Hang Ten Systems, the enterprise AI services company founded by former Infosys chief executive Vishal Sikka, has raised another $53 million. The second seed round closed just five weeks after the first and takes total funding to $85 million. The new money is led by Xora, the early-stage investment platform backed by Singapore’s Temasek, with Mayfield and Aramco Ventures both returning.
The headline number is the easy part. The harder question is what a four-month-old company with roughly two dozen employees is actually selling to enterprises that turn over more than $10 billion a year, and whether the delivery claims behind the round survive contact with the source documents.
This article works through the round, the investor list, the Hobie platform, the named customers, and the places where the company’s own press release, TechCrunch’s reporting and the trade coverage do not agree with one another. Every figure below traces to a named source, and the References section links all of them.
Table of contents
- What Hang Ten Systems Announced on 16 September
- Who Put the $53M In, and Why the Round Closed in Five Weeks
- Inside Hobie, the Hang Ten Systems Platform
- Hang Ten Systems Customers: Confirmed, Implied and Unnamed
- Where the Sources Disagree About Hang Ten Systems
- The Arithmetic Behind the 10x Claim
- What Hang Ten Systems Is Actually Competing Against
- Vishal Sikka’s Second Act, and What VianAI Tells You
- What This Means for Buyers of Enterprise IT Services
- What to Watch Over the Next Two Quarters
- References
What Hang Ten Systems Announced on 16 September
The announcement went out on 16 September 2026 under a Menlo Park, California dateline. Hang Ten Systems said it had closed an additional $53 million in seed funding led by Xora, a Temasek-backed fund, bringing total funding in the company to $85 million. Mayfield, which led the first seed round, and Aramco Ventures both participated again.
Sikka confirmed the mechanics to TechCrunch. The second round closed five weeks after the $32 million first round closed, and the valuation, which he declined to disclose, was in his words “a bump up” from the first.
The round in numbers
| Round | Announced | Amount | Lead | Participating | Cumulative |
|---|---|---|---|---|---|
| Seed I | 24 June 2026 | $32M | Mayfield | Aramco Ventures (strategic), angel investors | $32M |
| Seed II | 16 September 2026 | $53M | Xora (Temasek-backed) | Mayfield, Aramco Ventures | $85M |
Hang Ten Systems is careful to call both raises seed rounds. That framing matters, and not every outlet followed it, which the source comparison later in this article returns to.
The board change that came with it
Jerry Yang, Yahoo co-founder and founding partner of AME Cloud Ventures, has joined the Hang Ten Systems board of directors. He was already an investor from the first round, and TechCrunch’s June coverage noted his board involvement at launch, so the September release formalises a relationship that already existed rather than announcing a new one.
Who Put the $53M In, and Why the Round Closed in Five Weeks
The cap table is unusually operator-heavy for a seed-stage services company. Alongside the institutional leads sit Lip-Bu Tan, chief executive of Intel; Sanjay Mehrotra, chairman and chief executive of Micron Technology; and Yang. These are people who buy enterprise software at scale rather than people who only fund it.
What each investor brings
| Investor | Profile | Role | What it adds |
|---|---|---|---|
| Xora | Temasek-backed; invests across AI Infrastructure, Applied AI and Deep Tech | Led Seed II | Access to Temasek portfolio companies in Singapore and Southeast Asia |
| Mayfield | Early-stage firm investing $3 billion in AI companies, 70% at inception | Led Seed I, followed on | 120 IPOs and 225+ acquisitions over 56 years |
| Aramco Ventures | $7.5 billion committed across four venturing programs, HQ in Dhahran | Strategic in both rounds | Deployment opportunities inside Aramco itself |
| Angel operators | Lip-Bu Tan (Intel), Sanjay Mehrotra (Micron), Jerry Yang (Yahoo) | Individual | Buyer-side credibility; Yang now on the board |
According to Sikka, Xora approached Hang Ten Systems rather than the other way round, prompted by the speed of the early customer wins. Phil Inagaki, Xora’s managing partner and chief investment officer, framed the thesis around implementation rather than models: “The bottleneck is implementation: deploying it securely, economically and at speed. That takes a different kind of services partner.”
Mahdi Aladel, chief executive of Aramco Ventures, was more specific about breadth: “The more closely we look at their approach, the more functions and use cases we find across our own operations where it would fit. Teams across our business are keen to work with them, and projects are already being scoped.”
