February 25, 2026

๐—ช๐—ต๐—ฒ๐—ฟ๐—ฒ ๐—ถ๐˜€ ๐˜๐—ต๐—ฒ ๐—ถ๐—ป๐˜ƒ๐—ฒ๐˜€๐˜๐—ผ๐—ฟ๐˜€โ€™ ๐—ฎ๐˜๐˜๐—ฒ๐—ป๐˜๐—ถ๐—ผ๐—ป ๐—ต๐—ฒ๐—ฎ๐—ฑ๐—ถ๐—ป๐—ด ๐—ถ๐—ป ๐Ÿฎ๐Ÿฌ๐Ÿฎ๐Ÿฒ?

There are lots of reasons for investors to be distracted from looking at your deal:

๐Ÿ”น๐—œ๐—ป๐˜ƒ๐—ฒ๐˜€๐˜๐—ผ๐—ฟ๐˜€ ๐—ฎ๐—ฟ๐—ฒ ๐—ผ๐—ฏ๐˜€๐—ฒ๐˜€๐˜€๐—ถ๐—ป๐—ด ๐—ฎ๐—ฏ๐—ผ๐˜‚๐˜ ๐—ฒ๐˜…๐—ถ๐˜๐˜€ โ€“ Most are suffering from showing to little cash back to their LPs. Yes, the M&A market is experiencing a rebound, but this is driven by very large deals. Smaller transaction are still subdued. Yes, IPOs are making a comeback, but again it is driven by the very large companies.

๐Ÿ”น๐—œ๐—ป๐˜ƒ๐—ฒ๐˜€๐˜๐—ผ๐—ฟ๐˜€ ๐—ฎ๐—ฟ๐—ฒ ๐—ณ๐—ผ๐—ฐ๐˜‚๐˜€๐—ถ๐—ป๐—ด ๐—ผ๐—ป ๐—น๐—ฎ๐—ฟ๐—ด๐—ฒ ๐—ผ๐˜‚๐˜๐—ฐ๐—ผ๐—บ๐—ฒ๐˜€ โ€“ Product plays are great, but what investors really want is companies which have a shot at becoming unicorns (>$1bn valuation) or, even better, decacorns (>$10bn). In 2025 alone, more than 100 US companies became unicorns - approximately half the number of European unicorns today. And investors able to support these very large bets. In January came the news that Andreessen Horowitz had raised $15bn in new funding, representing close to 1/5th of all US venture funding in 2025.

From a sector standpoint, AI and defence are obviously all the rage:
โ€
๐Ÿ”น๐—”๐—œ is the investment theme that keeps giving (approximately half of the investorsโ€™ dollars went into AI in 2025):

โšซ๐—”๐—œ ๐—ถ๐˜๐˜€๐—ฒ๐—น๐—ณ: foundation models - with a big battle between โ€œclosed modelsโ€ (US) and open source models (a China speciality โ€“ with Mistral in Europe); application layers โ€“ with a key question mark as to the defensibility of the position as foundation models incorporates new applications; agentic AI โ€“ the new buzzword, but essentially it is giving the keys to AI to not only create nice outputs, but also actions (with or without a human in the loop).

โšซ๐—œ๐—ป๐—ณ๐—ฟ๐—ฎ๐˜€๐˜๐—ฟ๐˜‚๐—ฐ๐˜๐˜‚๐—ฟ๐—ฒ ๐˜๐—ผ ๐—ฑ๐—ฒ๐—น๐—ถ๐˜ƒ๐—ฒ๐—ฟ ๐—”๐—œ: Data Centers (DCs) of course, but also all that is required to fire up these DCs, including the chips (compute and memory) and the energy (lots of it โ€“ hence the renewal of nuclear energy). Beyond current technologies, many bets on tomorrowโ€™s, in energy (fusion), accelerated computing (what will challenge GPUs; how quickly optics will get into semis, etc), and of course quantum computing (hiding somewhat the application of quantum to sensing and communications).

๐Ÿ”น๐——๐—ฒ๐—ณ๐—ฒ๐—ป๐—ฐ๐—ฒ. The numbers are staggering: European defence expenses are set to grow from 2% to 5% of GDP per annum. Thatโ€™s at least another $500bn (about 10x the size of the annual VC market), which prime contractors (the cost-plus type) and new upstarts are vying for. Among the areas of interest: drones (600 startups in Europe and counting); cyber (making existing infrastructure more resilient to attacks); new space/telecoms (building the European alternative to SpaceX/Startlink), etc. Behind it all the idea to recycle off-the-shelf commercial technology (โ€œdual useโ€) and shaking up procurement process (easier said than done).

As usual, better swim with the tide!
โ€
Sources
โ€https://www.a16z.news/p/state-of-marketsโ€
โ€https://techcrunch.com/2026/01/12/at-least-36-new-tech-unicorns-were-minted-in-2025-so-far
/
โ€https://techcrunch.com/2026/01/09/the-venture-firm-that-ate-silicon-valley/