登录/注册
我们的世界很复杂. 没有简单的答案.
There is, however, magic (and power) in making the complex look simple.
We’re excited to announce 微米’s investment in Atlas AI, a spinoff from Stanford’s 可持续性 and Artificial Intelligence Lab simplifying land-based analytics and prediction from satellite imagery and other sensors.
现在, Atlas AI is using data and artificial intelligence to help some of the most important decision-makers in the world make important decisions.
But that’s not really why we invested in Atlas AI.
We invested because Atlas AI is the only company in the world with a scalable platform that delivers trusted value from satellite imagery and artificial intelligence.
好的团队，更好的时机
Professors Stefano Ermon, Marshall Burke and David Lobell are smart cookies. Burke and Lobell are known around the world as leaders in applying data science and machine learning to studying and predicting earth systems — weather patterns, 水资源, 作物产量. Ermon is at the epicenter of AI re搜索, specifically as it applies to real-world 应用程序 of value.
I first met Ermon as he was starting Atlas AI, and we talked about the challenges of growing a startup with an initial focus on solving problems in emerging markets — because it’s hard. How do you scale a Palo Alto, California, startup while selling to the poorest people in the world?
Someone like Rockefeller could pony up the dough to make a go of it, but to be scalable — to build technology that solves problems not just in emerging markets but in other global markets as well — Atlas AI would need to get mean, 无情甚至, and focus on winning everywhere it could. Atlas AI didn’t have that sense of ruthlessness. 而且它不需要钱. 所以美光没有投资.
All that changed when we met Victoria Coleman, Atlas AI’s first CEO. I knew right away we would invest.
科尔曼令人印象深刻-博士, 维基媒体基金会首席技术官, 在英特尔任职, 国防科学委员会, DARPA和洛克希德. But that doesn’t make you a good startup CEO. So, 我第一次见到维多利亚的时候, I came in hot: You don’t know what you’re doing, 影响力投资是一个神话, 你必须做得更多.
But here’s the thing about Victoria. 她听. 我们聊了聊. 我们制定了战略.
And when I came a-knocking, Atlas AI’s customers listened. 我们聊了聊. 我们制定了战略.
事实证明, customers have only recently discovered what it is they actually need — and Atlas AI has it.
在Atlas AI我们信任
Figure 1 (The Earth, by an Italian in 1450) and Figure 2 (The Earth, by a German in 1507)
In 1450, according to the internet, the 弗拉莫罗地图 (Figure 1) was one of the most accurate maps of the world (and difficult for us to read now because it’s upside down from the usual, 中国在左边).
仅仅42年后, in 1492, the Eurasia-African view of the world was forever altered when a Genoese sailor (Christopher Columbus) stumbled on the 新的世界. In 1507, 有人画了一张地图 (Figure 2) and called the new stuff “America.”
The 1450 map was valuable — people used it and navigators relied on it. But the “discovery” of the Americas, 新信息, immediately transformed the dynamics and power structure of all people on Earth.
The 1450 map wasn’t wrong, but it wasn’t complete. So too with analytics services based on satellite imagery. 是的, from satellite images alone we can extrapolate correlations on the ground — the number of oil drums, 人口:一个城市的人口, 房子的大小.
By leveraging ground-truth and transfer learning in combination with satellite imagery, Atlas AI has the only technology in the world able to accurately — and with unprecedented resolution — analyze and predict Earth’s surface as it actually works.
The company has started with Africa, satisfying much-needed demand for agriculture and infrastructure intelligence to both build and support a sustainable future.
Atlas AI is now entering new markets with customers needing that whole picture — a complete, reliable and accurate picture — to help them make better decisions.
这就是我们投资的原因.