The five-week detail is easy to misread
The two rounds closed five weeks apart. They were announced roughly twelve weeks apart, on 24 June and 16 September. The company’s own site describes “several multi-million-dollar enterprise engagements signed in the five weeks since launch”, which is the same five-week window seen from the commercial side rather than the financing side. Read the gap as closed quickly and announced on a normal press cadence, not as a round papered in a fortnight and shipped the next day.
Inside Hobie, the Hang Ten Systems Platform
The differentiator Hang Ten Systems claims is not a model. It is a platform called Hobie, which co-founder and chief design officer Sanjay Rajagopalan described to TechCrunch as an in-house framework packaging reusable AI “skills” for regulated industries and complex enterprise software projects.
The company breaks Hobie into five named layers. Requests flow down through them; grounded answers flow back up.
| Layer | Role | What it does |
|---|---|---|
| Shoreline | Customer-facing surface | Chat with memory, clarification and intent routing, in the browser or piped into tools like Claude and ChatGPT Desktop |
| Shaper | Builds and refines the agents | Reads live systems, then improves on a closed loop of evaluate, diagnose, fix, verify; mostly deterministic compilation, so results repeat |
| Lens | The semantic layer | A published, versioned model of the business: datasets, metrics, relationships and a plain-language vocabulary |
| Reef | Knowledge, queried in place | Runs queries against existing systems with nothing copied or migrated |
| Beacon | Audit and governance | Every answer traces to the exact data and model version behind it, with fail-closed, role-aware access |
Why query-in-place is the load-bearing claim
Reef is the layer that will decide whether Hang Ten Systems can sell into regulated buyers. Querying in place, with nothing copied or migrated, removes the data-residency argument that stalls most enterprise AI pilots. Beacon’s per-caller credentials and version-stamped answers are aimed squarely at audit committees. Whether all of that holds up across a pharmaceutical client and an energy major at the same time is not something any press release can settle.
The commercial model attached to it
Hang Ten Systems prices against a defined outcome agreed before work starts, and delivers through what it calls a small, senior forward-deployed team working alongside the client’s own people. Tao Liu, a co-founder, is senior vice president of forward deployed engineering, so the model is baked into the org chart rather than bolted on.
Hang Ten Systems Customers: Confirmed, Implied and Unnamed
Care is needed here, because the three available sources name customers differently, and two of them name different corporate entities.
| Customer | Source | Status | Note |
|---|---|---|---|
| Fresenius Kabi | TechCrunch, 16 Sep 2026 | Named | June coverage said only “Fresenius”, the parent rather than the Kabi division |
| Saudi Aramco | TechCrunch, 16 Sep 2026 | Named | Release says Hang Ten is working with Aramco on applications across several functions |
| Siemens Energy | TechCrunch, 16 Sep 2026 | Named | Release quotes Siemens Gamesa Renewable Energy, a Siemens Energy subsidiary |
| Siemens Gamesa Renewable Energy | Press release, 16 Sep 2026 | Named, with quote | CEO Vinod Philip is quoted directly |
| “Global life-sciences leader” | hangten.ai homepage | Unnamed | Logo-free placeholder |
| “Global pharmaceuticals and clinical nutrition leader” | hangten.ai homepage | Unnamed | Matches how Fresenius Kabi describes itself |
The pipeline figure, and what it is not
Sikka told TechCrunch that Hang Ten Systems is working with a mix of existing customers and late-stage prospects spanning 21 major enterprises, with some at proposal or contract-negotiation stage. That is a pipeline number, not a customer count, and the two get conflated quickly once a funding story is syndicated. The firmer claims are narrower: multiple seven-figure contracts secured, eight-figure deals being pursued, and customers across the United States, Europe, the Middle East and Asia.
The 25-day deal
The most striking operational claim is that the company signed a multimillion-dollar contract to build a mission-critical software system within 25 days of the first meeting with that customer. Sikka added that he had never seen enterprise deals move that quickly, which from someone who ran Infosys is a meaningful frame of reference rather than a throwaway line.
