WEBVTT

1
00:00:00.096 --> 00:00:02.842
This is episode 773 of Startuprad.io:

2
00:00:02.842 --> 00:00:05.892
Startup News August 2026, recorded for

3
00:00:05.892 --> 00:00:09.553
publication on Labor Day, Monday, September 7, 2026.
Ten,

4
00:00:09.746 --> 00:00:13.503
minus eleven, twenty-seven — those are three numbers
that

5
00:00:14.145 --> 00:00:17.774
frame the story of the German startup ecosystem this

6
00:00:17.854 --> 00:00:21.547
month: ten new unicorns in 2026 so far, eleven percent

7
00:00:21.547 --> 00:00:25.128
fewer financing rounds in the first half of the year,

8
00:00:25.562 --> 00:00:28.869
and twenty-seven percent of German tech startups

9
00:00:29.174 --> 00:00:32.916
saying AI caused them to forgo new hires. This is the

10
00:00:33.590 --> 00:00:37.074
concentration economy, and it changes everything.

11
00:00:37.122 --> 00:00:40.302
This is Startuprad.io. I'm Joe Menninger,

12
00:00:52.117 --> 00:00:55.674
joining you from Frankfurt am Main, Germany. This is
the

13
00:00:55.803 --> 00:00:59.472
startup news for Germany, Austria, and Switzerland for

14
00:00:59.568 --> 00:01:02.789
August 2026. As a tradition established with

15
00:01:03.414 --> 00:01:07.036
Chris, we publish our summer wrap-ups around Labor
Day.

16
00:01:07.340 --> 00:01:11.138
If you celebrate it, happy Labor Day. This is

17
00:01:11.522 --> 00:01:15.176
part two of our summer wrap-up, and this is

18
00:01:15.288 --> 00:01:18.830
episode 773 of the Startuprad.io franchise. This

19
00:01:18.830 --> 00:01:22.388
month's thesis is what I'm calling the

20
00:01:22.725 --> 00:01:26.507
concentration economy. For years, we have tracked this

21
00:01:26.571 --> 00:01:30.354
ecosystem's growth in aggregate: total funding,

22
00:01:30.755 --> 00:01:34.024
unicorn counts, deal flow. And the

23
00:01:34.665 --> 00:01:38.336
headline numbers still look strong. Germany has

24
00:01:38.913 --> 00:01:42.583
produced ten new unicorns in 2026 so far. Amazing.

25
00:01:43.032 --> 00:01:46.222
But look beneath the surface and you find a very
different story.

26
00:01:47.089 --> 00:01:50.781
EY's data for the first half of 2026 shows about 5.3
billion euros invested in

27
00:01:50.781 --> 00:01:53.654
German startups, up fourteen percent from the same
period last

28
00:01:54.088 --> 00:01:57.908
year, while the number of financing rounds fell eleven
percent to 354. Sixty-seven

29
00:01:58.422 --> 00:02:01.841
percent of all capital went into rounds above 50
million euros. And an

30
00:02:02.194 --> 00:02:05.501
August Bitkom survey found that twenty-seven percent
of German tech

31
00:02:05.501 --> 00:02:09.257
startups had forgone new hires because of AI; seven
percent had cut jobs because of AI,

32
00:02:09.257 --> 00:02:12.467
while sixteen percent had added staff because of it.
The boom is

33
00:02:12.644 --> 00:02:16.095
real, but its benefits are concentrating. That is the
concentration economy.

34
00:02:16.095 --> 00:02:19.847
Before we get into the detail, let's check our

35
00:02:19.847 --> 00:02:23.294
predictions from the July episode. Prediction

36
00:02:23.678 --> 00:02:27.510
one: Helsing reaches a 25 billion US dollar valuation

37
00:02:27.846 --> 00:02:31.261
within twelve months. Not yet. The defense AI

38
00:02:31.453 --> 00:02:35.092
company continues to expand across NATO markets,

39
00:02:35.092 --> 00:02:38.603
but no new funding round was announced in

40
00:02:38.603 --> 00:02:42.033
August. We are still watching. Prediction two: two

41
00:02:42.610 --> 00:02:46.265
or more additional German defense unicorns

42
00:02:46.281 --> 00:02:49.951
emerge by the end of 2026. This month Cambridge

43
00:02:50.144 --> 00:02:53.717
Aerospace reached a 3.4 billion US dollar valuation,

44
00:02:54.214 --> 00:02:57.627
which is impressive, but it is a UK company,

45
00:02:57.995 --> 00:03:01.777
not a German one. So that prediction remains open.

46
00:03:01.825 --> 00:03:05.559
We need additional German-headquartered defense
startups

47
00:03:06.007 --> 00:03:09.388
crossing the billion-dollar line. Prediction three:
the

48
00:03:09.388 --> 00:03:13.122
pension-reform pipeline reaches 5 billion by mid-2027.
This month

49
00:03:13.122 --> 00:03:16.025
gave us a useful product signal, but not yet a

50
00:03:16.332 --> 00:03:19.743
capital-flow confirmation. Scalable Capital is already

51
00:03:19.743 --> 00:03:23.590
allowing customers to add its planned
Altersvorsorgedepot and

52
00:03:23.590 --> 00:03:27.212
promoting the state-supported retirement product ahead
of the new

53
00:03:27.228 --> 00:03:30.786
framework taking effect in 2027. That is evidence of

54
00:03:30.786 --> 00:03:34.269
competitive preparation, but we still need to see
actual assets

55
00:03:34.560 --> 00:03:37.197
flow before calling the prediction fulfilled.

56
00:03:39.425 --> 00:03:43.228
We are also still tracking the June prediction that
Isar

57
00:03:43.228 --> 00:03:46.963
Aerospace reaches orbit by December 31. No decisive
update in August.

