# Inside Strava’s New HQ: A Runner’s Paradise and the Future of AI in Fitness
Strava, the leading fitness application globally, has revealed its latest headquarters in San Francisco, marking the beginning of a fresh chapter in innovation and expansion. With a user base of 135 million, 51 million recorded activities weekly, and over 1 million community groups, Strava transcends simple fitness tracking—it’s a worldwide phenomenon.
During my visit, I got a glimpse of Strava’s new “nerd cave” for runners, met with San Francisco Mayor Daniel Lurie, and conversed with Strava leaders about their cutting-edge AI project, Athlete Intelligence. Here’s what I learned regarding Strava’s direction and the evolution of its AI beyond traditional large language models (LLMs).
—
## **A Tour of Strava’s New HQ: A Celebration of Fitness**
Strava’s contemporary multi-story headquarters is crafted to immerse employees in the fitness culture. Meeting spaces are named after legendary running and cycling routes, such as Central Park and Alpe d’Huez, while a dynamic Strava Photo Map features stunning visuals from users around the globe.
One standout aspect was a real-time, chronological map of San Francisco that showcased live activity movements from local Strava users. Seeing the city animated with physical activity was enchanting, and I proposed to Strava’s leaders that they think about extending this feature to other major cities worldwide.
The office also boasts a library stocked with running literature, walls decorated with GPS art made by users, and displays of employees’ race bibs—transforming it into a fantasy workplace for any fitness aficionado.
—
## **Athlete Intelligence: Strava’s AI Evolution**
Strava’s Athlete Intelligence has recently transitioned out of beta, with over 80% of users deeming it “very helpful.” What differentiates it from other AI-driven fitness applications?
### **Beyond Basic LLMs**
CTO Rob Terrell highlighted that Athlete Intelligence is not merely another chatbot. “It was crucial to me that Strava achieves something far more substantial than the low-effort LLMs that merely repeat Google search results,” he shared.
Instead, Strava’s AI integrates comprehensive user data to generate deeper insights. For instance, it assists novices in understanding heart rate zones and pace fluctuations without needing to manually evaluate past workouts.
### **Expanding to More Activities**
Currently, Athlete Intelligence predominantly targets running, hiking, walking, and cycling. When inquired about branching out to other sports, CPO Matt Salazar stated that Strava aims to guarantee high-quality insights before initiating AI support for more activities.
### **From Reactive to Proactive AI**
At present, Athlete Intelligence assesses previous workouts, but is there potential for it to offer proactive training suggestions? Salazar indicated that Strava is probing methods to aid users in reaching their fitness aspirations beyond just earning challenge badges and kudos.
Terrell acknowledged that Strava has traditionally been a “backward-looking product” but expressed a desire to pivot towards more forward-thinking solutions. Although no formal announcements were made, a more anticipative AI could revolutionize the experience for both novices and dedicated athletes.
—
## **Strava’s GPS Accuracy Dilemma**
One of the most fervent discussions I had with Strava’s leaders centered on GPS accuracy—specifically, why Strava refrains from “correcting” race distances when watches fail to capture the entire course.
Numerous users have voiced concerns about missing personal records due to GPS inaccuracies, but Terrell explained that Strava’s hands are tied by the precision of users’ devices. “We strive to be as accurate as possible rather than manipulate the data,” he stated.
With 72% of Strava activities logged on smartphones—often less precise than GPS watches—Strava faces challenges in aligning accuracy with user expectations. While they are investigating potential remedies, a simple solution remains elusive.
—
## **Could Strava Introduce Race Leaderboards?**
Considering that over half of the participants in major marathons like NYC, Boston, and London use Strava to log their races, I inquired whether they might contemplate establishing official race leaderboards.
Salazar cautiously replied, “We can’t comment on a future roadmap, but… that use case holds significant importance for us.” This implies that Strava is at least mulling over ways to enhance the race day experience and assist users in recognizing their accomplishments.
A specialized race leaderboard could enable runners to compare their times against their past performances and fellow competitors, making race day even more engaging.
—
## **What’s Next for Strava?**
Strava has already rolled out several significant updates in 2024, including:
– **Athlete Intelligence** for AI-centered insights
– **Weekly and night heatmaps**
– **Dark mode**
– **Tighter integrations with Oura and Apple Fitness Plus**
– **3D FATMAP terrain data** aimed at winter sports aficionados
While Strava executives didn’t disclose particular future strategies,
