Clem Delangue 🤗

Top trends on US startups and VC from Carta - very insightful. I was kind of surprised to learn that new hires get way less equity than before. Now with fewer employees, shouldn't it be more equity for each hire? 𝟰) 𝗧𝗵𝗲 𝗮𝘃𝗲𝗿𝗮𝗴𝗲 𝗻𝗲𝘄 𝗵𝗶𝗿𝗲 𝗿𝗲𝗰𝗲𝗶𝘃𝗲𝘀 𝟱𝟬% 𝗹𝗲𝘀𝘀 𝗲𝗾𝘂𝗶𝘁𝘆 𝘁𝗼𝗱𝗮𝘆 𝘁𝗵𝗮𝗻 𝗶𝗻 𝟮𝟬𝟮𝟮 (𝘀𝗹𝗶𝗱𝗲 𝟴𝟰) Startups are building with fewer employees than ever - meaning the leverage in hiring conversations is on the side of the employer. Salaries haven't kept pace with inflation and equity has been halved in 3 years. The only employee group bucking this trend are AI/ML engineers.

8

« The top 20 companies generate 95% of the sector's revenue, while the rest compete for the remaining 5%. » It is therefore necessary to better source and detect/qualify future startups that will be successful. This is particularly true in deep tech. To do this, we need to invest earlier in the TRL. Anticipating future successes is our credo at InnoTechValue. That's why we have developed predictive (rather than just generative) AI to identify future Deeptech gems from the laboratory stage to the market. To do this, we analyzed more than 50,000 projects. To put science into finance and finance into science https://lnkd.in/epNa-ray

6

1 Comment

Complete ranking of the 25 most valuable GenAI startups in 2026: Just a few years ago, most of these were pitch decks. Today, the top 10 alone are worth over $1.74T combined. In just the first months of 2026, five of these companies either raised or went public - bringing in $51B+. 1️⃣ OpenAI: raised $110B at an $840B valuation The largest private funding round in history ever. Amazon put in $50B. Nvidia $30B. SoftBank $30B. 2️⃣ Anthropic: $30B Series G at a $380B valuation Their annualized revenue is now $14B, more than double what it was a year ago. Claude Code alone is doing $2.5B ARR. 3️⃣ xAI: raised $20B Series E, then got acquired by SpaceX A month after closing the round, SpaceX bought xAI outright, valuing it at $250B as part of a combined $1.25T entity. 4️⃣ Legora: $550M Series D at a $5.55B valuation Valuation tripled from October 2025 to today. 800 law firms now use it across 50+ markets. 5️⃣ MiniMax: IPO on Hong Kong Stock Exchange Raised $619M. Shares closed up 109% on debut day. Market cap hit $13.7B by end of day, up from a $4B pre-IPO valuation. Founded in 2022. Listed in 2026. The fastest AI company ever to go from founding to public markets. Some quick stats from the full ranking: - Combined valuation: ~$1.84T - Median valuation: ~$10B - Average company age: roughly 3-4 years -- Bonus: Fluently: AI English tutor 💰 $2.5M+ raised (YC W24) Improve your English speaking skills with AI 👉 GetFluently.app -- Vertical AI is climbing fast - Mercor jumped 10 spots, ElevenLabs 8, Lovable 6 - each owning one specific workflow deeply instead of trying to do everything. The application layer is catching up to the labs. Your thoughts?

81

2 Comments

VC Real Talk #5: Fighting Ghosts “No competitors” is a red flag… every business faces rivals or substitutes. Ignoring them makes you look naïve or unprepared. Share how you analyze competition or rebut the need to address rivals. Drop your experiences… challenge or add insights. Final episode: Are you chasing capital before customers? “Premature Raise” drops next. #Startups #Entrepreneurship #Innovation #VentureCapital #Leadership #VCRealTalk