58
00:03:46.963 --> 00:03:49.940
The launch target remains

59
00:03:50.652 --> 00:03:54.324
later this year. Let's talk

60
00:03:56.163 --> 00:03:59.873
startups. The macro view — maybe you call

61
00:04:00.050 --> 00:04:03.530
it the unicorn factory. Germany

62
00:04:03.836 --> 00:04:07.057
has produced ten new unicorns

63
00:04:07.798 --> 00:04:11.057
in 2026 by late August. Let

64
00:04:11.719 --> 00:04:15.350
that number sit for a moment. On the

65
00:04:15.543 --> 00:04:18.837
surface, this is an

66
00:04:18.853 --> 00:04:22.597
extraordinary success story. But when you

67
00:04:22.710 --> 00:04:26.421
look at who is crossing the

68
00:04:26.630 --> 00:04:30.069
billion-dollar line, a pattern emerges.

69
00:04:30.744 --> 00:04:34.247
These are largely companies that have achieved

70
00:04:34.584 --> 00:04:38.294
meaningful revenue, real enterprise traction, or
strategic

71
00:04:38.374 --> 00:04:42.222
relevance in areas such as defense, industrial
technology,

72
00:04:42.222 --> 00:04:45.622
and deep tech. This is not the 2021 vintage, when

73
00:04:46.071 --> 00:04:49.903
narrative and growth metrics alone could carry more of
the

74
00:04:50.384 --> 00:04:53.976
valuation story. Take Moss as an example. On August 5,

75
00:04:54.073 --> 00:04:57.616
the corporate-spend platform announced a 35 million

76
00:04:57.616 --> 00:05:01.144
euro Series C at a 1 billion euro valuation. Its

77
00:05:01.144 --> 00:05:04.239
product has expanded from corporate cards into

78
00:05:05.807 --> 00:05:08.849
invoice management, travel and expense workflows,

79
00:05:09.171 --> 00:05:12.536
and finance automation. Or look at Camunda, whose

80
00:05:14.003 --> 00:05:17.779
annual recurring revenue is nearing 200 million US
dollars

81
00:05:18.037 --> 00:05:21.555
and is expected to exceed that level before year-end.

82
00:05:21.781 --> 00:05:25.304
This is real software revenue at scale, not just a

83
00:05:25.690 --> 00:05:29.291
projection. The unicorn factory is producing a very

84
00:05:29.709 --> 00:05:33.476
specific kind of company: businesses with stronger
operating

85
00:05:33.637 --> 00:05:36.891
proof than the market demanded a few years ago.

86
00:05:37.019 --> 00:05:40.803
There's an important structural point here. In the

87
00:05:41.332 --> 00:05:44.570
2020 and 2021 boom, a billion-dollar valuation

88
00:05:44.858 --> 00:05:48.706
could be reached with a very different mix of growth,

89
00:05:48.786 --> 00:05:52.217
narrative, and capital-market expectations. In 2026,

90
00:05:52.329 --> 00:05:55.391
investors are putting more weight on revenue

91
00:05:55.936 --> 00:05:58.757
quality, unit economics, strategic

92
00:06:00.588 --> 00:06:04.203
relevance, and a credible path to durable scale. That
filter

93
00:06:04.203 --> 00:06:07.544
may be healthy in the long run, but it also

94
00:06:08.202 --> 00:06:12.009
makes the unicorn count less representative of broad

95
00:06:12.315 --> 00:06:16.138
ecosystem strength. Ten unicorns can sound like a
rising

96
00:06:16.250 --> 00:06:19.864
tide, while much of the market below them faces a
tougher

97
00:06:19.864 --> 00:06:22.659
funding environment. The funding

98
00:06:23.559 --> 00:06:27.397
paradox is where the concentration thesis really takes
shape.

99
00:06:27.397 --> 00:06:30.653
And the clearest structural data point comes

100
00:06:30.942 --> 00:06:34.358
from the first half of 2026, not from August alone.

101
00:06:34.454 --> 00:06:37.614
EY recorded about 5.3 billion euros of venture

102
00:06:38.143 --> 00:06:41.415
capital invested in German startups in the first half
of 2026,

103
00:06:41.688 --> 00:06:45.344
fourteen percent more than a year earlier. But the
number of

104
00:06:46.002 --> 00:06:49.755
financing rounds fell eleven percent to 354. Seventeen
rounds

105
00:06:49.803 --> 00:06:53.556
exceeded 50 million euros, and those large
transactions captured

106
00:06:53.556 --> 00:06:56.731
67 percent of all invested capital. Let me repeat

107
00:06:57.804 --> 00:07:01.249
that: roughly two-thirds of the capital went into
rounds above

108
00:07:01.713 --> 00:07:05.382
50 million euros. This is the funding paradox of the

109
00:07:05.814 --> 00:07:09.483
concentration economy: more capital than in the first
half of 2025, but

110
00:07:09.499 --> 00:07:13.328
spread across fewer financings and concentrated
heavily in

111
00:07:13.328 --> 00:07:16.788
large rounds. That does not mean every early-stage
segment

112
00:07:17.269 --> 00:07:21.114
follows the same pattern, but it clearly raises the
importance of

113
00:07:21.114 --> 00:07:24.687
differentiation, traction, and capital efficiency for
founders.

114
00:07:24.943 --> 00:07:28.536
And to put the shift into a longer perspective,

115
00:07:28.632 --> 00:07:31.862
compare this with the peak. In the first half

116
00:07:32.568 --> 00:07:35.717
of 2021, EY counted 588 German startup

117
00:07:36.424 --> 00:07:40.136
financing rounds and about 7.6 billion euros of

118
00:07:40.136 --> 00:07:43.879
investment. In the first half of 2026, it counted 354

119
00:07:44.313 --> 00:07:48.040
rounds and about 5.3 billion euros. That is roughly

120
00:07:48.040 --> 00:07:51.784
forty percent fewer rounds than at the 2021 peak.

