Two 23-year-olds from rural France just beat Google DeepMind and raised $4.1M.

Minitap makes mobile development 10x faster with AI agents that write, test, and ship features autonomously.

• $4.1M seed round co-led by Moxxie Ventures and Mercuri
• Beat Google DeepMind on AndroidWorld benchmark in 40 days
• 1,900 GitHub stars, 70M+ user apps as design partners

How it happened:
• Met in rural Burgundy, taught each other coding
• Built viral app Fuego to 10K users (bootstrapped)
• Combined AI research + mobile expertise + infrastructure skills
• Built mobile-use framework + minitap cloud
• Claimed #1 benchmark position in 40 days
• Open-sourced solution, attracted major design partners

The gap: AI transformed web development but mobile stayed slow. Minitap bridges that.

https://technews180.com/funding-news/how-two-23-year-olds-beat-google-and-raised-4-1m-for-minitap/

Takeaway: Technical depth + market timing + execution beats big tech advantages.

Source: TechNews180

Yoodli just closed $40M Series B with 900% revenue growth in 12 months.

What it is: AI roleplay platform for sales/communication training. Think "batting cage for conversations."

The metrics:
• $40M Series B (WestBridge Capital led)
• 900% revenue growth YoY
• Team tripled to 40+ people
• Customers: SAP, Google, Snowflake
• <1 year since $13.7M Series A

How it happened:
• Started consumer (public speaking practice)
• Pivoted to B2B sales training
• Built interactive roleplay vs passive slides
• Focused on ROI-measurable outcomes
• Expanded to "experiential learning"
• Landed enterprise logos early

Takeaway: Interactive beats passive. Replace training videos with hands-on practice.

Source: GeekWire

Brewed by Startup Beaker.

Google bets $100M on AI security cameras.

Verkada hit $5.8B valuation in CapitalG-led round — up $1.3B since February. The startup crossed $1B in annualized bookings.

• $100M funding led by Google's CapitalG
• $5.8B valuation, $1B+ bookings
• 60+ AI features launched in September
• 30,000 customers, 20M+ images processed hourly

How it happened:
• Filip Kaliszan couldn't find good home cameras (2016)
• Partnered with Meraki co-founder for hardware expertise
• Offered free re-installations to win early customers
• Expanded through channel partners
• Launched 60+ AI features to beat legacy players

Takeaway: Legacy hardware markets are ripe for AI disruption when founders solve their own problems first.

Source: The Tech Buzz

German AI startup Black Forest Labs just closed a $300M Series B at $3.25B valuation.

What it is: Visual AI company behind FLUX image generation models, launched August 2024.

The numbers:
• $300M Series B (total $450M raised)
• $3.25B valuation in 16 months
• Led by Salesforce Ventures + AMP
• Enterprise customers: Adobe, Canva, Meta, Microsoft

How it happened - Founder steps:
• Built superior tech: FLUX outperforms competitors on prompt adherence
• Dual strategy: Open-source for developers + premium enterprise offerings
• Strategic partnerships: Integrated with major platforms
• Enterprise focus: Targeted Fortune 500 workflows from day one
• Network effects: Open-source adoption drove enterprise demand

Takeaway: Technical superiority + dual go-to-market strategy can create billion-dollar outcomes in 16 months.

Source: Entrepreneur Loop

#startups #AI

Zego alumni just cracked the software selection problem.

Dragonfly raised £2.6M pre-seed and launched their AI-powered software selection tool. Founded December 2024 by Sean King and Sven Sabas.

Key metrics:
• 250,000+ software products catalogued
• Conversational AI answers in seconds vs months of research
• 12-person team, mostly ex-Zego talent
• Led by Episode 1

How they did it:
• Built world's largest software catalogue first
• Developed conversational AI layer on top
• Assembled experienced team before launch
• Secured funding pre-launch

Takeaway: Software selection paralysis is real. Build the data moat first, then the AI layer.

Source: https://startupsmagazine.co.uk/article-dragonfly-raises-ps26m-it-launches-ai-tool-simplify-software-selection

#startups #AI

Evalyze launched an AI investor matching engine that analyzes pitch decks and matches founders with relevant investors in minutes.

Built on 8,000+ successful funding applications and 12,000+ verified investors.

How it works:
• Upload pitch deck + company details
• Get investor-readiness score with improvement suggestions
• Receive curated investor list by stage, sector, geography

Founder insight: Ex-VC CEO Vahid Fakhr saw great teams waste months on repetitive fundraising loops. Solution: data-driven outreach over cold emails.

