C funding at
15:53
a 13.3 billion dollar valuation — not a 400
15:56
million dollar valuation. And Italy's
16:00
Bending Spoons entered a definitive agreement to
16:04
acquire Airtable in a transaction valuing
16:08
Airtable at 1.285 billion dollars in
16:12
enterprise value, with an implied equity
16:15
value of roughly 2.25 billion dollars. What do
16:18
these international deals mean for DACH? The
16:22
infrastructure and application layers that
16:26
European startups depend on are consolidating
16:30
quickly, and control is shifting toward a
16:33
smaller set of very well-capitalized platforms.
16:37
That creates both opportunity and risk. Better
16:41
infrastructure lowers the cost of building AI
16:44
applications. But dependence on a small number of model,
16:48
cloud, chip, routing, and development platforms
16:52
creates exposure to pricing changes, access rules,
16:56
and competitive shifts. German and European founders
17:00
building on top of those layers need to be explicit
17:04
about where their defensible moat actually sits. There
17:08
is also a European sovereignty angle, but it needs
17:12
precision. Not every major deal is American-on-American —
17:15
Bending Spoons is Italian, and Lovable is
17:19
Swedish. Still, many of the most critical model,
17:23
compute, cloud, and AI-routing layers remain concentrated in
17:26
US-controlled platforms. If European policymakers are
17:30
serious about digital sovereignty, infrastructure
17:34
is one of the layers that matters most.
17:37
The restructuring and insolvency wave tells
17:41
the other side of the concentration economy.
17:45
Recent examples include Cultimate Foods, the
17:49
Berlin cultivated-fat startup that entered
17:53
preliminary insolvency proceedings at the end of
17:56
April, and Sono Motors, which opened insolvency
18:00
proceedings in August and put the Sono Solar
18:04
business and intellectual property up for
18:08
sale. Glow25, the collagen and wellness
18:11
brand, also announced significant layoffs, with
18:15
management explicitly pointing to
18:18
greater use of AI and automation. And Pleo, the
18:22
Danish expense-management fintech with a
18:26
significant European footprint, faced public
18:29
scrutiny after former employees described a much
18:33
tougher culture, executive turnover, internal
18:36
friction, and concerns about the company's direction.
18:40
When former employees are telling
18:44
journalists that a company has 'lost its soul,'
18:47
you do not just have an HR story; you
18:51
have a potential brand, retention, and
18:55
execution story. The broader pattern is
18:59
consistent with the concentration thesis:
19:02
capital is concentrating at the top,
19:06
while weaker or more capital-intensive
19:10
businesses face harder financing decisions,
19:13
restructuring, or failure. The middle of the
19:17
distribution can get squeezed. The defense
19:20
rotation also continued in August. Cambridge
19:24
Aerospace raised 300 million US
19:28
dollars in a Series C at a 3.4 billion dollar
19:31
valuation. But precision matters: Cambridge
19:35
Aerospace is a UK company, not a German one.
19:39
So while Cambridge Aerospace
19:43
confirms the broader European defense-tech thesis,
19:46
it does not advance our prediction
19:50
about additional German defense unicorns.
19:54
Germany already has major defense-tech unicorns
19:58
including Helsing and STARK. The
20:01
question is whether the pipeline beyond
20:04
them — in drones, autonomous systems,
20:08
satellite communications, cyber, sensing,
20:12
and logistics — produces additional
20:15
billion-dollar companies before
20:19
year-end. Our prediction says two more after the
20:22
July baseline. The clock is ticking.
20:26
Lightning round: six stories, about thirty
20:29
seconds each. Number one: Julian Teicke,
20:33
the founder of wefox, has started a new
20:37
company called Attuned with Rebecca Godfrey and
20:41
Thomas Hübl. Teicke is building again.
20:44
Attuned is focused on real-time human co-regulation
20:48
rather than another AI assistant, and its public
20:52
beta is scheduled to open in September. The
20:55
serial-founder signal is clear: experienced European
20:59
founders are recycling back into the ecosystem and
21:03
starting again. Number two: Waymo is laying
21:06
the groundwork for a fully autonomous
21:10
ride-hailing service in Munich. It plans manual mapping and
21:14
validation first, and says it aims to open
21:18
commercial ride-hailing to the public toward the end of
21:21
2027. If that happens, Germany becomes a serious test
21:25
market for large-scale autonomous mobility. Number
21:29
three: smart glasses are facing renewed privacy
21:32
scrutiny in Germany. Legal experts quoted by German
21:36
public media say a ban could be conceivable
21:39
under certain circumstances because covert recording
21:43
creates difficult consent and privacy questions.