121
00:07:51.784 --> 00:07:55.318
The ecosystem still attracts substantial capital,

122
00:07:55.318 --> 00:07:58.369
but participation has narrowed. That

123
00:07:59.012 --> 00:08:02.834
matters beyond venture funds themselves. Law firms,

124
00:08:02.850 --> 00:08:06.191
recruiters, coworking providers, accelerators, and

125
00:08:06.239 --> 00:08:09.949
other service businesses often depend much more on

126
00:08:09.997 --> 00:08:13.562
the number of active companies and transactions

127
00:08:13.627 --> 00:08:17.015
than on the size of a handful of mega-rounds.

128
00:08:17.080 --> 00:08:19.954
The AI labor shock. A Bitkom

129
00:08:21.671 --> 00:08:25.023
survey released in August found that

130
00:08:25.023 --> 00:08:28.647
twenty-seven percent of German tech

131
00:08:29.113 --> 00:08:32.817
startups had forgone new hires during the

132
00:08:32.962 --> 00:08:36.602
previous twelve months because of artificial

133
00:08:36.811 --> 00:08:40.660
intelligence. Seven percent said they had

134
00:08:40.740 --> 00:08:44.445
cut jobs because of AI. But sixteen

135
00:08:44.974 --> 00:08:48.823
percent reported the opposite: they had hired

136
00:08:49.064 --> 00:08:52.863
additional staff because of AI. Half said AI

137
00:08:53.088 --> 00:08:56.871
had no effect on their personnel needs.

138
00:08:56.999 --> 00:09:00.509
So the signal is not simply 'AI kills

139
00:09:00.606 --> 00:09:04.308
startup jobs.' It is a bifurcation in what

140
00:09:04.629 --> 00:09:08.396
kinds of work startups need and how much

141
00:09:08.396 --> 00:09:11.842
output they expect from each employee.

142
00:09:11.923 --> 00:09:15.369
On one side, AI can reduce demand for some

143
00:09:15.369 --> 00:09:18.627
planned administrative, analytical, content,

144
00:09:19.205 --> 00:09:22.803
support, or software-development work. On the

145
00:09:23.221 --> 00:09:26.738
other, companies building, deploying,

146
00:09:26.770 --> 00:09:30.063
integrating, and governing AI systems need

147
00:09:30.545 --> 00:09:34.208
specialized people. Bitkom also reports that the
average

148
00:09:34.240 --> 00:09:37.806
German startup in its survey employed twelve

149
00:09:37.934 --> 00:09:41.371
people, down from thirteen a year earlier.

150
00:09:41.532 --> 00:09:45.160
That connects directly to the concentration thesis.

151
00:09:45.160 --> 00:09:48.995
If a seed-stage founder can build more with a

152
00:09:49.011 --> 00:09:52.767
smaller team, the company may need less capital and

153
00:09:52.767 --> 00:09:56.329
fewer hires to reach the same milestone. That

154
00:09:56.329 --> 00:09:59.972
can be genuine efficiency, while still reducing

155
00:10:00.069 --> 00:10:03.086
the number of funding rounds, offices,

156
00:10:04.016 --> 00:10:07.708
service contracts, and entry-level roles

157
00:10:07.772 --> 00:10:11.447
generated per startup. The ecosystem can grow in value

158
00:10:11.559 --> 00:10:15.250
without growing proportionally in participation.

159
00:10:15.250 --> 00:10:18.524
The infrastructure layer. Three

160
00:10:18.524 --> 00:10:21.830
developments deserve attention. First, High-Tech

161
00:10:22.055 --> 00:10:25.521
Gründerfonds announced fundraising for HTGF V, its

162
00:10:26.067 --> 00:10:29.614
fifth seed fund, on August 26. The important

163
00:10:29.614 --> 00:10:33.273
signal is that HTGF is positioning itself as a

164
00:10:33.385 --> 00:10:36.961
multi-stage platform that can connect seed investing

165
00:10:37.395 --> 00:10:41.139
with later-stage support, while continuing to

166
00:10:41.396 --> 00:10:44.900
combine public anchor investors with private

167
00:10:45.414 --> 00:10:48.692
capital. Second, Scalable Capital is already

168
00:10:49.271 --> 00:10:53.080
marketing and allowing customers to add its planned

169
00:10:53.611 --> 00:10:55.761
Altersvorsorgedepot ahead of

170
00:10:57.329 --> 00:11:00.727
the new state-supported retirement framework

171
00:11:01.702 --> 00:11:05.170
taking effect in 2027. That matters because a

172
00:11:05.780 --> 00:11:09.569
successful capital-market pension wrapper could

173
00:11:09.569 --> 00:11:12.909
eventually create a much broader distribution

174
00:11:13.342 --> 00:11:16.570
channel for long-term investment products.

175
00:11:16.634 --> 00:11:19.765
If Scalable or its competitors eventually

176
00:11:19.765 --> 00:11:23.265
build venture-adjacent exposure into these pension

177
00:11:23.441 --> 00:11:27.054
products, the potential distribution channel reaches

178
00:11:27.054 --> 00:11:30.459
millions of German savers. That would be a structural

179
00:11:30.459 --> 00:11:34.217
change in how long-term capital reaches the innovation

180
00:11:34.217 --> 00:11:37.815
economy, although that step has not happened yet.

