Imagine walking into a restaurant that has no menu, no dining room, and no kitchen, but a line of investors outside handing the chef sacks of cash based on the meals he cooked at his last job. That is roughly where we are with Hang Ten Systems, which just added $53 million to its seed round and now sits on $85 million total. The money is real. The product, from where I sit as someone who tests these things for a living, is still a rumor.
The startup was founded by Vishal Sikka, formerly CEO of Infosys, and the new capital was led by Temasek’s Xora with Mayfield and Aramco Ventures joining in. The stated focus is enterprise AI services. That is the extent of what anyone outside the company can say with confidence, and I want to be upfront about that before offering an opinion.
Why a reviewer cares about a funding round at all
Normally I would skip this story. Funding announcements are not products, and But $85 million in seed money for an enterprise AI services company matters to the people who read this site, because that money eventually turns into a sales motion aimed directly at them. In about a year, somebody on your team is going to forward you a pitch deck from this company, and the question you will be asked is whether it belongs in your stack.
So consider this an early note in the file rather than a review. There is nothing to score yet.
The founder-credential premium
The most interesting thing about this raise is what it tells you about how AI capital is being allocated right now. A former CEO of one of the largest IT services companies on earth can raise seed money at a scale that most Series B companies would envy, before there is a public product to evaluate. Investors are paying for a resume and a thesis.
That is not automatically a bad bet. Sikka’s background is genuinely relevant to the problem. Enterprise AI services is not a technical puzzle so much as an organizational one, and someone who has run a global services business at scale understands the messy parts: procurement cycles, integration debt, the gap between what a pilot demonstrates and what production tolerates. Plenty of well-funded AI companies fail precisely because they are staffed entirely by people who have never had to deploy anything inside a bank.
But the credential premium cuts the other way too. Money raised on reputation tends to buy time rather than validation. A company with $85 million and no shipped product does not get the brutal, useful feedback loop that a scrappier team gets when it has to charge customers by month four. I have reviewed enough tools built in well-funded isolation to know how that turns out: polished, ambitious, and slightly detached from what practitioners actually asked for.
What I will be looking for
When Hang Ten Systems eventually puts something in front of buyers, these are the questions I plan to ask, and I would suggest you keep the same list handy:
- Is it a product or a services engagement wearing a product’s clothes? “Enterprise AI services” is a phrase that can describe software you install or a team of consultants you rent. The pricing model will tell you which one you are buying.
- What does the deployment story look like on day thirty? Not the demo. The point where your own data, permissions, and legacy systems get involved.
- Who owns the model layer? Enterprise AI companies sit somewhere on a spectrum between building their own models and orchestrating other people’s. Both are valid. Which one you are dealing with determines your switching costs.
- Can a mid-sized company use this, or is it a Fortune 100 product? A funding round this size usually implies the second, which means the rest of us wait for the down-market version.
- What happens when it is wrong? Error handling and human review paths are where enterprise AI tools either earn trust or quietly get abandoned.
My honest read
Xora, Mayfield, and Aramco Ventures writing checks together says the thesis is credible to people with real diligence budgets. Aramco Ventures in particular suggests industrial and energy interest, which is a market with genuine appetite for AI services and very little tolerance for flaky software. That is a demanding first customer base, and demanding customers make better products.
Still, I would encourage the same posture I take with every well-funded newcomer: interest without commitment. Nobody outside the company has used this thing. The funding total tells you what investors believe, not what the software does. Those are different pieces of information, and conflating them is how procurement teams end up with expensive shelfware.
We will test it when there is something to test. Until then, this is a name to know and a claim to verify later.
🕒 Published: