What If AI Competed on Health Outcomes?

Why the age of superintelligence demands new benchmarks for human health

In my recent essay, Health Moonshots in the Age of Superintelligence, I argued that the innovation landscape has crossed a meaningful threshold. As we begin 2026, that reality feels unmistakable. A global ecosystem of founders, funders, and researchers is now focused on transforming human health — and the tools at their disposal have never been more powerful. AI, compute, data, and scientific innovation at unprecedented scale are turning health moonshots from distant ambition into achievable outcomes.

This essay is a follow-on extension of that belief — and the topic of conversation we’ll continue next week at APOLLO HOUSE, the summit we’re convening during JPM Healthcare Week in San Francisco. There, we’ll gather leaders from across technology, healthcare, and investment to explore a central question: how can this moment be leveraged to deliver meaningful progress for humanity?

I believe the answer lies in refocusing ‘the race’ — by defining a new Health Moonshots Race built around clear, health-specific benchmarks and real-world results that impact people’s wellbeing.

If the many forces driving the current AI frenzy want to keep the boom moving up and to the right, they will need to earn broad trust and deliver on a more definable (and consequential) promise for humanity. They can do so by competing on something that matters to everyone: human health and wellbeing.

We are living through a dynamic moment of acceleration. Capabilities that once took decades to develop are emerging in months. Investment, talent, and attention are pouring into artificial intelligence. And yet, as powerful as today’s AI systems have become, the way the industry benchmarks “winning” is still poorly aligned with humanity’s most urgent needs.

That misalignment matters.

If we are truly entering an age of superintelligence and abundance, then what we choose to measure, reward, and celebrate will shape where these tools are ultimately applied. Today’s race largely rewards performance on academic, synthetic, and technical benchmarks — useful tests of capability, but disconnected from many of the problems most people actually face. If we want to understand what these systems are truly capable of, we should challenge them (and the people creating them) to solve more of the problems that affect every human on earth.

This is the moment to change the game.

Healthcare already measures a great deal. Clinical trials track survival rates and biomarkers. Public health relies on population-level indicators. Health systems manage cost, efficiency, and access. Researchers benchmark molecular and diagnostic performance every day.

But these measures were designed for healthcare — and for a different era — before AI at scale, frontier compute, and a global technology race fueled by unprecedented capital and talent. If we want faster health transformation, we need new frameworks that pull the world’s most competitive innovators (from all industries, especially tech) into the health arena and “reward” them for winning there.

What we lack is an AI-first, outcome-driven race — one that moves beyond the boundaries of healthcare and unites the broader technology ecosystem around shared health goals, much like the space race once united governments, scientists, and innovators of every stripe.

Today, the global AI race is defined by speed, accuracy, scale, and performance. These benchmarks have accelerated progress, and they deserve credit. But they leave unanswered the question people increasingly ask: How does all of this help my family, my community, my future?

The answer won’t come from better test scores. It will come from visible, real-world progress on issues that shape people’s health and wellbeing — progress that families can recognize in their own lives.

We need a new class of benchmarks — impact-driven challenges that align the world’s most innovative people, tools and resources toward solving problems that affect everyone.

Health Moonshots, and AI-driven evaluations, with clear and measurable outcomes.

One lesson emerging from today’s AI research is that progress accelerates when benchmarks are narrow, practical, and grounded in real-world outcomes. The most useful evaluations are no longer broad measures of “intelligence,” but focused tests tied to specific domains and problems. In health, that could mean benchmarks for early disease detection, trial design efficiency, diagnostic accuracy for particular conditions, or reductions in time from discovery to patient impact.

These kinds of evaluations don’t replace moonshots — they are the stepping stones to help make them achievable. Big breakthroughs are rarely won in a single leap. They are reached by stacking many focused wins, each measurable, each hard to game, and each tied to outcomes that matter. In an age of superintelligence, the path to curing disease may run through hundreds of smaller, well-defined races that compound over time.

The opportunity now is to design these benchmarks intentionally — so that the competitive energy of the technology world is directed not toward abstract scores, but toward measurable progress in human health.

A Call for Benchmarks That Matter Outside the Bubble

Imagine if, alongside weekly and quarterly AI performance updates, the global innovation ecosystem rallied around dozens (or even hundreds) of shared health goals such as:

• Progress in early detection and disease-modifying interventions for Alzheimer’s and related dementias

• Reductions in the burden and cost of diabetes and cardiometabolic disease

• Faster, more accurate diagnosis and treatment of major diseases

• Demonstrable impact from food-as-medicine approaches

• Measurable gains in healthy lifespan — not just years lived, but years lived well

• Real reductions in cost, waste, and friction across healthcare delivery

These are not abstract ideals. They are outcomes that can be measured, tracked, and improved over time. And for the first time, if we change the game — so that the most powerful organizations (and people running them) are competing to improve human health — progress will accelerate dramatically.

The next great race should be AI, capital, talent, and ingenuity aligned in service of human health.

A Health Moonshots Race — designed for the age of superintelligence — that encourages competition where it matters, collaboration where it helps, and faster translation from discovery to real-world impact. Not a one-off challenge, but an ongoing framework that makes progress visible in a way the whole world can get behind.

The benchmarks don’t need to be perfect on day one. But they do need to be clear, public, and continuously updated — so the world knows what it’s racing toward. If we design them to be as easy to check as the weather or your favorite sports team, than people outside the bubble will start to pay attention.

An Invitation to Change the Game

As we kick off 2026, let’s push on what we want all this intelligence to be used for.

Benchmarks shape behavior. They define ambition. They tell the world what winning means.

We now have tools — and a global community of health innovators — capable of changing the trajectory of human health at scale. The task before us is to aim these mighty resources deliberately, and to measure success by how much healthier our communities and world become.

P.S. You can apply to join our APOLLO HOUSE waitlist here. Or learn how to join StartUp Health’s network here.

P.P.S. Ping me if you have ideas for what these benchmarks could / should be! Let’s get going.