Caesar Sengupta and his team have built a “digital private bank” called Arta Finance, but it won’t compete with other private banks or the advisors who work there.
Rather, he wants to widen the gates for investors who could benefit from the types of investment opportunities available at those firms—think private placements, structured products, various types of alternatives—but don’t have a high enough net worth to be a traditional client.
We’ve seen plenty of digital investment platforms come and go, and the “democratization of alternatives” is well-worn marketing copy by now. Firms like CAIS and iCapital are becoming deeply embedded in independent wealth management shops, and a variety of other technology platforms are promising to bridge financial planning and alternative investments. But Sengupta has big ambitions for his project, and I think Arta Finance is an endeavor worth keeping an eye on.
As I drank my morning coffee, I thanked him for introducing me to the platform at a time when many Singaporeans were more likely enjoying the nightlife along Clarke Quay. He said he did not mind talking about his baby whatever the time of day.
“Millennial millionaires are probably our sweet spot,” said Sengupta, the firm’s co-founder and CEO. For now, the firm’s clients are accredited investors, many of them in their 20s and 30s, making at least $200,000 annually and with net worths of at least $2 million. He’s opening the door for the so-called “HENRYs,” or the “high-earners-not-rich-yet” set.
While declining to share overall AUM or number of clients, he said the firm currently works with several thousand investors in the United States.
While Arta operates as an SEC-registered RIA, Sengupta said its offerings were more like those from UBS, Goldman Sachs or Morgan Stanley.
The origin story of Arta goes like this: He and his colleagues were wealthy millennials working in Silicon Valley at firms like Google on projects like Google Pay and Chromebook. With high salaries and potentially lucrative liquidity events on the horizon, they were in the position to invest for a long time horizon, yet they were frustrated with their experiences at traditional financial service firms.
“We were investors on our own,” he said. “We used some of the automated investing services and then started to use [traditional] private banks on our own to get access and invest in private equity, private real estate and other alternatives,” he said.
“Overall, it was a good experience but very expensive and you could only start using it in your 40s” because of the high minimum investment requirements. “What we are doing is bringing private banking to everyone, and doing it much more cost effectively,” he said.
It is often said that timing is everything. After 15 years at Google, he left to start what he called a “digital family office to the world” in a 2022 TechCrunch article. In that same article, he noted that Arta had raised more than $90 million in early rounds of funding from investors, including former Google CEO Eric Schmidt, Sequoia Capital India, Ribbit Capital and dozens of other individuals.
“We’re pretty well staffed here, and enough capital from Ribbit and others for now,” he said. His developer and technology team have grown to 45 people, based across both Silicon Valley and Singapore, a full half of his total employee count.
He laughed when I asked him if that time zone difference presented a challenge.
“The kind of AI talent available in the Valley has helped us get access to some of the best AI work going on in the Valley,” he said.
This brings me back to the element of good timing. The explosion in available AI tools and large language models is the jet fuel powering his team to rapidly build an investment platform sophisticated enough to give investors easy access to private placements, structured products and other alternatives, as well as direct indexing and tax-loss harvesting services.
Assets are custodied at BNY Pershing, and while much of the platform is automated—and it is technology that is driving the lowering of costs, minimums, and efficiency—clients can get time with a human advisor when needed, Sengupta said.
A “quant infrastructure” enables the management of the investments, with built-in direct indexing and tax-loss harvesting, across all, some or none of the portfolio, depending on the client’s choice, and a much lower minimum investment requirement than traditional funds available on larger established platforms.
“Let’s say Parametric has a $250,000 minimum. We have a $25,000 minimum,” Sengupta said.
The tax-loss harvesting and direct indexing capabilities were built in-house to meet client demand. “We got a lot of feedback from our members: ‘can I get the S&P 499 not 500, minus my company?’” he said. Many of the newly minted millionaires in Silicon Valley will have, by dint of their work, heavy allocations towards companies that are already, arguably, overweighted in the broader index, i.e., Meta, Google, Amazon, Tesla.
“Where it gets very interesting is our satellite portfolios where you can put any investment into an ETF-type structure,” he said.
Arta also sells private placement life insurance or PPLI, one of 50 or 60 RIAs that can do so in the U.S., he said.
Arta’s website is inviting, and the application’s frontend interface is beautiful. From a design perspective, it puts to shame most of the clunky, legacy wealth interfaces this technology journalist has seen over the years.
What is nearing completion but not yet available is Arta AI, which is supposed to roll out in Singapore this summer and later in the U.S., once the firm gets regulatory signoff.
At first, it will provide the Arta user with the ability to ask the AI to generate a portfolio recap, something a traditional human private banker generally does with clients quarterly.
In a demo version, Sengupta showed me the various styles of portfolio reporting a client could ask for, which included a “Wall Street” style, an “Easy Reading” style, with no financial jargon, and a “Yoda Mode,” which reads as though the Star Wars character might have authored the report.
What makes it possible are 12 different AI models working together, some from service providers, others open source.
“We are not building our own LLM but there is such amazing work in open source LLMs and the commercial LLMs that are available,” he said.
“Step one is for accredited investors, step two is to take it international—global both directly through us and through partners—and step three, take it to everyone,” he said, noting that it would take years to get there.
“In the long run someone with 1,000 [Indian] Rupees” will be able to invest with Arta, said Sengupta. Given his team’s success at Google (Google Pay is the second largest payment system in India) I’m hesitant to doubt him.