Germany is reporting records in startup formation, unicorns and venture capital. Yet the same figures show that capital is concentrating in ever fewer rounds, that the number of financings is falling, and that the biggest bets rest on state-driven demand. The boom is real, but it ends at the point where a startup would have to become a company capable of replacing an industry.

On Monday the German Startup Association presented its Startup Monitor 2026, and the message was unambiguous. More than 3,000 new startups had been added since the start of the year, more than ever before. Ten companies had become unicorns, that is, firms valued at more than one billion euros, bringing the total to a record 39. German startups had raised eight billion euros in venture capital between January and September, more than in the whole of the previous year. „The upswing comes at the right time,“ said association head Verena Pausder, „large parts of our industry are struggling to survive.“

That sentence contains the real expectation behind the report: what is being lost in industry is supposed to be replaced by founders. Whether the numbers support this can be checked, because most of them are public. The picture that emerges is less uniform than the headline.

What is being counted

Start with the new formations. The figure does not come from a survey but from commercial register data evaluated by the analytics firm startupdetector. The association reports 3,053 new startups for the first half of the year, 52 percent more than in the second half of 2025. The phrase „since the start of the year“ therefore refers to the half-year report from July, not to the first nine months. More than a third of the new formations are classed as AI-focused, and according to the association more than half of all startups now have AI at the core of their product.

This is a meaningful finding about direction and pace. What it does not say is how many of these firms generate revenue, hire staff or reach a second funding round. The press release does not disclose who defines „startup“ and how strictly. A record number of registrations is, first of all, a record number of beginnings.

Where the capital lands

The question of what follows the beginning is answered by the capital side, and there the picture sharpens. The consultancy EY counts a volume of 5.3 billion euros for the first half of 2026, 14 percent more than in the same period a year earlier. Over the same period the number of financing rounds fell by 11 percent to 354. Seventeen rounds of more than 50 million euros accounted for 67 percent of the volume, up from 55 percent a year earlier.

The regional picture shows the same thing. Baden-Württemberg leads with 1.6 billion euros, a rise of 475 percent that is almost entirely due to a single round: 1.2 billion euros for the robotics company Neura Robotics. Bavaria loses 46 percent, and Berlin records 30 percent fewer deals. Thomas Prüver of EY himself warns that investors are concentrating on a few large startups and that small ventures are losing out in the competition for capital. KfW, Germany’s state development bank, reaches a similar assessment for the second quarter: 72 percent of the capital went to scale-ups, that is, to companies already scaling, and the high volume is no proof of a broad revival of the early stage.

Adding up the three billion-euro rounds the report mentions gives a rough share. Neura Robotics, the drone maker Helsing and the drone manufacturer Quantum Systems have each raised around one billion euros or more in recent months. Together that is at least a good three of the eight billion euros, about two fifths. The periods covered by the sources are not identical, so the order of magnitude is my own rough calculation. It does fit what EY and KfW establish for the half year itself.

The wrong comparison with 2021

The yardstick is 2021, the previous peak. According to the association’s forecast, the volume in 2026 will reach the second-highest value at twelve billion euros. That sounds like a return to normal, but it conceals a difference. According to the association, the number of financing rounds is still roughly a quarter below the 2021 level. EY counted about 1,160 rounds in 2021 and 716 in 2025. The volume is thus approaching the record while the number of financed companies remains well below it.

In addition, the sources disagree about the previous year. EY puts the 2025 volume at 8.4 billion euros; the association, according to the available reports, is lower. The claim that 2026 has already surpassed the whole of 2025 therefore depends on how one counts and cannot be sustained with the EY figures.

Broader than assumed, narrower than claimed

One point speaks against the widespread reading that this is purely a defense boom. Of the ten new unicorns, only two belong to defense and robotics, namely Stark and Neura. The other eight are osapiens (supply-chain software), Dash0 (system monitoring software), CMBlu Energy (energy storage), Focused Energy and Proxima Fusion (fusion technology), FINN (car subscriptions), Moss (fintech) and Augustus (financial infrastructure). That is a mix of software, energy and deep tech, and it shows that the momentum reaches more than one segment.

At the same time, deep tech is exactly the field in which individual rounds are large and the number of participants is small. The breadth of the unicorns is not a breadth of financing. For the first half of the year, EY names hardware as the largest sector with more than 1.6 billion euros, then software and analytics with more than 1.2 billion euros and defense with 579 million euros.