Where the Sources Disagree About Hang Ten Systems
Four discrepancies are worth logging, because they will propagate into secondary coverage and none is resolved by a single authoritative document.
| Point | Reporting says | Press release says | Why it matters |
|---|---|---|---|
| Headquarters | “Palo Alto-headquartered” in September; “Headquartered in the Bay Area” in June | Menlo Park, Calif. dateline | Palo Alto and Menlo Park are adjacent but distinct; the dateline is the company’s own |
| Round type | “Second seed round” (TechCrunch) | “Additional $53 million in seed funding” | Dealroom labelled it a Series A; unite.ai called it a second seed round |
| Customer entity | Siemens Energy; Fresenius Kabi | Siemens Gamesa Renewable Energy; Aramco | Parent versus subsidiary naming differs between the two |
| The amount itself | Headline reads $53M | $53 million throughout | TechCrunch’s own URL slug reads adds-50m, an editing artefact rather than a second figure |
How to cite this story safely
Use the press release for the funding mechanics, the investor list and the quotes, since that is the company speaking on the record. Use TechCrunch for everything the release does not contain: headcount, the 21-enterprise pipeline, the 25-day deal, the acquisition approaches and the VianAI status. Where the two conflict on a name, the release is the better source for entities Hang Ten Systems is contractually engaged with, and the reporting is the better source for context.
The Arithmetic Behind the 10x Claim
Hang Ten Systems promises customers a tenfold improvement in cost, speed, or a combination of the two. Its own site is franker about this than the marketing line suggests: “On some programs it shows up mostly as speed, on others mostly as cost, and often as a mix of both. We agree which matters most to you before we scope the work.”
The three published inputs
The company cites three figures, each attributed to third-party research rather than to its own delivery data.
| Figure | What it measures | Attributed to | Caveat |
|---|---|---|---|
| 26-55% | Productivity lift from agentic development | METR (2025) | A range, not a point estimate |
| 2-5x | Reduction in the customization tax | GitHub / MIT Copilot productivity study | Applies to customisation effort, not whole programmes |
| Approximately $0 | Marginal cost of custom software | McKinsey, State of AI (2024) | A directional claim about marginal, not total, cost |
None of those three inputs multiplies out to ten on its own, and Hang Ten Systems does not claim they do. The 10x figure is a commercial commitment negotiated per engagement, not a derived result. That is a more defensible position than a headline multiplier, and worth noting before anyone treats it as an industry benchmark.
The staffing claim is the sharper number
Rajagopalan told TechCrunch that some projects can be staffed with two to four people where roughly thirty would previously have been required. Customers or independent third parties still handle final quality checks and certification, so the compression sits in build and iteration rather than across the whole assurance chain.
Sikka’s framing of why this is possible is blunt: “The build part itself has basically become close to zero marginal cost, close to zero time.” His argument is that AI shifts software work from writing code toward defining requirements and validating that systems behave as intended, which is also, conveniently, the part of the job that has always been billed as advisory.
What Hang Ten Systems Is Actually Competing Against
The market Hang Ten Systems has walked into is crowded from three directions at once, and the competitive answer differs for each.
| Competitor class | What they are doing | The stated answer |
|---|---|---|
| Model developers (OpenAI, Anthropic) | Expanding enterprise offerings and services arms | Independence from the underlying platform |
| Traditional systems integrators | Embedding generative AI into existing delivery | No legacy revenue to protect; smaller teams per programme |
| Incumbent IT services firms | Partnering with model vendors to retrofit AI | Built AI-native rather than retrofitted |
| In-house enterprise teams | Hiring and building internally | Forward-deployed senior teams plus a reusable skills library |
Sikka’s own line is the clearest statement of the positioning: “Enterprises still need trusted partners who are independent of the underlying platform and who work with and solely in the interest of the enterprise. There will always be room for companies like us.”
Replacement versus net-new work
This is the detail most coverage skipped. Many Hang Ten Systems engagements are replacing incumbent providers, but Sikka said more than half of current opportunities are new projects companies had previously put off, rather than work taken from existing vendors. If that ratio holds, the near-term commercial story is unlocked backlog rather than displaced spend, which is a very different threat profile for the incumbents.
The disruption argument, in context
Mayfield’s Navin Chaddha put the structural case plainly in June: “Traditional services scale linearly with headcount. Hang Ten is built so its leverage grows with every project.” Whether that leverage is real at a hundred engagements rather than a handful is the open question, and no amount of seed funding answers it. Anyone evaluating an IT outsourcing arrangement in 2026 now has to price in the possibility that the delivery-cost curve moves under them mid-contract.