24

Silicon Valley is only now waking up to brand marketing. For a place that prides itself on being the most innovative ecosystem in the world, I keep watching many pour tens of millions into startups with “revolutionary GTM” strategies… only to peek under the hood and find the most basic playbooks dressed up in jargon. Let’s talk about it. I’ve been in SF for a couple of weeks, and while the Agentic Commerce discourse is fascinating, the conversations around “building a brand” are surface-level. A brand is not a logo. It’s not a few nice photos on a website. And it’s definitely not edgy videos clipped for virality. Four years ago, I built a 10,000-person waitlist for a consumer fintech app before writing a single line of code. How? Branding. Before ChatGPT’s Sora, there was Reface. I used memes to personalize every welcome email -literally inserting each person’s face into content - to make the experience unforgettable and inherently shareable. It spread like wildfire. That playbook came from two places: 👉 Hollywood, where I cut my teeth as an actress 👉 And Jack Dorsey’s Cash App, which turned marketing into cultural currency. Jack Dorsey is, in my view, one of the greatest marketers of our time. By deeply understanding human behavior, he transformed Cash App from a money-losing product into a cultural juggernaut - blending utility with identity. It wasn’t just fintech; it was culture. In 2015–2016, Cashapp launched $Cashtags, permanent usernames like $KatGarcia. These worked like Venmo handles, but were public and brandable. Suddenly people could share their Cashtag on Twitter, merch, or music lyrics - virality loop. Example: Rappers would literally drop their Cashtags in songs or bios, and fans would tip them. He tapped into hip hop culture which led to network effects and virality loops non-dependent on Meta and Google. Today, I’m building Ground AI (AI growth layer for the Agentic Commerce era) and the emerging + enterprise clients we power understand this intuitively. Brand is their oxygen. Product is a commodity - but creativity isn’t. It’s what turns good products into cultural movements, and cultural movements into millions in revenue. As both consumer and B2B AI enters its renaissance, brand is once again becoming the ultimate differentiator. Just like Liquid Death convinced people to pay $8 for a can of water, the brands that master brand equity and brand identity - from product to price to positioning - will scale fastest. Because in a world where distribution is agentic and AI-driven, brand is the anchor that makes everything else work. Jack Dorsey knows this intuitively well and it’s why this is Day 3 of getting him to invest or advise me in my building of Ground AI.

49

16 Comments

We used product-market fit framework from Sequoia Capital to design our go-to-market. TL;DR: PMF comes in different types- depending on the type of market. 1️⃣ 𝐇𝐚𝐢𝐫 𝐨𝐧 𝐅𝐢𝐫𝐞 (urgent, obvious pain) Customers feel a sharp pain now. They are actively looking for a fix. 𝐒𝐢𝐠𝐧𝐚𝐥𝐬: inbound leads name the problem unprompted. Budget exists. High “do you have this?” intent. Short sales cycles. Clear willingness to switch. 𝐁𝐮𝐲𝐞𝐫𝐬 𝐰𝐚𝐧𝐭: fast relief, reliability, clear ROI, low implementation friction. 𝐖𝐡𝐚𝐭 𝐰𝐢𝐧𝐬: outcomes that feel meaningfully different, delivered on one focused beachhead workflow. Often, Distribution >> Product. 𝐘𝐨𝐮𝐫 𝐣𝐨𝐛: win one wedge end-to-end, then expand. 𝐆𝐓𝐌 𝐭𝐡𝐚𝐭 𝐟𝐢𝐭𝐬: tight positioning, aggressive proof, fast iteration off objections. 𝐅𝐚𝐢𝐥𝐮𝐫𝐞 𝐦𝐨𝐝𝐞: going broad too early, shipping “nice-to-have,” narrating the future instead of proving relief today. 2️⃣𝐇𝐚𝐫𝐝 𝐅𝐚𝐜𝐭 (status quo belief) [this is us] The market has normalized the pain. Buyers think it is “just how it is.” Not actively looking for a “solution” cos they think its unsolvable. 𝐒𝐢𝐠𝐧𝐚𝐥𝐬: buyers are not searching. They accept workarounds. They optimize around constraints. 𝐁𝐮𝐲𝐞𝐫𝐬 𝐰𝐚𝐧𝐭: safety, legitimacy, and a path that does not make them look reckless. 𝐖𝐡𝐚𝐭 𝐰𝐢𝐧𝐬: reframe the “hard fact” as solvable, then make change feel inevitable and safe. 𝐘𝐨𝐮𝐫 𝐣𝐨𝐛: educate the market (👀), prove the shift with evidence, and de-risk adoption with pilot → rollout. 𝐆𝐓𝐌 𝐭𝐡𝐚𝐭 𝐟𝐢𝐭𝐬: credible thought leadership (👀👀), references, data, internal champions, compliance-ready messaging. 𝐅𝐚𝐢𝐥𝐮𝐫𝐞 𝐦𝐨𝐝𝐞: manufacturing urgency, or asking for a big leap before trust exists. 3️⃣ 𝐅𝐮𝐭𝐮𝐫𝐞 𝐕𝐢𝐬𝐢𝐨𝐧 (new paradigm) Buyers cannot fully imagine the new world yet, even if the value is real. 𝐒𝐢𝐠𝐧𝐚𝐥𝐬: high curiosity, slow decisions, “interesting,” unclear budgets, frequent “not now,” hard internal alignment. 𝐁𝐮𝐲𝐞𝐫𝐬 𝐰𝐚𝐧𝐭: belief the future is real, plus proof you can get there without disruption, while solving something today 𝐖𝐡𝐚𝐭 𝐰𝐢𝐧𝐬: endurance plus small “pit stops” that create real commercial traction. 𝐘𝐨𝐮𝐫 𝐣𝐨𝐛: turn the big vision into provable steps with clear milestones. 𝐆𝐓𝐌 𝐭𝐡𝐚𝐭 𝐟𝐢𝐭𝐬: design partners, prototypes-in-production, narrow use cases that ladder up, narrative backed by demos. 𝐅𝐚𝐢𝐥𝐮𝐫𝐞 𝐦𝐨𝐝𝐞: using Hair on Fire tactics before the market believes. Here’s what founders miss: 𝐇𝐚𝐫𝐝 𝐅𝐚𝐜𝐭: education first. Aggressive selling backfires. 𝐅𝐮𝐭𝐮𝐫𝐞 𝐕𝐢𝐬𝐢𝐨𝐧: break the story into milestones that prove traction (often in adjacent segments first). 🎯 PMF is not one-size-fits-all. Which one are you in right now? PS: Follow-up post tomorrow ;) #founder #pmf #gtm