21:47
This is the kind of issue where German law and
21:51
enforcement could influence how ambient-computing
21:55
hardware develops across Europe. Number four:
21:59
Carlsen Verlag — not chess champion Magnus Carlsen —
22:04
is suing OpenAI together with author Marc-Uwe
22:08
Kling and illustrator Astrid Henn. The case
22:11
concerns alleged infringement involving their work
22:16
'Das NEINhorn' and goes directly to the question
22:19
of how copyrighted text and illustrations may be
22:23
used or reproduced by generative AI.
22:27
Number five: Bavaria is committing roughly 30
22:31
million euros to WERK1 for 2027 through 2032.
22:35
Again, that is startup-infrastructure
22:38
funding, not a new venture fund. Number six:
22:41
BaFin's crypto supervision. In an August interview on
22:45
MiCAR, BaFin expert Ruth Burkert said that
22:48
a BaFin license is viewed in the market as a
22:52
'seal of quality.' For crypto companies, that
22:56
underscores how regulatory credibility can become
22:59
a commercial asset across the European market.
23:03
Regulation, in this case, can become part of market
23:07
positioning. And one more story cuts directly to the
23:11
thesis of this episode. The Business Angels
23:14
Club Berlin-Brandenburg published a position paper
23:18
warning about the deterioration of early-stage
23:22
funding in Berlin. According to the BACB, pre-seed
23:25
funding volume in Berlin has fallen 65 percent
23:29
since 2022, back to roughly 2017 levels. The median
23:33
angel ticket has fallen from 50,000 euros to 38,000
23:37
euros. Startup insolvencies have nearly doubled since
23:41
2022, even while startup formations have increased.
23:45
The BACB's argument is straightforward: the base
23:49
of the funding pyramid is under pressure even while
23:52
the top of the ecosystem produces headline successes.
23:57
That is the concentration economy in one chart:
24:01
ten new German unicorns in 2026 at the top, while
24:05
early-stage financing indicators in Berlin point in the
24:09
opposite direction. Without angels there are fewer
24:12
startups reaching market entry; without enough
24:16
startups entering the funnel, there are fewer scale-ups
24:19
later. We will link the BACB position paper in the
24:23
show notes. Now, the takeaways for operators and
24:27
investors. If you are a founder or startup
24:31
operator, here are three things this month's signals tell you.
24:35
First, if you are raising capital, assume investors can be
24:39
highly selective. The data shows capital concentrating into
24:43
fewer and larger rounds. If your metrics are good but not
24:46
exceptional, extending runway through revenue, disciplined
24:50
burn, or bridge financing may be strategically better than
24:53
forcing a full round at the wrong moment. Second,
24:57
audit your hiring plan against AI capability.
25:01
Twenty-seven percent of the German tech startups in Bitkom's
25:05
survey said they had forgone new hires because of AI, while
25:09
sixteen percent had added staff because of AI. If you have not
25:13
assessed which roles should be automated, augmented,
25:16
redesigned, or expanded, you are missing a material operating shift.
25:20
This is not simply about cutting people. It is
25:23
about allocating human capital to work where
25:27
judgment, relationships, creativity, accountability,
25:31
and domain expertise create the most value.
25:35
Third, if you operate in fintech infrastructure,
25:39
robotics, physical AI, or defense, capital and strategic
25:43
activity are currently unusually strong. That
25:47
does not make every company in those sectors
25:51
fundable, but it does mean the market is rewarding credible
25:54
traction and strategic relevance. If you are
25:58
an investor, the signal is equally clear: European
26:02
venture outcomes are becoming more concentrated, so
26:06
portfolio construction, follow-on reserves,
26:09
and conviction around category leaders matter more.