Key numbers:
• 6,000+ founders already using it
• Focuses on pre-seed/seed software and AI startups

Takeaway: Fundraising is pattern recognition. Use data from thousands of successful applications before you start pitching.

https://www.evalyze.ai

#startups #fundraising

PearAI got "cancelled" on day 1 of Y Combinator for cloning an open-source code editor. 3 months later: $1.25M seed round.

What happened:
• Released MVP that was essentially a copy of Continue editor
• Got roasted for licensing violations and bravado tweets
• Apologized within 24 hours, fixed licensing
• Pivoted from editor to AI coding framework
• Used feedback to build tool that integrates multiple AI coding platforms
• Graduated YC Winter 2024 with modified product
• Raised from Goodwater Capital, Multimodal Ventures, others

How they recovered:
→ Listened to criticism instead of defending
→ Fixed technical issues immediately
→ Pivoted based on market feedback
→ Turned controversy into product validation

Takeaway: Sometimes getting "cancelled" forces the pivot you needed.

Brewed by Startup Beaker.

Ask Sage cracked government AI with $17M Series A.

GenAI platform for defense/government sectors. Founded by Nicolas Chaillan (ex-Air Force/Space Force Chief Software Officer).

• $17M Series A, Sapphire Ventures lead
• 30,000+ DoD users, 14,000+ gov teams
• First GenAI with DoD IL5 + FedRAMP High certs

How it happened:
• Leveraged insider gov knowledge from military roles
• Built security-first, compliance-ready platform
• Secured massive regulatory barriers (IL5/FedRAMP)
• Proved PMF with 30K users before fundraising
• Used gov traction for top-tier VC validation

Takeaway: Domain expertise + regulatory moats = defensible govtech.

Brewed by Startup Beaker.

🔥 Fireside Chat in Tashkent / December 12 at C-Space

After the Digital Startup Awards ceremony, MOST invites you to an open discussion about the future of the venture market in the region.

📍 Location: MOST C-Space Hub
🕓 Time: December 12 at 16:00
🎙 Topic: “Central Asia Investment Hub: The Role of Uzbekistan in the Regional Venture Landscape”

Uzbekistan is rapidly strengthening its position in Central Asia as a new center for investment and technological innovation. During the Fireside Chat, we will discuss how the regional venture ecosystem is forming, why Uzbekistan is becoming a gateway for investors into Central Asia, and what opportunities are emerging for startups and corporations.

Speakers:

Pavel Koktyshev — CEO, MOST Holding

Bakht Niyazov — Founder of Justart VC Studio, venture investor

Uchkun Tulavov — Founder & CEO of BITO Ecosystem, entrepreneur, expert in business digital transformation

Dilmurod Khodiev — CEO, C-Space

We welcome venture investors, entrepreneurs and startup founders, corporate representatives developing CVC and partnerships, as well as everyone interested in venture trends and the evolution of Tech & Innovation in Central Asia.

🎁 Bonus for guests
Exclusive pre-sale of tickets to CEVF 2026 Tashkent will be available only at this event—lowest prices with a 30% discount.

Registration: https://forms.gle/2Lp7NFUVhfApwDHm7

Dragonfly just raised £2.6M and launched an AI that picks your tech stack in seconds.

Zego alumni Sean King and Sven Sabas built a conversational AI that cuts software selection from months to seconds. Instead of endless demos and outdated marketplaces, you ask questions in plain English.

• 250,000+ software tools catalogued
• Handles simple to complex enterprise queries
• Led by Episode 1, angels include QuantumBlack founder Sam Bourton

How it happened:
• Founders experienced tech selection pain at Zego firsthand
• Built world's largest software catalogue from trusted sources
• Launched public tool December 2024 alongside funding
• Planning enterprise offering with "digital fingerprinting" for late 2025

Takeaway: Software selection paralysis is real. AI-powered curation beats endless research.

https://startupsmagazine.co.uk/article-dragonfly-raises-ps26m-it-launches-ai-tool-simplify-software-selection

#startups #AI

Anchor co-founders just raised $16M Series A for AI course generator.

Oboe creates personalized learning courses from any topic using AI. Users define a goal, AI generates chapters, quizzes, and audio content.

• $16M Series A led by a16z
• 3 months after public launch
• Previously raised $4M seed in 2024
• Now offers unlimited course generation
• $15/month for deep chapters, $40/month Pro

How it happened:
• Built on Anchor/Spotify credibility
• Launched with instant content generation (no loading screens)
• Focused on STEM topics with quality sourcing
• Pivoted pricing to freemium model
• Secured major label partnerships for scale

Takeaway: Speed beats perfection in AI products.

