OpenAI, Anthropic give $800M in free AI credits to YC startups

Prime Intellect announced a $130 million Series A, led by Andreessen Horowitz and Sequoia, to build a platform where firms can train and deploy their own agentic AI systems. The move aims to break the reliance on frontier AI labs, giving companies direct control over high‑impact AI agents and opening new revenue streams.

OpenAI promised each Spring 2026 YC startup $2 million in API credits tied to equity, while Anthropic upped its offer to $500k with no strings. Both firms now extend up to $500k no‑equity credits plus optional $1.5 million equity tranches, potentially dumping $800 million of compute per year to lock in the next wave of AI companies.

The Swedish AI startup Lovable is courting a $300M raise led by Menlo Ventures, which would double its valuation to $13.2B after hitting $500M annualized revenue run rate. Its vibe‑coding platform powers sites for founders and enterprises like Nvidia, signaling AI‑driven web‑building is a hot market.

General Intuition just closed a $320 million Series A at a $2.3 billion valuation, betting that millions of hours of video‑game play can train a foundation model for physical AI. The model can run a quadrupedal robot after only eight minutes of real‑world fine‑tuning, suggesting a shortcut to broader robot development. If successful, companies could skip data collection and build smarter machines faster.

A new Harvard Business School working paper shows AI‑focused YC startups are about a quarter smaller than their non‑AI peers and employ roughly 15% fewer entry‑level staff. The leaner, flatter teams translate into significantly higher valuations per employee, signaling a talent‑intensive shift for founders and investors.

Even though SaaStr built its own AI agents that sit on top of Salesforce, most companies won’t replace their CRM with home‑grown code. Compliance, deep integrations, scaling to hundreds of reps, and the maintenance load make roll‑your‑own solutions brittle beyond the startup stage. The post outlines four concrete reasons why the trend will stay niche.

Jason Lemkin argues that protecting net‑revenue‑retention by letting a logo churn is short‑sighted. A well‑priced renewal discount can keep a long‑standing customer, buy another year to improve the product, and ultimately boost growth despite a temporary dip in NRR. The NRR premium is shrinking, so the trade‑off now favors retention.

In the SaaStr AI 2026 AMA, founders argued that B2B AI agents should deliver 120% of a top salesperson's output, not merely replicate it. When agents exceed human performance, they unlock faster outbound targeting, qualified meetings, and real‑time support, reshaping GTM speed and effectiveness.

Zain Jaffer, who turned Vungle into a $780M Blackstone exit, says the SaaS growth hacks that got him there now choke AI‑native startups. He’s flattening org charts, betting on proprietary data as the real moat, and swapping easy‑revenue agency models for a platform that can scale without price wars. The shift shows how old VC instincts must evolve for the next wave of AI businesses.

A senior engineer admits to “LLM burnout” after hours of reading AI‑generated code and text. The constant style quirks, hallucinations, and repetitive errors are eroding productivity and morale, a warning sign for teams leaning heavily on LLM‑driven workflows. Ignoring it could hurt adoption and staff well‑being.

In the past year AI companies have earmarked $9.75 billion for forward‑deployed engineering (FDE) teams, turning embedded engineers into an industry standard. Structured as balance‑sheet units, standalone funds, or partner ecosystems, these investments lock in customer knowledge and create costly switching barriers.

AI can exploit codebases with clear, consistent patterns, cutting token usage and boosting output quality. That flips the cost‑benefit calculus, turning previously uneconomic rewrites into attractive projects. Legacy or proprietary stacks lose this edge, so firms should refactor toward AI‑friendly designs to stay ahead.

Paradigm closed a $1.2 billion fourth fund that will back AI, robotics, and other frontier tech while still backing crypto infrastructure. The move signals a strategic pivot as the firm rides the AI boom and diversifies beyond a market that’s facing headwinds. Early bets already include drone delivery startup Zipline and space venture True Anomaly.