181
00:11:37.911 --> 00:11:40.609
Third, Bavaria announced roughly 30 million

182
00:11:41.059 --> 00:11:43.982
euros of support for WERK1 in Munich for the

183
00:11:44.592 --> 00:11:48.431
period from 2027 to 2032. That is not a new 30 million

184
00:11:48.431 --> 00:11:52.173
euro venture fund; it is infrastructure funding for
one

185
00:11:52.286 --> 00:11:55.255
of Bavaria's central startup hubs. The

186
00:11:55.640 --> 00:11:59.153
broader signal is still relevant: regions are
competing

187
00:11:59.185 --> 00:12:03.003
for startup activity not only with branding, but with

188
00:12:03.195 --> 00:12:06.692
long-term ecosystem infrastructure. Top signal

189
00:12:07.206 --> 00:12:10.205
segments. Segment one: Moss and the fintech pulse.

190
00:12:10.815 --> 00:12:14.649
Moss crossed the 1 billion euro valuation

191
00:12:14.681 --> 00:12:17.825
mark on August 5, and it is worth

192
00:12:18.290 --> 00:12:21.868
understanding why this particular unicorn

193
00:12:21.868 --> 00:12:25.318
matters for the broader fintech thesis. Moss

194
00:12:25.318 --> 00:12:29.073
started with corporate cards and has

195
00:12:29.587 --> 00:12:33.246
expanded into a broader finance platform,

196
00:12:33.438 --> 00:12:36.985
including invoice management, travel and

197
00:12:36.985 --> 00:12:40.612
expense workflows, and AI-supported finance

198
00:12:41.061 --> 00:12:44.688
operations. The strategic lesson is

199
00:12:45.234 --> 00:12:49.021
cross-sell: win a finance team with one product,

200
00:12:49.021 --> 00:12:51.781
build trust inside the workflow,

201
00:12:52.390 --> 00:12:55.788
then add adjacent products on top of the

202
00:12:56.381 --> 00:12:59.956
existing relationship. Meanwhile,

203
00:13:00.389 --> 00:13:03.963
Berlin-based Pliant has surpassed 100 million US

204
00:13:04.236 --> 00:13:07.682
dollars in annual recurring revenue and

205
00:13:07.682 --> 00:13:11.353
has expanded into the United States. That

206
00:13:12.075 --> 00:13:15.922
creates a serious competitive dynamic in

207
00:13:16.483 --> 00:13:20.058
European B2B payments and spend management.

208
00:13:20.346 --> 00:13:24.081
The fintech sector is moving beyond an era

209
00:13:24.370 --> 00:13:28.137
of many small players jockeying for position.

210
00:13:28.746 --> 00:13:32.449
The stronger companies are increasingly

211
00:13:32.449 --> 00:13:35.912
competing through platform breadth,

212
00:13:37.162 --> 00:13:40.897
infrastructure, cross-sell, and international

213
00:13:41.891 --> 00:13:44.953
scale. Segment two: the robotics surge.

214
00:13:45.433 --> 00:13:48.848
If there is one DACH sector that moved

215
00:13:48.960 --> 00:13:52.487
from interesting to unmissable this

216
00:13:52.487 --> 00:13:56.304
month, it is robotics and physical AI. NEURA

217
00:13:56.497 --> 00:13:59.628
Robotics expanded its platform through

218
00:14:00.254 --> 00:14:04.027
two concrete acquisitions: ACTIVE Shuttle from

219
00:14:04.027 --> 00:14:07.704
Bosch Rexroth and 100 percent of ADLATUS

220
00:14:08.202 --> 00:14:11.767
Robotics. The strategy is clear — bring

221
00:14:11.831 --> 00:14:15.203
different robotic applications onto a

222
00:14:15.203 --> 00:14:18.703
shared physical-AI infrastructure rather than

223
00:14:18.815 --> 00:14:22.475
scaling only one machine or one use case.

224
00:14:22.652 --> 00:14:26.327
Gravis Robotics added another major

225
00:14:26.343 --> 00:14:30.179
signal. The Swiss ETH spin-off announced a 200

226
00:14:30.211 --> 00:14:33.966
million US dollar Series A on August 17 to

227
00:14:33.966 --> 00:14:37.112
scale autonomous heavy machinery, and

228
00:14:37.593 --> 00:14:41.284
ETH reported that the round put the company

229
00:14:41.284 --> 00:14:44.943
at unicorn status. So this is not only a

230
00:14:44.992 --> 00:14:48.475
German robotics story; it is a DACH

231
00:14:48.828 --> 00:14:52.664
physical-AI story. Capital is backing systems

232
00:14:52.664 --> 00:14:55.971
that combine software, autonomy, sensors,

233
00:14:56.357 --> 00:14:59.631
and heavy industrial hardware.

234
00:14:59.648 --> 00:15:03.099
Segment three: AI infrastructure

235
00:15:03.099 --> 00:15:06.727
consolidation. By late August, reports said NVIDIA was

236
00:15:07.192 --> 00:15:10.820
in talks to acquire Hugging Face. That

237
00:15:10.820 --> 00:15:14.544
transaction was formally announced on

238
00:15:14.656 --> 00:15:18.005
September 3, after the August cut-off, at

239
00:15:18.582 --> 00:15:21.530
roughly 13 billion US dollars. So for

240
00:15:21.562 --> 00:15:24.976
August, the correct signal is

241
00:15:25.008 --> 00:15:28.613
acquisition talks, not a completed funding round.

