Macroeconomics From Outer Space: Why Satellite Imagery Now Beats the Fed
Last week I sat down with Barry Moline on The Great Ideas Podcast to discuss something that, on the surface, sounds like science fiction: predicting GDP using satellite imagery.
To my pleasant surprise, the conversation pushed into exactly the themes I think matter most—how the world measures economic activity, why our current tools are structurally broken, and what happens when forecasting leaves the basement and enters orbit.
Below is a distilled version of the conversation and its implications for markets, businesses, and anyone making decisions in a volatile macro environment.
1. The Core Question: What If We Could Literally See the Economy Instead of Merely Inferring It?
Traditional macroeconomic forecasting is slow, noisy, and heavily anecdotal. GDP arrives with a three-month lag. Inflation numbers bounce with revisions. The Federal Reserve still uses the Beige Book—essentially a nationwide “how are things going?” phone tree.
Meanwhile, satellites orbit Earth every 90 minutes, revisiting the same coordinates every ~5 days, capturing unbiased physical evidence of economic expansion.
The founding insight behind Atlas Analytics is simple:
During my time at Harvard and Oxford—studying economics by day and remote sensing/computer vision by night—I built a patent-pending algorithm, ROY (Remote Orbital Yield), that extracts the expansion of the built environment from satellite imagery.
Highways, factories, rooftops, logistics hubs, commercial floorspace, ports—the literal physical footprint of economic activity.
And it turns out we see something quite powerful:
The expansion of the built environment is one of the best leading indicators of GDP growth we’ve ever measured.
2. Why ROY Works: Seven of the Top Ten Predictors Are From Space
One of Barry’s questions that stood out in particular was: “How does AI weigh satellite data versus traditional indicators?”
Here’s the short answer:
Seven of the top ten variables in ROY’s machine-learning model come from satellite imagery—not lagged GDP, not CPI, not housing starts.
That’s a staggering result for the field of macroeconomics, and it explains why ROY consistently outperforms the Federal Reserve and major Wall Street consensus forecasts.
The underlying reason is structural: economic activity happens in areas of agglomeration—dense clusters where 90% of the world’s economic output takes place on just 16% of its land1. Satellite imagery sees those expansions the moment they occur.
Governments, by contrast, wait three months… then revise… then revise again.
3. Why Weekly Data Changes the Game
As mentioned in the episode, the U.S. government uses a highly manual, antiquated process to produce GDP four times a year. The data is pulled from a myriad of sources: surveys, census data, panels, sector reports, state records, conference calls, and more. The process is incredibly labor intensive, and the estimates are produced with significant, inconsistent lag, and several revisions.
Atlas provides our clients with 50+ points on a weekly basis.
By running ROY weekly, we convert a lumpy, outdated quarterly metric into a dynamic signal that reveals:
Micro-movements of the business cycle
Inflection points before they appear in official data
Early warnings of recessions
Evidence of overheating or overvaluation in specific macro-exposed financial indices
While government statistics arrive with long lags and revisions, Atlas models detect economic turning points as they happen.
4. Yes, Market Timing Is Now Possible (Sorry Warren Buffett)
One of the bigger reactions during the interview came when I said something that would get me thrown out of some economics departments:
Market timing is no longer impossible. It is now quantifiable.
Using ROY signals, Atlas Capital—my own portfolio where I trade exclusively on Atlas data—has returned roughly 60% year-to-date (~54% to be exact).
We don’t pick individual stocks. We avoid idiosyncratic, equity-specific risk. Instead, we trade macro-exposed ETFs whose performance is tightly linked to national economic cycles.
If ROY says tech (XLK) is fundamentally overpriced relative to GDP conditions, we act accordingly.
If ROY says we’re entering a pullback, we reposition before the market catches up.
It’s macroeconomics with a telescope instead of a rear-view mirror.
5. Applications Go Far Beyond Finance
Yes, hedge funds and family offices are our primary clients today—but this technology applies to:
Corporations deciding factory output, capital expenditures, or hiring
Local governments validating growth or decline without waiting for state agencies
Real estate investors identifying states or metros that are accelerating
Supply chain operators monitoring industrial expansions
Policymakers assessing regional development in real time
My favorite metaphor from the interview:
Whether you pay attention to the weather or not, it will still rain on you.
Atlas is simply offering you the umbrella first.
6. “Aren’t Satellite Images Public? Can Others Copy You?”
Yes, the imagery is public.
No, the algorithm is not.
Atlas filed patents in 2021 covering the ROY methodology—band math, feature extraction, spectral analysis, computer-vision architecture, and more patents are on the way.
As I told Barry:
Every Coke needs a Pepsi. But Pepsi still can’t crack Coke’s formula.
Competition will come. It should—this is a new field.
But Atlas has opened the frontier.
7. For Those Wondering About an Atlas Investment Fund…
People often ask whether we plan to launch an Atlas fund. While we sometimes use ‘Atlas Capital’ as a purely illustrative, non-commercial example, there is no fund today—and we do not trade on private signals that we don’t share with our clients. Instead, we use our proprietary signals combined with our economic and financial intuition to time the market.
If or when we explore a formal investment vehicle, it would be fully separate from our analytics business, and our readers would be the first to hear about it.
8. Want to Follow the Data?
We publish:
Quarterly GDP predictions (ahead of the BEA) — Free
Weekly macro-market commentary — Premium
Video explainers & deep dives — YouTube
Custom consulting for corporations & funds — White glove
Follow along with us on socials:
A Closing Thought
Macroeconomics is entering its “weather forecast moment.” For the first time, we can measure—not guess—the state of the economy in near-real time.
Satellite-powered forecasting is not the future.
It is the present.
And for those paying attention, it’s a serious advantage.
This article was written with valuable research assistance from Morgan Reppert, Executive Operations Associate for Atlas Analytics.