What Figma and Databricks are Really Buying in Their Most Recent Acquisition
PLUS: Former Google Engineer Building a New Platform Real-Time Voice and Video AI
Episode of the Week
This week, the Geeks of the Valley podcast released Episode #112 featuring Ivan Casadevall, CEO of ARYADUTA Hotel Group, operating 8+ upscale hotels as Indonesia’s most iconic hospitality brand. With nearly three decades of leadership across Southeast Asia - including roles at Pan Pacific, Minor International, and MNC Land - Ivan brings deep insight into the future of luxury hospitality in a post-pandemic world.
In this episode, Ivan shares how ARYADUTA is expanding into experiential destinations like Ubud and Labuan Bajo, blending modern technology with timeless service. From sustainability to brand reinvention, he explores what it takes to thrive in a rapidly evolving travel market. Whether you're a founder, operator, or just passionate about hospitality, this episode is full of perspective. Tune in!
This edition is brought to you by VARA Network
Built on the Gear Protocol - Vara Network is a revolutionary Web3 application platform, removing many of the barriers associated with building, deploying, and operating dApps. Vara’s computer science paradigms like persistent memory and actor model, along with parallel data processing, allows for deep scalability, new design patterns and features. Vara’s unique architecture makes it completely frictionless for users. Learn more at Vara.Network
Venture Radar
Last week marked a pivotal moment for two companies we featured early on - Modyfi (featured on 18th September 2022) and Fennel (featured on 9th July 2022) - both of which were acquired by Figma and Databricks respectively. We’ve pulled together a few thoughts on why these startups originally caught our attention, and why their acquisitions may signal something bigger on the horizon.
At Geeks of the Valley, we've been closely tracking the intersection of generative AI and creative tooling - and this past week, a major signal emerged from that convergence. Figma, the collaborative design giant, has acquired Modyfi. While the acquisition was relatively quiet in media cycles, we believe this deal speaks volumes about where design is headed - and why early bets on AI-native infrastructure are beginning to pay off.
A Generative Design Engine Built for Today’s Creative Teams
Modyfi set itself apart early by embracing generative design at its core. Their platform combined real-time multiplayer editing, raster and vector support, and AI-driven image creation - all in-browser. This wasn’t AI as a bolt-on feature; it was deeply embedded in the workflow, allowing users to prompt, manipulate, and iterate on visual content without switching tools or contexts.
For those of us watching the design-tech space, Modyfi stood out not just for its clean interface or speed, but for how naturally it brought prompt-based creativity into the professional design stack. It positioned itself somewhere between Photoshop, Midjourney, and Figma - and made that fusion feel seamless.
Why This Acquisition Makes Strategic Sense
From our vantage point, Figma’s acquisition of Modyfi is a timely response to 2 converging pressures:
The design brief is changing: What used to be a multi-step process involving ideation, drafting, asset creation, and prototyping is now collapsing into fewer, faster, and more iterative loops. Tools like Modyfi are architected for that velocity. Figma’s core stack wasn’t - until now.
Collaboration is no longer enough: Figma won the first battle by making design multiplayer. But the next wave is about making design intelligent - enabling teams to move from idea to asset in seconds, not hours. This acquisition accelerates Figma’s leap into that space.
The Bigger Signal: The Design Stack is Being Rewritten
What excites us at Geeks of the Valley is what this move signals for the rest of the ecosystem. Creative teams are no longer satisfied with static design tools or cloud-based versions of desktop workflows. They want systems that feel alive - responsive, generative, and embedded with context. And they're willing to shift tooling fast when something better emerges.
Real-time data infrastructure is having its moment - not because it’s new, but because AI is finally making it necessary. With LLM-powered apps, personalization engines, and recommendation systems demanding sub-second responses, the underlying data stack needs to evolve. That’s why Databricks’ recent acquisition of Fennel feels like more than just a tactical move - it’s a clear signal that the era of reactive AI infrastructure has officially arrived.
Fennel’s Vision: Streaming Infrastructure Without the Headaches
Fennel quickly stood out by tackling one of the least glamorous, but most critical problems in modern data systems: real-time feature engineering. Built for teams deploying recommendation engines, personalization, fraud detection, and more, Fennel offered a single platform that unified batch and streaming pipelines - while maintaining low latency, consistency, and observability.
Its appeal? The magic was in the simplicity. Fennel let teams define features in a declarative way - and then handled everything under the hood: DAG construction, backfills, versioning, and materialization. No more glue code, no more managing Kafka, Flink, and Redis independently. Just one system that worked end-to-end, designed to serve fresh, accurate features at scale.
Why Databricks Made the Move
From our perspective, the acquisition of Fennel is less about competitive defense and more about strategic depth. Databricks has built its reputation on unifying data lakes and warehouses with AI and ML - but when it comes to real-time operational AI, it has historically lagged behind purpose-built platforms.
Fennel brings precisely what Databricks was missing: an elegant, developer-first layer for real-time ML infrastructure. With demand surging for AI-native products - from TikTok-style feeds to personalized LLM agents - embedding Fennel into the Databricks ecosystem could turn what was once a batch-heavy stack into a continuous, low-latency intelligence engine.