Source: TechCrunch

Brewed by Startup Beaker.

H1-B to $5.5B: Harness raises $200M at $5.5B valuation, making founder Jyoti Bansal a new billionaire.

Harness automates software deployment with AI. Customers like United Airlines see 75% faster deployments. Growing 50% YoY with $250M+ revenue.

Bansal's playbook:
• Moved from India with few hundred dollars on H1-B
• Worked 7 years as engineer waiting for green card
• Built AppDynamics, sold to Cisco for $3.7B in 2017
• Got bored on safari after 6 months of retirement
• Founded Harness to solve adjacent problem in same domain
• Now worth $2.3B personally

Key insight: "Writing code is 30% of the job. Testing and deploying is 70%."

Serial founders with domain expertise build faster.

https://www.forbes.com/sites/simonemelvin/2025/12/11/how-this-indian-immigrant-went-from-an-h1-b-visa-to-building-a-55-billion-business/

Brewed by Startup Beaker.

Ex-OpenAI product manager who helped launch ChatGPT just raised $9.3M seed for Worktrace AI.

Angela Jiang (GPT-3.5/GPT-4 launch team) + Deepak Vasisht (MIT PhD) are automating business workflows by watching employees work.

• $9.3M from Conviction, 8VC, OpenAI Fund
• 5 design partners already using the agent
• 4 full-time + 3 part-time team
• Spots workflow patterns within 1 week
• 95% of enterprise AI tools fail to scale (MIT study)

How they built it:
• Met at South Park Commons hatchery
• 8 months from founding to launch
• Agent watches desktop work passively
• Ranks automation opportunities by impact
• Provides code-ready files for agent builders

Takeaway: Don't just build AI tools. Build AI that learns your users' actual workflows.

Brewed by Startup Beaker.

AI coding tools promised speed but delivered friction instead.

Kilo Code raised $8M seed (Cota Capital) after hitting #1 on OpenRouter and 750K downloads in under a year.

Key metrics:
• $8M from Cota Capital, Breakers, General Catalyst
• 750K+ downloads since early 2025
• 6.1 trillion tokens processed monthly
• 500+ models vs competitors' 3-5

Founder playbook (Scott Breitenother + Sid Sijbrandij):
• Identified incumbent friction: rate limits, confusing pricing
• Built transparent alternative with open pricing
• Used product velocity as moat: shipped 5 major features in one month
• Leveraged co-founder credibility (GitLab's Sid)

Lesson: When incumbents add friction, speed + transparency wins.

#startups #ai

Brewed by Startup Beaker.

Databricks just closed 2024's largest VC round: $10B Series J at $62B valuation.

Key metrics:
• $3B+ revenue run rate (60% YoY growth)
• 500+ customers paying $1M+ annually
• First-time positive free cash flow expected Q4
• 10,000+ organizations using platform

How Ali Ghodsi built this:
• Started with Apache Spark at UC Berkeley (2008) → Created open-source foundation
• Founded Databricks (2013) → Commercialized Spark vs dominant Hadoop
• Listened to customers → Built Delta Lake and MLflow for real needs
• Created Lakehouse architecture → Unified data lakes and warehouses
• Went cloud-first → Enabled scalability despite skepticism
• Expanded globally → Opened hubs worldwide

Funding led by Thrive Capital will fuel AI products, acquisitions, and international expansion.

Takeaway: Build on open-source, listen obsessively to customers, create unified platforms.

https://www.databricks.com/company/newsroom/press-releases/databricks-raising-10b-series-j-investment-62b-valuation

Brewed by Startup Beaker.

Lovable tripled its valuation to $6.6B in 5 months with a $330M Series B.

The Swedish "vibe-coding" startup lets anyone build apps with text prompts. Key metrics: $200M ARR in 18 months, 100k+ daily projects, 25M+ total projects created.

Founder lesson: CEO Anton Osika refused Silicon Valley relocation pressure, stayed in Stockholm. "You can build a global AI company from this country," he said at Slush.

Result: Access to European talent, $100M ARR in 8 months, Fortune 500 customers like Klarna and Uber who cut prototyping from 6 weeks to 3 hours.

The vibe-coding category is exploding alongside Cursor's $29B valuation.

#startups #AI

Brewed by Startup Beaker.

Lawhive just raised $40M Series A to disrupt "main street" law firms with AI.

What it is: AI-powered SaaS platform that automates legal work for small law firms, reducing costs by up to 50%.