The demand set by the state

That leaves the question of where the revenue comes from that is supposed to justify the valuations. In the defense segment the answer is clear: from the state. In February 2026 the budget committee of the Bundestag approved an initial procurement of 269 million euros each from Helsing and Stark, 536 million euros in total, for so-called loitering munition, that is, kamikaze drones that can circle over a target. At the same time it cut the planned framework contract, the ceiling for possible follow-on orders, from 4.3 to 2 billion euros, with a cap of one billion per company. Beyond that, new parliamentary approval is required.

At Quantum Systems the picture is smaller than the valuation suggests. According to the trade press, the framework contract for up to 747 Twister reconnaissance drones is worth 85 million euros over seven years, with 147 systems ordered immediately. I was unable to check reports of a further order of about 210 million euros in detail. Helsing, in turn, sells exclusively to governments according to Sacra, besides the Bundeswehr to British customers and to Ukraine. None of the sources gives a current revenue figure; the available number dates from 2023.

Two things follow provisionally. First, demand in this segment is politically set and politically limited, as the cut to the framework shows. Second, part of the valuations rests on expected procurement rather than on revenue already earned. How large that part is cannot be determined from the public figures. Bruegel, a Brussels think tank, also describes how in Germany 67 to 90 percent of defense procurement goes to the ten largest contractors, compared with under 40 percent in the United States. Young suppliers first have to fight for access.

Where the chain breaks

At the end of the chain stands the exit from startup status. A KfW analysis of 986 exits since 2005 reaches a clear finding: 92 percent were acquisitions and only 4 percent were IPOs. In 57 percent of acquisitions the buyer came from abroad, in almost a quarter from the United States. 46 percent of IPOs took place abroad, a third of them on Nasdaq. In the association’s survey, 62 percent of founders would prefer an IPO in the United States and only 27 percent in Germany.

Cooperation with the established industrial base is also weakening. According to one report, only 54.1 percent of startups still cooperate with corporations, the lowest value since 2020. Only one in four startups with initial funding is said to reach Series A, the first larger growth round. These two figures come from a media report and not from a primary source. They are consistent, however, with what the EY and KfW figures show.

The comparison with billion-euro rounds in the United States makes the gap visible. According to Pausder, the German rounds of Helsing, Quantum Systems and Neura would not make the top ten there. She says: „We have to counter this, because this is exactly where it is decided whether our companies make the leap to the top.“

What this means

The association’s survey of about 1,800 companies contains a finding that is easily overlooked. Only about a third rate Germany as an attractive location for founding a company, six percentage points fewer than the year before, while almost twice as many rate their own company location positively. Cited are proximity to universities, founder networks and skilled workers. Attachment to a location thus arises locally, not through the system as a whole. For regions with strong industry and universities nearby this is an opportunity but no guarantee, because that is also where the gap between initial funding and scaling is largest.

Provisionally, this much can be stated: Germany is founding more, but capital is concentrating in a few large rounds while the number of financed companies falls. The biggest bets in the defense sector hang on politically steered demand. And for a large share, the path from startup to a listed or independently growing company leads abroad. This interpretation is provisionally corroborated, not verified. It rests on half-year data and on individual cases, and it could be tested by current revenue figures of the billion-euro companies or a breakdown of capital flows by stage.

If startups are to replace established industry, as the report suggests, then what is still missing is exactly what makes industry what it is: breadth, employment and the ability to grow independently over many years. The number of beginnings is growing. Whether continuations follow will be decided less by the founders than by capital, exits and the customers who place the orders.

Ralf Keuper

Sources

• Startup-Verband: press release on the first half of 2026
• EY Startup-Barometer Deutschland, July 2026 (H1)
• EY Startup-Barometer, January 2026 (2025 annual figures)
• it-boltwise: Startup Monitor 2026, record in company formation
• it-boltwise: German startup boom, 3,000 new formations and 8 bn euros venture capital
• ms-aktuell: Ten billion-euro firms, Germany’s startups gain weight
• quasa.io: 3.4 billion euros VC, large rounds shape Germany’s Q2 (KfW data)
• Business Punk: Germany founds, America lists
• KfW Fokus Nr. 521: Start-ups, growth and exit routes
• deutsche-startups.de: Investments and exits in 2026
• DroneXL: Germany approves 536 million euro Helsing and Stark drone deal
• ESUT: Quantum Systems to deliver up to 747 Twister reconnaissance drones
• Sacra: Helsing revenue, valuation and funding
• Börsen-Zeitung: Helsing valuation rises to more than 15 bn euros
• Bruegel: Reforming European defence procurement to boost military innovation and startups
• Investing.com/Bloomberg: Quantum Systems wins drone order worth 210 million euros