Vishal Sikka's Second Act, and What VianAI Tells You
Sikka spent twelve years building enterprise software at SAP, later sat on Oracle’s board, and stepped down as Infosys chief executive in 2017. He founded VianAI in 2019, which emerged from stealth with $50 million in seed funding and raised $140 million in a 2021 round led by SoftBank Vision Fund 2. He left VianAI in April 2026 and told TechCrunch that company is “going through a transaction”, declining to elaborate.
Why this is not VianAI again
Chaddha drew the distinction explicitly. VianAI focused on enterprise AI applications and analytics tools for decision-making, whereas Hang Ten Systems describes itself as an enterprise AI services company built around agentic code generation, a reusable skills library and domain expertise. One sold software; this one sells delivery. The early crew includes executives who worked with Sikka across SAP, Infosys and VianAI, among them co-founders Navin Budhiraja as chief technology officer, Rajagopalan as chief design officer and Liu leading forward deployed engineering.
Where the name comes from
Hang Ten Systems takes its name from the surfing manoeuvre in which a rider puts all ten toes over the nose of the board. Sikka has used the metaphor from the start, describing the aim as helping enterprises ride what he called “probably the biggest wave of our lives”.
What This Means for Buyers of Enterprise IT Services
For an enterprise technology buyer, the interesting content of this round is not the money. It is the set of commitments a well-funded new entrant is now willing to make in writing.
Four things to take to your next vendor conversation
The claims below are Hang Ten Systems’ own, and they are useful precisely because they can be put to incumbents as questions.
| Commitment | What it means in practice | The question it licenses |
|---|---|---|
| Outcome-based pricing | Price agreed against a defined outcome before work starts | Will you price this programme against an outcome rather than a rate card? |
| Query in place | No data copied or migrated out of existing systems | Can your AI delivery run without extracting our data? |
| Traceable answers | Every answer traces to an exact data and model version | Can you version-stamp and audit every generated artefact? |
| Small senior teams | Two to four people where thirty were typical | What is your headcount assumption, and what happens to the price if it falls? |
None of this requires believing Hang Ten Systems will succeed. A credible competitor making these commitments changes what is negotiable, and that is true whether or not the company is on your shortlist. The same logic applies to any digital transformation programme currently being scoped on traditional assumptions about team size and timeline.
The counterweight
Two dozen people cannot deliver twenty-one enterprise programmes. Hang Ten Systems currently has roughly 20 to 25 employees across the United States, the Middle East and Australia, and plans to hire in Europe and India. Delivery capacity, not demand and not capital, is the visible constraint, and it is exactly what the $53 million is earmarked for alongside engineering, consulting and sales expansion.
What to Watch Over the Next Two Quarters
Four signals will tell you more than the next funding headline.
Whether headcount tracks the pipeline
If Hang Ten Systems is still at 20 to 25 people in six months while claiming a larger pipeline, the forward-deployed model is either extraordinarily leveraged or the pipeline is not converting. Hiring in Europe and India is the tell, since India is where the custom software development delivery capacity that incumbents rely on actually sits.
Whether an eight-figure deal lands
Seven-figure contracts are signed; eight-figure deals are described as being pursued. A disclosed eight-figure engagement would be the first hard evidence that the model scales past pilot-adjacent work.
Whether the acquisition approaches return
Sikka says Hang Ten Systems has received serious acquisition offers from “very big companies” within four months of launch and has turned them down for now. He declined to name them. An acquisition at this stage would say more about incumbent anxiety than about the company.
Whether named customers go on the record again
Siemens Gamesa has already widened the scope of its work after two months of delivery, according to its chief executive. A second named reference from Fresenius Kabi or Aramco would move the story from one enthusiastic customer to a pattern.
References
Former Infosys chief’s AI startup nabs another $53M
Hang Ten Systems raises additional $53 million led by Xora, bringing total funding to $85 million
Former Infosys chief has a new startup that wants to challenge the IT services world
Hang Ten Systems: Software built to your specification
Hang Ten Systems Raises $32 Million Seed Round to Accelerate Enterprise AI Adoption
More AI coverage: explore Progressive Robot's AI Models, Tools & Releases hub — hands-on reviews, setup guides and benchmarks in one place.