Ex-Infosys chief's AI outsourcing venture: the one move Japanese companies should make now

For Japanese companies outsourcing development to Indian IT contractors, a move has emerged that shouldn't be overlooked. Vishal Sikka, former CEO of Infosys, launched a new company, Hang Ten Systems, in late May 2026, and announced on June 24, 2026 that it had raised $32 million in a seed round. The aim is to replace the traditional IT outsourcing model, where effort piles up in proportion to headcount, with reusable "AI skills" and agentic code generation. Sikka himself once led Infosys, a major outsourcing player. The fact that this very person now says he wants to "break the conventions of outsourcing" suggests that the service content of Indian IT companies themselves will change within a few years. If a client company is going to rethink the premises of vendor selection, now — before its counterpart moves to an AI-native delivery model — is the time to do it.

What is Sikka's Hang Ten Systems?

Hang Ten Systems is an AI services company that aims to continuously "build, modify, and operate" enterprise software. At its core are agentic code generation, reusable AI skills that can be carried across projects, and domain expertise. Functions and know-how built once are stored as skills and combined for use in the next project. This is the fundamental difference from conventional outsourcing.

Founding members include CTO Navin Budhiraja, Chief Design Officer Sanjay Rajagopalan, and Tao Liu, who leads forward-deployed engineering. Yahoo co-founder Jerry Yang has joined as a director. Sikka himself spent 12 years building enterprise software at SAP, served as CEO of Infosys until 2017, and has since led his own AI company, VianAI.

Why is a former outsourcing chief now denying outsourcing?

Behind this is a broader "departure from the man-month business" underway across India's IT industry. Major players such as TCS, Infosys, Wipro, and HCLTech are rolling out AI literacy training across hundreds of thousands of employees at once. According to reports, a majority of Infosys employees have reportedly acquired AI literacy, and HCLTech is said to have won large AI-related contracts. The shift from writing code through sheer manpower to having AI write code while people focus on design, verification, and operations is becoming the industry mainstream.

What sets Sikka's Hang Ten apart is that it pursues this shift not as "improvement from within an existing outsourcing company" but as "a new company built AI-first from the ground up." Mayfield, which led the funding round, commented: "Traditional services only scale in proportion to headcount. Hang Ten is designed so that leverage increases with every project." It is an idea that reverses the labor-intensive nature of outsourcing itself: reusable assets accumulate the more projects are completed, raising productivity on the next one.

The scale and names behind the funding

Converted at roughly ¥155 to the dollar (a mid-2026 guide), $32 million comes to about ¥4.9 billion. That's large for the seed stage, and the initial customer list already includes real major enterprises — German renewable-energy giant Siemens Gamesa and healthcare company Fresenius. Landing real customers barely a month after founding shows that this is not just a concept but that a track record of delivery is already in motion.

Reactions from the front lines of outsourcing

Reactions on both the client and vendor side show a temperature gap. In India's engineering community, there is a positive view that "if the effort of writing code decreases, the valued skill shifts to requirements definition and review," alongside a sense of crisis that "if projects earning revenue on unit price times headcount decline, mid-tier and smaller outsourcing companies will struggle."

On the Japanese client side too, differences in thinking about quality tend to become a point of contention. It has long been pointed out that there is a cultural gap between Japan, which wants near-100% completeness before starting, and the Indian side, which moves forward at around 80% and fixes as it goes. As AI makes it possible to produce a first draft at high speed, this "launch at 80%" style is, if anything, likely to accelerate. If the client side sticks to its conventional acceptance-testing process, it will not fully capture the benefit of faster delivery.

Major players are not sitting still either. Global system integrators, including TCS and Infosys, are moving quickly toward an AI-native delivery setup, for instance by joining as certified partners in support programs for generative AI code-generation platforms. Sikka's new company is trying to strike at the leading edge of that same trend with the agility of a startup.

Implications for Japanese companies with Indian IT outsourcing

To narrow it down to one concrete move companies currently holding outsourcing contracts should make, it is to include, in the next quote request for a renewal or additional order, a question asking "how would the workload and unit price change under AI-native delivery," as part of a competitive quote comparison. Get the current vendor's quote based on conventional effort as usual, and in parallel obtain a re-quote assuming an AI-first approach. The gap between the two directly reflects how serious that vendor is about moving to AI. Now that players like Hang Ten have emerged with a design where "leverage increases the more it's used," a quote that keeps man-month pricing unchanged will only become relatively more expensive over time if left as is.

What should also be confirmed at this point is ownership of reusable assets. If AI skills and agents are reused across projects, it needs to be settled contractually who captures the resulting productivity gains. Does the accumulated asset make your own development cheaper, or is it diverted to other clients' projects while your own unit price stays flat? Leave this ambiguous, and the AI-driven gains risk being captured by the vendor side alone.

Impact on the market

What Sikka's entry symbolizes is that the axis of competition in Indian IT is shifting from "how cheaply and how many people can you supply" to "how cleverly can you reuse assets." As this progresses, the evaluation criteria on the client side also change. Selection shifts away from a focus on "the vendor's headcount and site scale" toward looking at "how much reusable asset does the vendor hold" and "how are skills from past projects carried forward into the next one."

In the Japanese market, TCS and Infosys are strengthening their offensive by forming blended teams with local skilled talent, and an AI-first delivery setup will further push up the productivity of these blended teams. If clients keep structuring contracts around the old model of on-site staffing and man-months, they risk falling behind rivals who have switched to AI-native competitors on both cost and speed.

Practical information and related links

AI-related capital movements by Indian tech companies and their practical points of contact with Japan are advancing at the same time. Related moves worth keeping an eye on include the following.

• As an example of an AI company M&A in the marketing space, MoEngage's acquisition of Aampe a useful reference.
• As a concrete example of Japan and India teaming up on contract manufacturing, Kaga Electronics and Syrma SGS's EMS partnership can be cited.
• On the semiconductor front, for Japan-India contact points, Kaynes's move to supply chips to the Japanese market is also worth checking alongside these.

Summary: add one line about "AI premise" to your next quote request

Sikka's Hang Ten Systems is still a startup that has just closed its seed round. But the fact that the very person who once led the flagship of outsourcing has raised $32 million around a model of "growing assets, not headcount" is close to a starting gun for a shift in how Indian IT delivers services. A concrete move Japanese companies with outsourcing contracts can make now is to add one line to their next quote request for a renewal or additional development — asking about the workload, unit price, and ownership of reusable assets under AI-native delivery — and have the current vendor answer it. How crisply that answer comes back is, on its own, the material for deciding whether to switch or to continue. Locking in terms before your counterpart finishes its AI transition gives you more negotiating leverage than moving after they already have.

Sources

• TechCrunch — Former Infosys chief has a new startup that wants to challenge the IT services world
• Entrackr — Former Infosys CEO Vishal Sikka's Hang Ten Systems raises $32 Mn Seed round