Billionaire founder who scaled up on data

While the rest of Silicon Valley were chasing viral chatbots and AI generative models, teenager Alexndr Wang saw a bottleneck: AI models needed better data to get smarter but that input was getting scarce. So, in 2016 he founded Scale AI, a startup that built the data supply chain for the AI world.

(Yes, his name is written as Alexandr, without the letter e. Some sites attribute this to numerology as eight is a number that signifies wealth in Chinese language.)

Now, Meta has paid a staggering 14.3 billion US dollars for a 49 percent stake in the startup. And 28-year-old Wang is to lead Meta’s “Superintelligence Lab”– a team of 50 handpicked by Mark Zuckerberg and tasked with building AI that’s smarter than humans.

“This is a very expensive acquihire of Alexandr Wang,” wrote tech analyst Ben Thompson, founder of Stratech newsletter. Talent grabs are nothing new in Silicon Valley. Microsoft dropped 650 million US dollars on Inflection in 2023 – officially a “licensing deal”, but everyone knew they were really buying Mustafa Suleyman and his crew. Google followed suit, throwing 2.7 billion US dollars at Character.AI to snag founder Noam Shazeer.

But Meta’s move? This is different. Zuckerberg isn’t just hunting for another coding wizard. He’s going after something way more fundamental – something that every AI company is going to scramble for soon: quality data.

Boring, but Vital

AI models are only as good as what you feed them and they are running out of good data. Whatever there was on the internet has been devoured and new data is needed desperately. That means humans have to sit there, laboriously label massive datasets. It’s boring, it’s expensive, but it’s absolutely critical.

Large Language Models (LLMs) need a lot of data to train and retrain. When it all started, they were fed what was available, but as their capacity grew, the data had to be more nuanced. Now it is like teaching a post graduate level candidate who needs more quality input than what a school student needs.

While the tech world was chasing flashy consumer apps and sexy AI products, Wang saw more value in the data pipeline. Now Scale AI works with PhDs, historians, software engineers and global gig workers across the world, including India, to annotate some of the most complex datasets in the world. One annotation could cost as much as 100 US dollars.

His business acumen proved right and now his startup has contracts with major tech companies and government agencies, including the US Department of Defence. Scale AI’s services became essential in training models for self-driving cars, facial recognition, logistics, and even language processing tools like ChatGPT.

Los Alamos Kid

Born in New Mexico to Chinese immigrant parents, Wang ticked all the boxes even in his school days as he made it to the Math Olympiad, US Physics Team, US Computing Olympiad. (A bit of an irony here, given Trump administration’s threat to target Chinese students in the US.) He started coding at Quora while still in high school.

After high school, he moved to Silicon Valley to work at the investment firm Addepar. He briefly attended the Massachusetts Institute of Technology and had a stint as an algorithm developer at the high-frequency trading firm Hudson River Trading before he dropped out to co-found Scale AI in 2016 with Lucy Gao. Scale AI then made it to that year’s Summer Cohort of Y Combinator.

“The great thing about Scale is that it is taking advantage of an opportunity that is hiding in plain sight,” said Y Combinator director Jared Friedman in a blog that year.

The Y Combinator stint helped Wang develop powerful connections and during the height of the Covid-19 pandemic, Wang was roommates with his friend Sam Altman, the chief executive officer of OpenAI.

Scale AI’s business model was brilliantly simple: AI companies need clean, labelled data, and Wang’s startup supplied it through platforms like Remotasks and Outlier. Scale AI basically became the matchmaker between tech giants and thousands of gig workers worldwide, mostly from the Philippines.

Lucy Guo, Wang’s co-founder – another child of Chinese immigrants who taught herself to code as a teenager – left the firm in 2018 as she had “differences in product vision”. But she kept her equity stake that is now worth 1.25 billion US dollars. That made her the youngest self-made female billionaire at 30, snatching that title from Taylor Swift.

Reflecting on those early days, Wang admitted on a South Park Commons podcast: “The first half of our batch…was probably … [what] we refer to as a squiggle – a lot of existential angst; you don’t really know what you’re doing with your life.”

But that uncertainty didn’t last long. Scale AI started landing contracts with everyone who mattered. Want to train self-driving cars? Scale AI. Need facial recognition? Scale AI. Building the next ChatGPT? You guessed it – Scale AI.

By 2019, Peter Thiel’s Founders Fund was throwing 100 million US dollars at Scale AI to help it hit unicorn status. In just five years, Wang had gone from MIT dropout to the youngest self-made billionaire on the planet and now his worth is estimated at 3.6 billion US dollars.

The Dark Side

But it was not a squeaky-clean journey all the way. A Washington Post investigation threw some serious shade on Scale AI’s operations, especially in the Philippines. The report said thousands of workers on Remotasks platform were getting paid peanuts, with the company regularly delaying payments and offering basically zero recourse when things went wrong.

Labour rights groups weren’t having it, calling out Scale AI and other American AI companies for treating overseas workers like garbage. Something many Indian workers of Silicon Valley-based firms also allege.

Yet, as the startup’s business grew, so did the profile of Wang. He became a voice that carried weight and had testified before Congress about how Chinese AI companies are catching up with the American ones. Scale AI even took out a full-page Washington Post ad with an open letter to Trump saying “America must win the A.I. War.”