242
00:15:28.757 --> 00:15:32.202
There were other major infrastructure deals

243
00:15:32.523 --> 00:15:35.904
around the same time. Stripe was reported

244
00:15:35.904 --> 00:15:39.509
to have finalized a deal to acquire

245
00:15:39.509 --> 00:15:42.009
OpenRouter for more than 7 billion US

246
00:15:42.009 --> 00:15:45.438
dollars, although Stripe itself had not

247
00:15:45.839 --> 00:15:49.572
publicly confirmed the transaction. Lovable raised

248
00:15:49.685 --> 00:15:53.178
400 million US dollars in Series C funding at

249
00:15:53.210 --> 00:15:56.976
a 13.3 billion dollar valuation — not a 400

250
00:15:56.976 --> 00:16:00.213
million dollar valuation. And Italy's

251
00:16:00.645 --> 00:16:04.155
Bending Spoons entered a definitive agreement to

252
00:16:04.587 --> 00:16:07.842
acquire Airtable in a transaction valuing

253
00:16:08.677 --> 00:16:12.049
Airtable at 1.285 billion dollars in

254
00:16:12.049 --> 00:16:15.228
enterprise value, with an implied equity

255
00:16:15.404 --> 00:16:18.953
value of roughly 2.25 billion dollars. What do

256
00:16:18.953 --> 00:16:22.550
these international deals mean for DACH? The

257
00:16:22.710 --> 00:16:26.516
infrastructure and application layers that

258
00:16:26.564 --> 00:16:29.759
European startups depend on are consolidating

259
00:16:30.257 --> 00:16:33.612
quickly, and control is shifting toward a

260
00:16:33.869 --> 00:16:37.642
smaller set of very well-capitalized platforms.

261
00:16:37.642 --> 00:16:41.351
That creates both opportunity and risk. Better

262
00:16:41.351 --> 00:16:44.690
infrastructure lowers the cost of building AI

263
00:16:44.947 --> 00:16:48.286
applications. But dependence on a small number of
model,

264
00:16:48.960 --> 00:16:52.492
cloud, chip, routing, and development platforms

265
00:16:52.508 --> 00:16:55.879
creates exposure to pricing changes, access rules,

266
00:16:56.602 --> 00:17:00.407
and competitive shifts. German and European founders

267
00:17:00.455 --> 00:17:04.036
building on top of those layers need to be explicit

268
00:17:04.212 --> 00:17:08.050
about where their defensible moat actually sits. There

269
00:17:08.242 --> 00:17:12.080
is also a European sovereignty angle, but it needs

270
00:17:12.080 --> 00:17:15.869
precision. Not every major deal is
American-on-American —

271
00:17:15.869 --> 00:17:19.129
Bending Spoons is Italian, and Lovable is

272
00:17:19.225 --> 00:17:22.774
Swedish. Still, many of the most critical model,

273
00:17:23.015 --> 00:17:26.774
compute, cloud, and AI-routing layers remain
concentrated in

274
00:17:26.790 --> 00:17:30.631
US-controlled platforms. If European policymakers are

275
00:17:30.631 --> 00:17:33.992
serious about digital sovereignty, infrastructure

276
00:17:34.170 --> 00:17:37.032
is one of the layers that matters most.

277
00:17:37.258 --> 00:17:40.702
The restructuring and insolvency wave tells

278
00:17:41.123 --> 00:17:44.824
the other side of the concentration economy.

279
00:17:45.405 --> 00:17:48.507
Recent examples include Cultimate Foods, the

280
00:17:49.217 --> 00:17:53.004
Berlin cultivated-fat startup that entered

281
00:17:53.453 --> 00:17:56.891
preliminary insolvency proceedings at the end of

282
00:17:56.891 --> 00:18:00.633
April, and Sono Motors, which opened insolvency

283
00:18:00.730 --> 00:18:04.327
proceedings in August and put the Sono Solar

284
00:18:04.761 --> 00:18:07.942
business and intellectual property up for

285
00:18:08.038 --> 00:18:11.395
sale. Glow25, the collagen and wellness

286
00:18:11.813 --> 00:18:15.186
brand, also announced significant layoffs, with

287
00:18:15.266 --> 00:18:18.430
management explicitly pointing to

288
00:18:18.639 --> 00:18:22.028
greater use of AI and automation. And Pleo, the

289
00:18:22.430 --> 00:18:26.012
Danish expense-management fintech with a

290
00:18:26.285 --> 00:18:29.658
significant European footprint, faced public

291
00:18:29.658 --> 00:18:33.432
scrutiny after former employees described a much

292
00:18:33.432 --> 00:18:36.934
tougher culture, executive turnover, internal

293
00:18:36.966 --> 00:18:40.789
friction, and concerns about the company's direction.

294
00:18:40.821 --> 00:18:44.001
When former employees are telling

295
00:18:44.178 --> 00:18:47.869
journalists that a company has 'lost its soul,'

296
00:18:47.869 --> 00:18:51.301
you do not just have an HR story; you

297
00:18:51.317 --> 00:18:54.540
have a potential brand, retention, and

298
00:18:55.246 --> 00:18:58.838
execution story. The broader pattern is

299
00:18:59.015 --> 00:19:02.206
consistent with the concentration thesis:

300
00:19:02.719 --> 00:19:06.295
capital is concentrating at the top,

301
00:19:06.343 --> 00:19:10.016
while weaker or more capital-intensive

302
00:19:10.016 --> 00:19:13.769
businesses face harder financing decisions,

303
00:19:13.769 --> 00:19:17.234
restructuring, or failure. The middle of the

304
00:19:17.234 --> 00:19:20.891
distribution can get squeezed. The defense

305
00:19:20.907 --> 00:19:24.372
rotation also continued in August. Cambridge

306
00:19:24.918 --> 00:19:27.918
Aerospace raised 300 million US

307
00:19:28.062 --> 00:19:31.591
dollars in a Series C at a 3.4 billion dollar

308
00:19:31.591 --> 00:19:35.329
valuation. But precision matters: Cambridge

309
00:19:35.345 --> 00:19:39.147
Aerospace is a UK company, not a German one.

310
00:19:39.292 --> 00:19:42.464
So while Cambridge Aerospace

311
00:19:43.057 --> 00:19:46.726
confirms the broader European defense-tech thesis,

312
00:19:46.726 --> 00:19:50.267
it does not advance our prediction

313
00:19:50.267 --> 00:19:53.937
about additional German defense unicorns.