26:13
That argues for fewer, higher-conviction bets in
26:17
some strategies and for enough follow-on capacity
26:21
to support the winners. One more investor note:
26:25
insolvencies and restructurings in cleantech, foodtech, and
26:29
other capital-intensive sectors can create
26:33
distressed-asset opportunities. Companies such as Cultimate
26:37
Foods and Sono Solar developed real technology and
26:40
intellectual property, even though their financing
26:44
structures or runway did not hold. In some cases, the
26:48
failure of the company does not imply that the
26:52
underlying technology is worthless. For patient capital,
26:55
strategic buyers, or specialist investors,
26:59
assets coming out of restructurings may be
27:02
priced very differently from their long-term technical
27:06
value. That is not a recommendation; it is an observation
27:10
about where market inefficiencies can emerge.
27:15
What to watch next. Four things as we move into the fall.
27:18
Number one: the Q3 and second-half deal-count data.
27:22
The verified baseline is that Germany recorded 354
27:26
financing rounds in the first half of 2026, down eleven percent
27:30
year over year. The question is whether that decline
27:34
continues into the second half or stabilizes. Our
27:38
explicit prediction remains that quarterly deal count falls below
27:42
150 by year-end. Number two: HTGF V. Watch the
27:45
fundraising, the mix of public anchor investors and private
27:49
investors, and how HTGF's multi-stage strategy
27:53
develops. That matters because HTGF remains one of the central
27:57
institutional players in German early-stage technology
28:00
investing. Number three: whether Scalable Capital's
28:04
Altersvorsorgedepot attracts fast-moving competitors.
28:07
My prediction says at least three competing
28:11
products are announced or launched by Q1 2027. Watch
28:14
Trade Republic, direct banks, neobanks, and
28:18
established asset managers. The first products can shape
28:22
customer expectations around fees, asset
28:26
allocation, and user experience. Number four: the Isar
28:30
Aerospace launch window. We have been tracking
28:34
this since June. An orbital launch before year-end
28:37
would be a landmark moment for European deep tech
28:41
and would strengthen the case that Germany can
28:45
produce globally competitive hardware-first
28:49
companies, not only software platforms. Now, the closing.
28:53
That is the August 2026 startup news for Germany, Austria, and Switzerland.
28:56
The concentration economy is the thesis, and the evidence is visible across
29:00
capital allocation, hiring, sector leadership, and early-stage stress. Ten
29:03
new unicorns sit at the top. At the same time, twenty-seven percent of German tech
29:07
startups in Bitkom's survey said they had forgone new hires because of AI, while
29:11
sixteen percent added staff because of it. This is not simply a crisis story. It
29:14
is a transition in how capital, labor, and strategic value are distributed. The
29:18
companies and investors that understand the new rules will be better
29:21
positioned to benefit. We have seen concentration phases in other mature
29:25
technology ecosystems as well. The important question is what happens after the concentration.
29:29
What defines whether an ecosystem emerges stronger or
29:33
weaker is whether capital and experience recycle through
29:36
exits, secondaries, employee wealth creation, and founders
29:40
starting again. Germany has not yet proven that it can do that
29:44
at the scale of Silicon Valley. The next twelve to
29:48
eighteen months will tell us whether this concentration is a
29:52
temporary phase or a more permanent restructuring. Let me
29:55
restate the three predictions. One: the decline in deal
29:59
count accelerates through Q4, with quarterly German startup
30:03
financing rounds falling below 150 by year-end. Two: at
30:07
least two of Germany's ten 2026 unicorns show a down
30:11
round or a meaningful distress signal within eighteen
30:14
months. Three: Scalable Capital's Altersvorsorgedepot triggers
30:18
at least three competing pension products by Q1 2027.
30:22
Check us against those numbers in the months ahead. If you
30:26
missed the July episode, go back and listen. We covered the Helsing
30:30
valuation trajectory, the defense-tech rotation, and the
30:33
early pension-reform signal that developed further this
30:36
month. This is the second of two end-of-summer
30:40
episodes. Our next monthly news will be the September edition,
30:44
covering the signals that emerge in September. This has been
30:48
Startuprad.io. I'm Joe Menninger, reporting from Frankfurt am
30:51
Main, Germany. Everyone celebrated the ten unicorns.
30:57
Nobody counted the deals that did not happen.
31:01
That is the gap between the headline economy and the
31:05
real one.