**Podcast Summary: A16Z Podcast**
**Host:** Joe Schmidt
**Guest:** Dharmesh Shah (Co-founder and CTO of HubSpot)
**Publication Date:** 2026-11-10
### **Market Context and the Builder's Mindset**
* **Best of Times, Worst of Times:** Running a successful public company during this transformation brings a revitalized mission to serve customers with AI, similar to how HubSpot was founded 20 years ago to help SMBs leverage the internet.
* **The "Yellow Brick Road" and Frontier Labs:** Dharmesh Shah agrees with Joe Schmidt’s essay about building software outside the path that Frontier AI labs are racing down. Startups and established companies should stay off the "yellow brick road" of direct frontier model competition unless they have structural advantages, finding unique pockets of opportunity instead.
* **Abstraction and Speed:** AI acts as a new tool for higher-level abstraction, letting builders iterate and learn much faster than before, significantly shortening the loop and reducing the cost of failure.
* **Avoid Competing with Sam Altman:** Dharmesh jokingly cites one of his business rules as "never compete with Sam Altman," though he notes that open-source models, commercial alternatives, and open router frameworks mean builders no longer rely on a single company for essential AI ingredients.
### **HubSpot’s History, Evolution, and Ecosystem**
* **Origins and the SMB Wedge:** HubSpot started out of business school in the Boston area. Originally, they targeted ERP software for law firms (briefly called "LegalSpot" as an aperture-narrowing exercise), but learned that law firms were difficult to sell to and cared more about marketing. They pivoted to SMB marketing software because it combines the charging power of enterprise with the scale of consumer.
* **The Platform and CRM Journey:** HubSpot started as a suite of marketing applications (content management, SEO, landing pages, forms) to solve the core problem of fragmentation. Later, they discovered that 60% of their marketing customers were not using any CRM at all because existing options were too intimidating and expensive. To capture that segment, they launched a free, easy-to-use CRM, which eventually evolved into a platform.
* **Ecosystem Philosophy:** Influenced by historical platform/ecosystem missteps (such as Twitter closing down its developer APIs over time), HubSpot treats its developer ecosystem like a product, aiming to reduce "time to joy" and treating partners openly and fairly.
### **AI, Interfaces, and Architecture**
* **The Interface vs. the Value:** The interface is a means to an end. Business applications sit closer to Uber Eats/DoorDash (where the underlying work and network matter more than the raw app UI) rather than products where the interface is 99% of the value.
* **New OS / Browsers:** Chat GPT and Cloud Cowork effectively function as the new operating systems or browsers through which users consume value.
* **Determinism vs. Inference:** Dharmesh states that anything that can be solved with a structured query deterministically should be solved that way, rather than throwing a 300 IQ model at it via inference.
* **AI Architecture & Codebases:** With agentic coding, incumbents can leverage past data and even play back alternative historical universes of code to see what different architectural primitives would have yielded. HubSpot uses structured data stores alongside unstructured data contexts (like a context graph/second brain) and runs automated "evals" to test new models.
### **HubSpot Next and "YouSpot"**
* **HubSpot Next:** An internal innovation lab and "startup within a startup" designed to pursue new product categories, markets, or technologies without the baggage of the core organization.
* **YouSpot:** A project launched exactly 30 days prior to recording that targets one-person businesses (solopreneurs) with a CRM designed specifically for them. It tests early monetization, consumption-based pricing, and a second-brain/graph model architecture.
* **Acquiring Success:** Successful internal projects can either remain standalone, be absorbed into the HubSpot mothership via an internal "acqui-hire," or contribute core architectural learnings back to the main platform.
### **The Current State of Software and Entrepreneurship**
* **Company Formations:** In 2025, the U.S. saw a record-shattering 5.6 to 6 million new company formations, with another record pace tracking through August.
* **Solopreneurs:** Over 70% of new company formations are predicted to remain one-person companies (not hiring employees within the first two years), signaling a massive, growing market for solopreneur-focused software.
* **The SaaS Apocalypse Counter-Thesis:** While some traditional software companies will die if they fail to adapt, the overall opportunity set for software is larger than ever because AI enables solutions for problems that software was never previously equipped to handle.
* **Call to Action:** Dharmesh encourages high-agency, entrepreneurial builders to dive in, noting that it is the most fun and opportune time to build software in commercial history.
### **Lightning Round Answers**
* **Coffee preference:** Dunkin (though Dharmesh doesn't drink coffee or caffeine and drinks only water).
* **Shopping preference:** Instinct.
* **Animal to be:** A mammal, specifically an elephant.
* **Most Boston thing about HubSpot:** Pragmatism, solving for the customer, and balancing East Coast grind/pragmatism with West Coast ambition.
* **Best day ever at HubSpot:** About a week prior to recording, as new ideas and projects came together and unraveled clearly.