Here’s what’s on our plate today:

• 🧪 OpenAI’s ChatGPT Health and the vertical AI land grab.

• 🧠 Private LLMs, Trump’s AI voice, hurricane robots.

• 📊 Reader poll on who should own health AI.

• ✏️ Prompt of the Day for founders picking verticals.

Let’s dive in. No floaties needed…

The Laboratory

How OpenAI is setting the stage for the vertical AI race

It can often take years for a technology to exert its true influence on the lives of its consumers. Take the example of automobiles. The earliest automobiles were powered by steam engines. And though the steam engine was never truly successful, it did instigate the idea for internal combustion engines that made the automobile industry flourish throughout the last century.

In 2023, OpenAI introduced a similar idea into public consciousness. The idea: artificial intelligence models can be used to power conversation chatbots, generate content, all while emulating human speech and text. The development set off a race to develop artificial general intelligence (AGI). However, while companies competed to make their models bigger and better, OpenAI has quietly been focusing on another key aspect that could transform AI into one of the most consequential technological developments of all time.

The company recently announced the launch of ChatGPT Health, representing a pivotal moment not just for the company, but for the entire AI industry. By creating a dedicated health experience within ChatGPT that connects to medical records and wellness apps, OpenAI has demonstrated a playbook that AI companies across the industry are likely to follow: moving from general-purpose chatbots into domain-specific verticals with deep data integrations.

The announcement

OpenAI, in a blog post, announced the health features for ChatGPT, allowing users to connect medical records and popular health and fitness apps to have more personalized health-related conversations with the AI chatbot.

The feature lets users upload lab results, visit summaries, and clinical histories, and link apps such as Apple Health, MyFitnessPal, Weight Watchers, and Peloton. ChatGPT can then summarize reports, track trends like cholesterol level,s and help users prepare for medical appointments.

OpenAI said users can control how ChatGPT uses their health information through dedicated settings. The move expands ChatGPT into health-focused use cases, bringing the chatbot closer to sensitive areas where data privacy and regulation are likely to draw close attention.

The motive for launching the vertical is not difficult to fathom. Over 230 million people ask health and wellness questions on ChatGPT every week. That organic behavior, millions uploading lab results, asking about medications, decoding insurance jargon, created the business case for a dedicated product. OpenAI didn't create demand; it formalized what was already happening.

However, this also means that the playbook is now visible to the entire industry. First, observe where users are already applying your general-purpose AI. Second, build a domain-specific evaluation infrastructure (OpenAI's HealthBench, developed with 262 physicians). Third, hire executives with deep industry expertise. Fourth, partner for data connectivity rather than building integrations from scratch. Finally, create architectural separation with enhanced privacy protections for sensitive data.

This template applies far beyond healthcare. Finance, legal services, education, and insurance all exhibit the same characteristics: complex information scattered across multiple systems, frustrated consumers struggling to navigate bureaucracy, and professionals drowning in administrative work. Any AI company watching OpenAI's health launch should be asking: what's our vertical play?

The market opportunity nobody can ignore

The financial stakes explain the urgency. Vertical AI solutions captured $3.5 billion in 2025, nearly 3x the $1.2 billion invested in 2024. Menlo Ventures Healthcare led this surge, capturing nearly half of all vertical AI spending, more than the next four verticals combined.

The broader healthcare AI market tells an even more dramatic story. The market was valued at $26.57 billion in 2024 and is projected to reach $505.59 billion by 2033. Grand View Research: That's not a market opportunity; it's a gold rush.

Medical documentation and back-office revenue cycle management combined account for 60% of healthcare IT spend, creating a $38 billion opportunity. These are workflows that existed before AI but were never automatable: paperwork, billing codes, and prior authorizations. AI changes what's possible, and whoever captures these workflows captures durable revenue.

The investment community has noticed. Healthcare AI spending hit $1.4 billion in 2025, nearly tripling 2024's investment. Menlo Ventures' research found that healthcare AI has produced more unicorns than any other vertical AI segment, eight companies valued above $1 billion, and many more approaching that threshold.

The competitive scramble

OpenAI may have moved first, but its early lead in consumer health AI is unlikely to last. Rivals already have much of the infrastructure needed to respond quickly.