• $40M Series A co-led by GV and TQ Ventures
• $11.9M seed in April → $40M Series A in 8 months
• AI "Lawrence" passed part one of UK Solicitors Qualifying Exam
• Targets $1T US legal market vs Big Law focus of competitors
• Claims 50% cost reduction for small firms
• Founded 2019, now expanding to US market

How it happened:
• Started with underserved market (small firms vs Big Law)
• Built proprietary AI models for legal automation
• Proved product-market fit in UK first
• Secured Google Ventures as lead investor
• Combined consumer brand with deep tech R&D
• Positioned as first mover in main street legal AI

Takeaway: While competitors chase Big Law, Lawhive found gold in the underserved small firm market.

Source: TechCrunch

Brewed by Startup Beaker.

Ex-Yahoo CEO pivots from failed startup to AI personal assistants

Marissa Mayer shuttered photo-sharing startup Sunshine after 6 years to launch Dazzle, an AI personal assistant company that just raised $8M seed at $35M valuation.

• $8M led by Forerunner's Kirsten Green
• Investors: Kleiner Perkins, Greycroft, Slow Ventures
• Sunshine investors got 10% Dazzle equity
• Coming out of stealth early 2026

How it happened:
• Team began prototyping Dazzle summer 2024
• Realized AI had "much bigger impact" than photo sharing
• Mayer admitted Sunshine problems too "mundane"
• Landed Green (Warby Parker investor) as lead
• Dissolved $20M Sunshine to focus on Dazzle

Takeaway: Even seasoned founders pivot for bigger opportunities.

https://techcrunch.com/2025/12/23/marissa-mayers-new-startup-dazzle-raises-8m-led-by-forerunners-kirsten-green/

#startups #funding

While everyone builds AI companions, 222 bets on real human connection.

The YC W23 startup just raised $10.1M to scale their "IRL social facilitator" that uses AI to match strangers for offline experiences.

• Founded by 3 friends (ages 22-24) who met in high school/college
• Started as Zoom research project → backyard dinners → full platform
• 16 employees, moved LA → NYC for expansion
• $17 per event or $22/month subscription model
• 100-question compatibility survey powers ML matching

How it happened:
• Built MVP during Web3 hype (investors kept asking "have you tokenized your event?")
• Secured initial $10K from Z Fellows + backing from 1517 Fund/General Catalyst
• Applied to YC twice, got in on second try
• Hit 18-month goals in 2 months through YC acceleration
• Leveraged AI hype cycle for Series A ("company has ML core to it")
• Focused on investors who believed in IRL experiences vs digital-first VCs

The contrarian bet: making money only when people meet IRL and enjoy it enough to return.

Takeaway: Sometimes the best tech play is using AI to get people offline.

https://www.wsj.com/articles/in-an-age-of-ai-this-startup-is-betting-on-irl-69608cd4

Brewed by Startup Beaker.

Lemon Slice just raised $10.5M from Matrix Partners and YC to fix the "uncanny valley" problem in AI avatars.

What it is: A 20-billion-parameter diffusion model that creates interactive digital avatars from a single image, streaming at 20 FPS on one GPU.

• $10.5M seed round led by Matrix Partners
• YC, Dropbox CTO, Twitch CEO, The Chainsmokers invested
• 20B parameter model runs on single GPU
• API + embeddable widget (one line of code)
• Works for human and non-human characters

How it happened:
• Founded 2024, started with video model experiments
• Built proprietary diffusion model vs using existing solutions
• Focused on real-time interaction (20 FPS streaming)
• Launched API-first approach for easy integration
• Secured enterprise customers in education, e-commerce, training
• Raised from top-tier investors who saw technical differentiation

https://techcrunch.com/2025/12/23/lemon-slice-nabs-10-5m-from-yc-and-matrix-to-build-out-its-digital-avatar-tech/

Takeaway: Technical moats in AI come from building your own models, not integrating existing ones.

Brewed by Startup Beaker.

Dragonfly raised £2.6M and launched an AI that picks your tech stack in seconds.

Two Zego alumni built the world's largest software catalog (250,000+ tools) wrapped in conversational AI. Ask it "best project management for 5-person team with Google Calendar integration" and get ranked recommendations with explanations.

Key moves:
• Founded December 2024 by Sean King & Sven Sabas
• Assembled 12-person team from Zego/Peppy/Permutive alumni
• Built proprietary "digital fingerprinting" for tech stack mapping
• Secured Episode 1 as lead + QuantumBlack CTO, Bolt CEO as angels

Why it matters: Tech stack decisions make or break startups. This democratizes solutions architect expertise.

https://startupsmagazine.co.uk/article-dragonfly-raises-ps26m-it-launches-ai-tool-simplify-software-selection

#startups #ai