18

Excited to partner with Respan (formerly Keywords AI), but more importantly, excited for people to try the platform! It really is a step function for developing real, production ready agents (gateway, routing, evals & observability in one). More here: https://lnkd.in/dab_i78P Give it a try when you're ready to move from a personal Claude agent that works 70% of the time, to something that's ready for every day use and serving real customers (or just more reliable for your own use!). Looking forward to working together, Andy Li, Hendrix Liu and Raymond Huang

21

2 Comments

70 million users. $100M ARR. 50 employees. Most AI companies would need 200+ people to hit these numbers. Gamma did it with a team smaller than most Series A startups. In less than 5 years, Grant Lee grew Gamma from $0 to $100 million in ARR and 70 million users. They’ve recently raised $68M at a Series B at a $2.1B valuation, led by Sarah Wang and Andreessen Horowitz. Here’s what NOBODY knows about Gamma: Gamma wasn’t an overnight success. They built a content engine where every leader became a creator, sharing tactics, frameworks, and behind-the-scenes insights that turned LinkedIn into their primary growth channel. Not only this, but they turned their network of active users into a community growth channel. And it's how they turned a product an investor called "the dumbest idea I've ever heard" into a $2.1B company. Just take a look at some of their key players: Grant Lee, their Co-Founder & CEO, went from hearing "the dumbest idea I've ever heard" on a pitch call to building a $2.1B company with $100M ARR. He posts thought leadership content to over 92K followers, sharing insights on profitable growth, lean teams, and building AI-native products. Jon Noronha, their Co-Founder & CPO, pioneered A/B testing at Optimizely and brought that experimentation DNA to Gamma. With his Harvard CS background and Microsoft pedigree, he shares product insights with his network about building AI products that people actually love using, not just testing. Kristin Fracchia, their Head of Marketing, orchestrates an "always-on" creator strategy with 150+ micro-influencers across platforms. She posts hot takes on growth and marketing for AI-native startups to over 5K followers, sharing how Gamma achieved 1.36M+ impressions through creator partnerships and turned viral content into a science. Zach Leach, their Head of Design, leads a team that punches way above its weight, using AI to synthesize feedback from 60+ countries, analyze 550 weekly responses, and maintain brand consistency at global scale. He shares design insights about building AI-powered products that feel magical, not mechanical. Diana Kimball Berlin, leading Core Product, joined Gamma after 4.5 years as a Partner at Matrix. Her background? She literally worked on PowerPoint at Microsoft, launched Quip Slides, and has been obsessed with "desktop publishing" since high school. Now she's building what she calls "the best AI-powered presentations & storytelling tool in all the land" to her 6.9K followers. Ravish Agrawal, on Growth, is a YC alum and Forbes 30u30 honoree who previously scaled Able (YC W20) to 25K+ placements. He shares insights on AI automation, growth tactics, and building scalable systems to 14.8K followers. Together, they make Gamma impossible to ignore. Every post adds credibility, and every conversation drives visibility back to the product. This is how you properly do EGC in 2026. Take notes!

128

21 Comments

The fastest way to kill a startup is to try to solve too many problems. In the early stages, founders often believe that adding more value will accelerate growth. So the product expands. More features. More use cases. More audiences. At first, this feels like progress. But expansion before focus usually creates confusion. Customers struggle to understand who the product is really for. Sales conversations become longer and more complicated. Marketing messaging loses clarity. And when positioning becomes unclear, the business starts paying the price. Customer acquisition cost rises. Conversion rates drop. Retention begins to soften. Not because the product is bad. Because the product is unclear. Strong startups usually begin with painful focus. One problem. One audience. One core outcome. This level of clarity simplifies everything. Customers understand the value immediately. Sales cycles become shorter. Product decisions become easier. But the temptation to expand appears early. A large client requests additional features. A new segment shows interest. An investor suggests pursuing a broader market. The instinct is to say yes. But strong founders pause and ask a different question: Does this strengthen our core — or dilute it? Startups rarely fail because they focused too much. They fail because they expanded before their foundation was strong enough.

7