314
00:19:54.337 --> 00:19:58.103
Germany already has major defense-tech unicorns

315
00:19:58.103 --> 00:20:01.179
including Helsing and STARK. The

316
00:20:01.179 --> 00:20:04.576
question is whether the pipeline beyond

317
00:20:04.832 --> 00:20:08.506
them — in drones, autonomous systems,

318
00:20:08.522 --> 00:20:11.961
satellite communications, cyber, sensing,

319
00:20:12.411 --> 00:20:15.336
and logistics — produces additional

320
00:20:15.352 --> 00:20:18.962
billion-dollar companies before

321
00:20:19.140 --> 00:20:22.932
year-end. Our prediction says two more after the

322
00:20:22.932 --> 00:20:26.429
July baseline. The clock is ticking.

323
00:20:26.429 --> 00:20:29.717
Lightning round: six stories, about thirty

324
00:20:29.911 --> 00:20:33.553
seconds each. Number one: Julian Teicke,

325
00:20:33.729 --> 00:20:37.520
the founder of wefox, has started a new

326
00:20:37.520 --> 00:20:41.278
company called Attuned with Rebecca Godfrey and

327
00:20:41.358 --> 00:20:44.843
Thomas Hübl. Teicke is building again.

328
00:20:44.843 --> 00:20:48.569
Attuned is focused on real-time human co-regulation

329
00:20:48.746 --> 00:20:52.376
rather than another AI assistant, and its public

330
00:20:52.456 --> 00:20:55.732
beta is scheduled to open in September. The

331
00:20:55.732 --> 00:20:59.378
serial-founder signal is clear: experienced European

332
00:20:59.250 --> 00:21:02.812
founders are recycling back into the ecosystem and

333
00:21:03.775 --> 00:21:06.904
starting again. Number two: Waymo is laying

334
00:21:06.904 --> 00:21:10.290
the groundwork for a fully autonomous

335
00:21:10.579 --> 00:21:14.286
ride-hailing service in Munich. It plans manual
mapping and

336
00:21:14.286 --> 00:21:17.785
validation first, and says it aims to open

337
00:21:18.122 --> 00:21:21.748
commercial ride-hailing to the public toward the end
of

338
00:21:21.748 --> 00:21:25.488
2027. If that happens, Germany becomes a serious test

339
00:21:25.520 --> 00:21:29.116
market for large-scale autonomous mobility. Number

340
00:21:29.373 --> 00:21:32.922
three: smart glasses are facing renewed privacy

341
00:21:32.970 --> 00:21:36.599
scrutiny in Germany. Legal experts quoted by German

342
00:21:36.599 --> 00:21:39.826
public media say a ban could be conceivable

343
00:21:39.875 --> 00:21:43.712
under certain circumstances because covert recording

344
00:21:43.825 --> 00:21:47.213
creates difficult consent and privacy questions.

345
00:21:47.277 --> 00:21:50.938
This is the kind of issue where German law and

346
00:21:51.227 --> 00:21:54.951
enforcement could influence how ambient-computing

347
00:21:55.287 --> 00:21:59.069
hardware develops across Europe. Number four:

348
00:21:59.470 --> 00:22:03.188
Carlsen Verlag — not chess champion Magnus Carlsen —

349
00:22:04.437 --> 00:22:08.283
is suing OpenAI together with author Marc-Uwe

350
00:22:08.283 --> 00:22:11.761
Kling and illustrator Astrid Henn. The case

351
00:22:11.969 --> 00:22:15.543
concerns alleged infringement involving their work

352
00:22:16.039 --> 00:22:19.853
'Das NEINhorn' and goes directly to the question

353
00:22:19.853 --> 00:22:23.637
of how copyrighted text and illustrations may be

354
00:22:23.989 --> 00:22:27.083
used or reproduced by generative AI.

355
00:22:27.500 --> 00:22:30.754
Number five: Bavaria is committing roughly 30

356
00:22:31.203 --> 00:22:35.018
million euros to WERK1 for 2027 through 2032.

357
00:22:35.131 --> 00:22:38.273
Again, that is startup-infrastructure

358
00:22:38.273 --> 00:22:41.479
funding, not a new venture fund. Number six:

359
00:22:41.527 --> 00:22:45.310
BaFin's crypto supervision. In an August interview on

360
00:22:45.310 --> 00:22:48.681
MiCAR, BaFin expert Ruth Burkert said that

361
00:22:48.761 --> 00:22:52.518
a BaFin license is viewed in the market as a

362
00:22:52.598 --> 00:22:56.308
'seal of quality.' For crypto companies, that

363
00:22:56.308 --> 00:22:59.615
underscores how regulatory credibility can become

364
00:22:59.615 --> 00:23:03.196
a commercial asset across the European market.

365
00:23:03.453 --> 00:23:07.290
Regulation, in this case, can become part of market

366
00:23:07.290 --> 00:23:10.726
positioning. And one more story cuts directly to the

367
00:23:11.224 --> 00:23:14.514
thesis of this episode. The Business Angels

368
00:23:14.852 --> 00:23:18.222
Club Berlin-Brandenburg published a position paper

369
00:23:18.222 --> 00:23:21.640
warning about the deterioration of early-stage

370
00:23:22.074 --> 00:23:25.733
funding in Berlin. According to the BACB, pre-seed

371
00:23:25.910 --> 00:23:29.103
funding volume in Berlin has fallen 65 percent

372
00:23:29.842 --> 00:23:33.437
since 2022, back to roughly 2017 levels. The median

373
00:23:33.870 --> 00:23:37.449
angel ticket has fallen from 50,000 euros to 38,000

374
00:23:37.883 --> 00:23:41.670
euros. Startup insolvencies have nearly doubled since

375
00:23:41.799 --> 00:23:45.281
2022, even while startup formations have increased.

