Collect 100% MRR cash, AI playbook, Claude Opus 5 halves cost

Jason Lemkin warns that SaaS startups often miss cash on invoiced deals, ending up with only 60‑70% of their MRR in the bank. He proposes a simple KPI: collect at least 100% (ideally 110%) of monthly recurring revenue each month, which can extend runway dramatically and prevent hidden cash‑flow gaps.

At SaaStr AI 2026 three AI veterans, Anthropic’s Eleanor Dorfman, Atlassian’s Sharif Mansour, and Scale’s Rory O’Driscoll, urged SaaS founders to build on the software they already own instead of chasing every new model. Their playbook shows how to embed chat, dedicated UI, and custom workflows on existing products, delivering value faster and with less risk.

A new review of clinical trial data shows the failure rate has hovered around 90% for the past three decades, barely better than the 80% seen in the 1970s‑80s. For biotech founders and investors, this steady high attrition underscores the persistent risk in drug development and the need for smarter target selection.

Data from ICONIQ shows three‑year contracts fell from 28% to 23% of new logos between 2023‑2026, while sub‑one‑year deals rose to 13%. Rapid AI replacement cycles make long terms risky, and only companies with >110% NRR can secure them without discounts. Focus on quick ROI and renewal quality instead of contract length.

The essay draws a direct line from Kubernetes’s rise to today’s open‑weight model surge, arguing that a neutral, open‑weight substrate will drive the next wave of AI infrastructure innovation. If the analogy holds, startups and vendors will rally around open‑weight models, reshaping the market’s competitive landscape.

Anthropic launched Claude Opus 5, a model that matches the intelligence of its flagship Fable 5 at roughly half the price. Benchmarks show Opus 5 setting new state‑of‑the‑art scores on coding, knowledge work, and scientific tasks while using far fewer tokens, making it the default on Claude Max and the strongest option on Claude Pro.