Hi Seedradar Squad! December closed the year with a massive pivot toward the physical realities of the AI age. Capital didn't just flow into software; it flooded the foundational bedrock—from Databricks' colossal war chest to the kinetic defense of Castelion and the geothermal power of Fervo Energy.

Simultaneously, the model wars matured into a battle for specialization. OpenAI and Google split the atom of intelligence, launching distinct "thinking" and "instant" models to separate deep reasoning from raw speed, while AWS entered the fray with a versatile new family of models. The focus has firmly shifted from mere chat interfaces to deep, autonomous execution.

December delivered a monumental surge of capital into the absolute bedrock of the AI economy. Anchored by a record-breaking infusion for the world’s leading data platform, the month saw massive commitments flow into next-generation geothermal energy, hypersonic defense, and the critical security layers required to protect the modern enterprise.

On December 16, Databricks announced that it had secured $4 billion in a Series L funding round, a transaction that vaulted its valuation to $134 billion. This massive capital injection reinforces its position as one of the world's most valuable private software companies.

The company plans to use the funds to aggressively advance product development across its data and AI portfolio. Additionally, the capital will be used to provide liquidity for employees and support future AI acquisitions, further expanding its technological capabilities.

Read More

At Seedradar Ventures, we back relentless founders at the earliest stage, investing in teams solving big problems with solutions that deliver a 10–20x greater impact.

Our syndicate offers investors access to a curated pipeline of early-stage startups, sourced through our network.

👉 Ready to learn more? Explore the Seedradar Ventures Syndicate on AngelList.

A decisive evolution in model specialization and autonomous capability has arrived. OpenAI and Google escalated the arms race with dual-track releases that separate deep reasoning from instant speed, while AWS expanded the infrastructure frontier with a versatile new model family. Simultaneously, the boundaries of generative media shattered as Disney became the first major studio to integrate its iconic IP directly into the video generation ecosystem.

OpenAI has released GPT-5.2, a major update designed to compete directly with Google’s Gemini 3. The release features two distinct variants: a "Thinking" model engineered for deep strategic insights and complex reasoning, and an "Instant" model optimized for rapid, everyday tasks.

Both models are now integrated with Microsoft’s Work IQ to provide enhanced contextual understanding for enterprise users.

Read More

Early-stage capital is fueling a dynamic class of startups poised to be tomorrow's giants. This elite group includes pioneers reinventing the physics of computing to solve the global energy crisis, alongside ventures deploying autonomous scientists to accelerate biological discovery. From cinema-quality generative audio to real-time defense against deepfakes, these innovators are building the essential hardware and security layers for the next era.

In one of the largest seed rounds in history, Unconventional AI has raised $475 million (at a $4.5 billion valuation) to build a new computing architecture based on "thermodynamic annealing." Founded by Naveen Rao (ex-Databricks/Nervana) and researchers from MIT and Stanford, the company is moving beyond digital logic to design analog chips that leverage the "intrinsic physics" of silicon.

Backed by Lightspeed, a16z, and Jeff Bezos, they aim to solve the global data center power shortage by running LLMs with 99% less energy than GPUs, targeting "biology-scale" efficiency.

Read More

The industry’s leading minds converged to define the next phase of autonomous capability and physical infrastructure. From the agentic breakthroughs at AWS re:Invent to the "System 2" reasoning focus at NeurIPS, these events marked a pivotal moment in accelerating the shift from simple pattern matching to a new era of deliberate, energy-efficient, and governable intelligence.

At its annual Las Vegas conference, Amazon Web Services launched the "Amazon Nova 2" model family, featuring the "Sonic" speech-to-speech model and "Lite" reasoning model.

The event also introduced "Kiro," a new autonomous agent capable of handling complex workflows, signaling a major shift in AWS's strategy toward agentic AI and reduced infrastructure overhead for developers.

Read More