376
00:23:45.281 --> 00:23:48.780
The BACB's argument is straightforward: the base

377
00:23:49.358 --> 00:23:52.937
of the funding pyramid is under pressure even while

378
00:23:52.985 --> 00:23:56.740
the top of the ecosystem produces headline successes.

379
00:23:57.318 --> 00:24:00.753
That is the concentration economy in one chart:

380
00:24:01.298 --> 00:24:04.737
ten new German unicorns in 2026 at the top, while

381
00:24:05.172 --> 00:24:08.732
early-stage financing indicators in Berlin point in
the

382
00:24:09.119 --> 00:24:12.551
opposite direction. Without angels there are fewer

383
00:24:12.746 --> 00:24:16.048
startups reaching market entry; without enough

384
00:24:16.388 --> 00:24:19.900
startups entering the funnel, there are fewer
scale-ups

385
00:24:19.949 --> 00:24:23.362
later. We will link the BACB position paper in the

386
00:24:23.426 --> 00:24:26.708
show notes. Now, the takeaways for operators and

387
00:24:27.930 --> 00:24:31.331
investors. If you are a founder or startup

388
00:24:31.829 --> 00:24:35.618
operator, here are three things this month's signals
tell you.

389
00:24:35.618 --> 00:24:39.199
First, if you are raising capital, assume investors
can be

390
00:24:39.231 --> 00:24:42.908
highly selective. The data shows capital concentrating
into

391
00:24:43.309 --> 00:24:46.585
fewer and larger rounds. If your metrics are good but
not

392
00:24:46.617 --> 00:24:50.230
exceptional, extending runway through revenue,
disciplined

393
00:24:50.230 --> 00:24:53.778
burn, or bridge financing may be strategically better
than

394
00:24:53.842 --> 00:24:56.941
forcing a full round at the wrong moment. Second,

395
00:24:57.616 --> 00:25:00.538
audit your hiring plan against AI capability.

396
00:25:01.806 --> 00:25:05.338
Twenty-seven percent of the German tech startups in
Bitkom's

397
00:25:05.547 --> 00:25:09.208
survey said they had forgone new hires because of AI,
while

398
00:25:09.256 --> 00:25:13.029
sixteen percent had added staff because of AI. If you
have not

399
00:25:13.029 --> 00:25:16.417
assessed which roles should be automated, augmented,

400
00:25:16.561 --> 00:25:20.254
redesigned, or expanded, you are missing a material
operating shift.

401
00:25:20.334 --> 00:25:23.752
This is not simply about cutting people. It is

402
00:25:23.865 --> 00:25:27.441
about allocating human capital to work where

403
00:25:27.553 --> 00:25:31.178
judgment, relationships, creativity, accountability,

404
00:25:31.178 --> 00:25:34.545
and domain expertise create the most value.

405
00:25:35.331 --> 00:25:38.940
Third, if you operate in fintech infrastructure,

406
00:25:39.661 --> 00:25:43.430
robotics, physical AI, or defense, capital and
strategic

407
00:25:43.430 --> 00:25:46.846
activity are currently unusually strong. That

408
00:25:47.279 --> 00:25:51.015
does not make every company in those sectors

409
00:25:51.112 --> 00:25:54.897
fundable, but it does mean the market is rewarding
credible

410
00:25:54.897 --> 00:25:58.056
traction and strategic relevance. If you are

411
00:25:58.826 --> 00:26:02.482
an investor, the signal is equally clear: European

412
00:26:02.755 --> 00:26:06.524
venture outcomes are becoming more concentrated, so

413
00:26:06.524 --> 00:26:09.875
portfolio construction, follow-on reserves,

414
00:26:09.875 --> 00:26:13.548
and conviction around category leaders matter more.

415
00:26:13.580 --> 00:26:17.074
That argues for fewer, higher-conviction bets in

416
00:26:17.827 --> 00:26:21.144
some strategies and for enough follow-on capacity

417
00:26:21.657 --> 00:26:25.407
to support the winners. One more investor note:

418
00:26:25.439 --> 00:26:29.253
insolvencies and restructurings in cleantech,
foodtech, and

419
00:26:29.365 --> 00:26:33.035
other capital-intensive sectors can create

420
00:26:33.211 --> 00:26:36.945
distressed-asset opportunities. Companies such as
Cultimate

421
00:26:37.137 --> 00:26:40.504
Foods and Sono Solar developed real technology and

422
00:26:40.970 --> 00:26:44.664
intellectual property, even though their financing

423
00:26:44.680 --> 00:26:48.374
structures or runway did not hold. In some cases, the

424
00:26:48.374 --> 00:26:52.068
failure of the company does not imply that the

425
00:26:52.068 --> 00:26:55.457
underlying technology is worthless. For patient
capital,

426
00:26:55.843 --> 00:26:59.151
strategic buyers, or specialist investors,

427
00:26:59.280 --> 00:27:02.331
assets coming out of restructurings may be

428
00:27:02.508 --> 00:27:06.235
priced very differently from their long-term technical

429
00:27:06.284 --> 00:27:10.084
value. That is not a recommendation; it is an
observation

430
00:27:10.149 --> 00:27:13.335
about where market inefficiencies can emerge.

431
00:27:15.070 --> 00:27:18.912
What to watch next. Four things as we move into the
fall.

432
00:27:18.912 --> 00:27:22.318
Number one: the Q3 and second-half deal-count data.