Zooming Out: AI is Getting Operational - Fast
What excites us most about this deal isn’t just the tech integration. It’s the signal it sends. As AI goes from experimentation to deployment, the bottleneck is no longer model quality - it’s infrastructure. Feature freshness, latency, and consistency are becoming the differentiators between good and great user experiences.
By acquiring Fennel, Databricks isn’t just upgrading its stack - it’s preparing for a world where every AI model needs real-time context, and every product becomes more intelligent by the second.
Geeks of the Week
Startup Name: Outspeed
Geography: US
One-liner: Platform for Realtime Voice and Video AI
Founder(s) Background: Senior Data Scientist at AutoGrid (acquired by Schneider Electric), Software Engineer (Infrastructure) at Google.
Thoughts:
Outspeed is building the infrastructure layer for real-time, multimodal AI, addressing a fast-growing need as voice and video become dominant interfaces across telehealth, customer service, education, and beyond.
Its developer-first approach reduces technical barriers with a PyTorch-inspired SDK and clean API that abstracts away latency, GPU optimization, and WebRTC complexity - allowing teams to ship in days, not months.
Positioned for defensibility, the platform isn’t just a toolkit - it’s enterprise-grade infrastructure, offering SOC2, GDPR, and soon HIPAA compliance, making it a strong fit for regulated industries.
Under the hood, Outspeed powers real-time transcription, translation, agentic AI, and vector embeddings with production-grade performance - enabling the next generation of AI-native user experiences.
Founder(s) building in stealth
Chief of Staff / Chief Strategy Officer at Casavo (Series D).
Pre-Seed Investments / Venture Development at Rocket Internet.
Deals of the Week
Venture funding activity held strong last week across early and mid-stage rounds, with AI infrastructure, cybersecurity, fintech, and biotech continuing to dominate investor interest. Notable activity was seen across the US, Europe, and Asia as VCs doubled down on startups tackling core infrastructure, productivity, and health-related challenges.
Auradine – the Santa Clara-based blockchain and AI infrastructure startup – raised $153M in a Series C round led by StepStone Group, with participation from Samsung Catalyst Fund and Qualcomm Ventures. Auradine is building secure, energy-efficient computing platforms for digital assets and AI workloads.
Glycomine – a San Carlos-based biotech firm – secured $115M in Series C funding from CTI Life Sciences Fund, Novo Holdings, and Sanofi Ventures. The startup is advancing clinical development of GLM101, a therapy for PMM2-CDG, a rare genetic metabolic disorder.
Exaforce – a San Jose-based security operations startup – raised $75M in Series A funding to scale its AI-native platform designed to automate and protect enterprise cyber defense workflows.
Chapter – a New York-based Medicare navigation platform – raised $75M in a Series D round led by Stripes, but continues to operate like a Series B/C-stage company in terms of product-market expansion and infrastructure. It’s building a vertically integrated tech stack for healthcare navigation.
Crux – a New York-based climate finance platform – secured $50M in Series B funding led by Lowercarbon Capital and Andreessen Horowitz. Crux facilitates the transfer of clean energy tax credits, unlocking new liquidity in renewable markets.
Marshmallow – a London-based digital insurer – raised $90M in Series C funding led by Portage Capital. The company uses AI and alternative data to underwrite auto insurance products, targeting underserved customer segments in the UK.
Vizzy – a London-based AI-driven video editing startup – secured £3.65M in seed funding led by Adjuvo. Vizzy helps creators automate professional-grade edits with natural language prompts.
DELLI – a UK-based social commerce platform for independent food producers – raised $6M in funding to expand its curated marketplace and creator ecosystem across Europe.
SplxAI – a Croatian AI security startup – raised $7M in a seed round led by Launchub Ventures. The startup simulates adversarial attacks on LLMs to identify vulnerabilities before deployment.
Notable Mergers and Acquisitions
Strategic consolidation continues to shape the startup ecosystem, with early to growth-stage companies actively engaging in acquisitions to strengthen core capabilities, expand product offerings, and deepen their competitive moats. The past week saw notable M&A activity in AI infrastructure, autonomous systems, fitness tech, and data analytics.
One standout move came from Databricks, which acquired Fennel, a real-time analytics startup focused on low-latency data pipelines and event tracking. The acquisition boosts Databricks’ position in the enterprise AI stack by enhancing its real-time analytics capabilities, a growing demand across customer-facing and AI-native products.
In the AI developer ecosystem, OpenAI is reportedly in late-stage talks to acquire Windsurf, a code-generation assistant startup, for over $3 billion. Though not yet finalized, the potential acquisition highlights OpenAI’s ambition to vertically integrate key developer tools built on its own foundation models.
Meanwhile, in autonomous systems, Applied Intuition acquired EpiSci, a startup known for building autonomous software across air and sea platforms, including unmanned drones for DARPA and the U.S. Navy. This move extends Applied Intuition’s reach beyond automotive autonomy into defense and aerospace domains.
On the consumer front, Strava announced its acquisition of Runna, a UK-based fitness coaching app popular among runners. By integrating Runna’s personalized coaching tools, Strava aims to diversify beyond activity tracking and tap into the growing demand for AI-enhanced training programs.
These deals reflect a broader trend where startups are no longer just acquisition targets - they’re increasingly becoming strategic buyers in their own right, using M&A as a lever for speed, scope, and differentiation in hypercompetitive markets.
This edition is brought to you in partnership with Stella Capital.