Google is well-positioned. In October 2025, it partnered with b.well, a health data platform that already connects more than 2.2 million healthcare providers and also underpins the medical records feature in ChatGPT Health. That means Google’s Gemini does not need to build health data pipelines from scratch. The plumbing is already in place, making it easier for Google to roll out consumer-facing health tools if it chooses to do so.

Microsoft is taking a different route. Through Dragon Copilot and its partnership with electronic health records giant Epic, Microsoft is focused on doctors rather than patients. Its AI tools are designed to streamline clinical workflows inside hospitals and clinics. Over time, that provider-centric position could allow Microsoft to extend its AI capabilities to patients without competing head-on with ChatGPT in the consumer market.

Anthropic is the wildcard. Claude’s emphasis on safety and strong guardrails makes it attractive for high-risk sectors like healthcare, where mistakes can have serious consequences. While Anthropic has so far focused on enterprise and professional users, healthcare may be difficult to ignore, especially as demand grows for AI systems that prioritize reliability over speed or scale.

However, it is not just big-ticket companies that are eyeing the healthcare industry. Startups, too, will look to capitalize on the shift, even though, for them, the stakes will be very different.

What it means for startups

OpenAI’s push into healthcare creates a mixed picture for startups. On one hand, it opens the door to partnerships. OpenAI is not building everything itself. It relies on companies like b.well for medical records and on platforms such as Apple Health and MyFitnessPal for fitness and nutrition data. Startups with proprietary datasets, niche workflows, or deep clinical expertise could plug into these platforms rather than compete with them.

At the same time, the competitive pressure is real. Frontier AI labs no longer appear content to simply provide models and let others build applications. They are increasingly shipping full, production-ready tools in sensitive domains like healthcare, law, and finance. That shift puts startups at risk of being crowded out.

The path to survival lies in specialization. In healthcare, AI may be the engine, but domain expertise is the moat. Startups that go deep into specific workflows, such as prior authorization, clinical documentation, or medication management, can build defensibility that broad platforms struggle to replicate.

Investors are already leaning this way. Retention tends to be strongest where software becomes core infrastructure rather than a narrow feature. Companies that sit at the center of workflows, acting as systems of record or orchestration layers, are more likely to endure as platforms expand around them.

For larger companies, the time is also ripe. OpenAI’s health launch is how well it aligns with the current regulatory mood. For now, the environment is unusually supportive. U.S. states such as Utah have approved AI-assisted prescription renewals, while the FDA has signaled it will take a relatively light-touch approach to regulating wellness software and wearable health tools. That creates room for experimentation.

OpenAI is also trying to shape the rules itself. The company has published policy proposals calling for clearer regulatory pathways for consumer health AI and even floated the idea of broader access to global medical data. A more detailed health AI policy framework is expected soon.

This timing matters. Companies that launch early can demonstrate value before stricter rules arrive and, in doing so, influence how those rules are written. First movers do not just win users; they help set the regulatory baseline.

More broadly, ChatGPT Health reflects a shift across AI. The industry is moving away from general-purpose chatbots toward vertical, high-stakes applications.

The bigger shift

Healthcare follows education and shopping as OpenAI bets on becoming the interface between users and entire industries. With enterprise spending already shifting toward AI applications, the advantage now lies with companies that go deep into specific domains, secure data partnerships, and embed themselves so tightly into workflows that switching becomes difficult.

OpenAI has already shown it is capable of dominating the market for developing the underlying models upon which AI applications are built. Now, the company seems poised to establish its dominance in the field of AI-powered applications rather than allow other companies take advantage of the underlying tech to build their tools.

Prompt Of The Day

Bite-Sized Brains

• Private-by-default ChatGPT rival: Signal founder Moxie Marlinspike’s new app, Confer, runs models in secure hardware, encrypts chats end-to-end, and promises no data harvesting, logging, or ad training.

• Trump’s AI voice sells housing: A new Fannie Mae ad uses an AI-generated clone of Donald Trump’s voice (with his permission), spotlighting how deepfake-style tech is creeping into government-adjacent messaging.

• Ocean robot surfs a Cat 5: Oshen’s autonomous “Oshen S” surface robot rode out Category 5 Hurricane Beryl while streaming real-time wave and wind data to improve future storm forecasting.

Tuesday Poll

Rate This Edition