433
00:27:22.447 --> 00:27:26.008
The verified baseline is that Germany recorded 354

434
00:27:26.056 --> 00:27:29.852
financing rounds in the first half of 2026, down
eleven percent

435
00:27:30.737 --> 00:27:34.362
year over year. The question is whether that decline

436
00:27:34.362 --> 00:27:37.779
continues into the second half or stabilizes. Our

437
00:27:38.148 --> 00:27:41.965
explicit prediction remains that quarterly deal count
falls below

438
00:27:42.157 --> 00:27:45.574
150 by year-end. Number two: HTGF V. Watch the

439
00:27:45.670 --> 00:27:49.263
fundraising, the mix of public anchor investors and
private

440
00:27:49.776 --> 00:27:53.353
investors, and how HTGF's multi-stage strategy

441
00:27:53.465 --> 00:27:57.042
develops. That matters because HTGF remains one of the
central

442
00:27:57.074 --> 00:28:00.811
institutional players in German early-stage technology

443
00:28:00.827 --> 00:28:04.240
investing. Number three: whether Scalable Capital's

444
00:28:04.288 --> 00:28:07.573
Altersvorsorgedepot attracts fast-moving competitors.

445
00:28:07.573 --> 00:28:11.018
My prediction says at least three competing

446
00:28:11.018 --> 00:28:14.832
products are announced or launched by Q1 2027. Watch

447
00:28:14.928 --> 00:28:18.629
Trade Republic, direct banks, neobanks, and

448
00:28:18.629 --> 00:28:22.395
established asset managers. The first products can
shape

449
00:28:22.395 --> 00:28:26.161
customer expectations around fees, asset

450
00:28:26.161 --> 00:28:29.802
allocation, and user experience. Number four: the Isar

451
00:28:30.171 --> 00:28:33.749
Aerospace launch window. We have been tracking

452
00:28:34.022 --> 00:28:37.792
this since June. An orbital launch before year-end

453
00:28:37.792 --> 00:28:41.418
would be a landmark moment for European deep tech

454
00:28:41.482 --> 00:28:45.027
and would strengthen the case that Germany can

455
00:28:45.396 --> 00:28:49.054
produce globally competitive hardware-first

456
00:28:49.359 --> 00:28:53.161
companies, not only software platforms. Now, the
closing.

457
00:28:53.194 --> 00:28:56.739
That is the August 2026 startup news for Germany,
Austria, and Switzerland.

458
00:28:56.787 --> 00:29:00.300
The concentration economy is the thesis, and the
evidence is visible across

459
00:29:00.397 --> 00:29:03.589
capital allocation, hiring, sector leadership, and
early-stage stress. Ten

460
00:29:03.718 --> 00:29:07.359
new unicorns sit at the top. At the same time,
twenty-seven percent of German tech

461
00:29:07.455 --> 00:29:11.242
startups in Bitkom's survey said they had forgone new
hires because of AI, while

462
00:29:11.242 --> 00:29:14.819
sixteen percent added staff because of it. This is not
simply a crisis story. It

463
00:29:14.835 --> 00:29:18.445
is a transition in how capital, labor, and strategic
value are distributed. The

464
00:29:18.461 --> 00:29:21.523
companies and investors that understand the new rules
will be better

465
00:29:21.667 --> 00:29:25.322
positioned to benefit. We have seen concentration
phases in other mature

466
00:29:25.819 --> 00:29:29.666
technology ecosystems as well. The important question
is what happens after the concentration.

467
00:29:29.666 --> 00:29:33.145
What defines whether an ecosystem emerges stronger or

468
00:29:33.289 --> 00:29:36.784
weaker is whether capital and experience recycle
through

469
00:29:36.880 --> 00:29:40.551
exits, secondaries, employee wealth creation, and
founders

470
00:29:40.679 --> 00:29:44.495
starting again. Germany has not yet proven that it can
do that

471
00:29:44.591 --> 00:29:48.280
at the scale of Silicon Valley. The next twelve to

472
00:29:48.425 --> 00:29:52.179
eighteen months will tell us whether this
concentration is a

473
00:29:52.371 --> 00:29:55.644
temporary phase or a more permanent restructuring. Let
me

474
00:29:55.644 --> 00:29:59.381
restate the three predictions. One: the decline in
deal

475
00:29:59.895 --> 00:30:03.360
count accelerates through Q4, with quarterly German
startup

476
00:30:03.504 --> 00:30:07.017
financing rounds falling below 150 by year-end. Two:
at

477
00:30:07.675 --> 00:30:11.092
least two of Germany's ten 2026 unicorns show a down

478
00:30:11.204 --> 00:30:14.961
round or a meaningful distress signal within eighteen

479
00:30:14.993 --> 00:30:18.605
months. Three: Scalable Capital's Altersvorsorgedepot
triggers

480
00:30:18.669 --> 00:30:21.944
at least three competing pension products by Q1 2027.

481
00:30:22.362 --> 00:30:25.878
Check us against those numbers in the months ahead. If
you

482
00:30:26.070 --> 00:30:29.891
missed the July episode, go back and listen. We
covered the Helsing

483
00:30:30.004 --> 00:30:33.375
valuation trajectory, the defense-tech rotation, and
the

484
00:30:33.375 --> 00:30:36.634
early pension-reform signal that developed further
this

485
00:30:36.907 --> 00:30:40.124
month. This is the second of two end-of-summer

486
00:30:40.672 --> 00:30:44.230
episodes. Our next monthly news will be the September
edition,

487
00:30:44.697 --> 00:30:48.115
covering the signals that emerge in September. This
has been

488
00:30:48.115 --> 00:30:51.749
Startuprad.io. I'm Joe Menninger, reporting from
Frankfurt am

489
00:30:51.749 --> 00:30:54.713
Main, Germany. Everyone celebrated the ten unicorns.

490
00:30:57.478 --> 00:31:01.020
Nobody counted the deals that did not happen.

491
00:31:01.295 --> 00:31:05.051
That is the gap between the headline economy and the

492
00:31:05.099 --> 00:31:05.729
real one.