Wang is apparently not a fan of Beijing and in a 2024 interview to ChinaTalk was very vocal about the threat China poses to the US interests and warning about the capability that Beijing is achieving in both AI software and hardware. In a revealing admission, he also says his parents “hate the CCP”, which is about the only bit of information that is readily available about his parents. 

The Meta Mega-Deal

Wang’s new role at Meta gives him a more prominent role in shaping the AI sector in the US and he has clear vision of how the country should put a data moat to defend against adversaries like China.

“You can see this with hypersonic missiles, as well as a lot of developments in space technology – we’ve waited for China to catch up a little bit before reigniting the flames. I have a real concern that it will take a crisis to make us take this stuff seriously,” he told ChinaTalk.

“If you believe we are on this fast take-off scenario and that we’re imminently close to AGI – then you need to prevent all of our secrets from going over to our adversaries and you need to lock down the labs.”

Opinions of influential Tech Bros now carry a lot of weight in Washington though the effect of the fallout between President Donald Trump and Tesla chief Elon Musk could alter that. The Meta deal could also see Wang join the elite group and it will be interesting to see how he fares, especially given the US-China tech duel. 

Alarm Bells

But the Meta-Scale deal has already started sending ripples through the AI world. Google, which was Scale’s biggest customer, is basically saying they are looking for alternatives. That’s opened the floodgates for Scale AI rivals.

“The Meta-Scale deal marks a turning point,” says Jonathan Siddharth from Turing, another data supply firm. “Leading AI labs are realising neutrality is no longer optional, it’s essential.”

Labelbox CEO Manu Sharma is practically rubbing his hands together, telling Reuters they’ll “probably generate hundreds of millions of new revenue” from customers fleeing Scale AI. Over at Handshake, CEO Garrett Lord is watching demand explode: “Our demand has tripled overnight after the news.”

The drive for new data also puts the focus on larger countries like India and China as larger population generates more data. But moulding it into quality data needs a lot of work. While China has been putting pieces together methodically both in terms of energy and data, India is yet to make a concerted effort on these fronts though there have been attempts to scale up computing power.

Like they say data is the new oil and India’s data-rich environment is a goldmine – but without structure and standards, it remains crude oil waiting for refinement.

Wang faced the same challenge in the early days. He told Business Insider that many investors rejected him because he wasn’t pitching a product filled with the buzzwords they wanted to hear. But the principle he followed can serve as a guiding light for any startup: “Focus on building the business, and the rest will kind of take care of itself.”

AI is on everyone’s mind

Our last newsletter on the impact of Artificial Intelligence on the job market drew some interesting responses – from those worried about getting sacked soon to others who dismissed it all as hype. But it’s heartening to see that AI is on everyone’s radar, whether you believe in it or not. Even more encouraging is the Kerala government’s growing interest in using AI to make governance more efficient and tapping local talent for solutions. Last week, the Kerala State IT Mission, in collaboration with the Kerala Startup Mission (KSUM) and the ICT Academy of Kerala, hosted K-AI: The Guiding Hand of Smart Governance at Trivandrum Technopark. Startups and various government departments connected at the event – and hopefully, we’ll see some results soon.

Another Technopark event was a Generative AI Hackathon 2025 focused on how the technology can transform public administration across key sectors. It was organised by the Boston Institute of Analytics (BIA) in association with Prathidhwani, the welfare and recreation forum of IT employees at Technopark. More than 200 students from institutions across Kerala took part. But some habits die hard. The press note issued by Technopark had no details about the winners or their innovations – but made sure the names of the functionaries who attended were there. Thank God for small mercies, eh?

Chinese water torture continues

Another month, another reminder of China’s grip on the global supply chain. Ever since Beijing imposed restrictions on the export of rare metals, electric car manufacturers have been in panic mode. Carmakers in Europe and the US have sounded the alarm, while in Japan, Suzuki has already slowed EV production due to the scarcity of a specific magnet – a cheap component but crucial for energy conversion. Indian firms making electric two-wheelers are also in trouble, as they are heavily dependent on Chinese suppliers for materials. In a detailed Economic Times report, former CMD of Scooters India Ltd, Arunaditya Sahay, points out that while India has the fifth-largest reserves of rare earth elements, it lacks the advanced extraction and refining capabilities.

Own a slice in Dubai

Buying an apartment is an expensive affair – even in second-tier cities like Trivandrum or Cochin, let alone in Dubai or Singapore. But Dubai is testing a new model where a group of people can co-own a property. Through a tokenised real estate project, individual apartments are split into slices and sold as blockchain-backed tokens for as little as 2,000 dirhams (47,000 rupees). Investors earn returns through rent or resale. The Dubai Land Department has already sold two such apartments in high-end neighbourhoods. One listing drew over 10,700 investors and was eventually sold to 149 buyers from 35 countries, according to the news agency WAM. Could this model catch on globally?

Pizzas are the new tea leaves

What do military leaders do when war looms? Apparently, they order pizza. According to a Futurism report, pizza orders from the Pentagon spiked just before Israel launched its attack on Iran. The piece quotes Alex Selby-Boothroyd, head of data journalism at The Economist who says the “Pentagon Pizza Index” has been a surprisingly accurate predictor of global flashpoints – from coups to wars – since the 1980s. “On the night of August 1st 1990, for example, the CIA ordered 21 pizzas in a single night just before the Iraqi invasion of Kuwait (a new record),” he added. “Who says pie charts aren't useful?”Tea leaf reading is officially outdated now.