A factory is an argument between eras. A new vision system watches labels fly past at speed. Three metres away, an old press communicates in a dialect last fashionable when dial-up internet still squealed. Operators add context on paper. An ERP system offers a neat answer several hours later. Every machine may be telling the truth, yet the plant still lacks a version of events that everyone can use.
Vinay Nathan has built his second career around sorting out that argument. The CEO and co-founder of Altizon is an engineer who wandered through product management, marketing and global sales before arriving, in 2013, at the industrial internet. He and co-founders Yogesh Kulkarni and Ranjit Nair started with Datonis, a platform designed to connect industrial equipment, collect its signals and turn them into applications. Thirteen years later, the vocabulary has moved through IoT, Industry 4.0, digital factories and industrial AI. The awkward machines remain.
This is why Nathan's story is more useful than a standard tour of technology fashion. He has not merely watched the labels change. He has had to sell through each one, install behind it and persuade manufacturers that a promising demonstration can survive a Monday morning shift.
The four-day delay
The original insight was local and wonderfully unglamorous. Pune is surrounded by auto-component manufacturing. Nathan saw plants full of equipment assembled from different vendors and generations. Sensors existed, but meaningful connections were scarce. A tier-two supplier might learn its production details four days after the production happened. By then, analysis was less a management tool than a post-mortem.
Nathan described the opportunity in plain terms: connect the machines, drive data from them and build applications around predictive maintenance and equipment efficiency. Management did not need more numbers to admire. It needed an operational view while intervention still mattered. His sharper formulation was that people should enter the analysis when a system finds an anomaly, not spend their days manually hunting through normal behaviour.
“It is no good collecting the data, if human is involved in analyzing. Human should only be involved when there are anomalies in the system.”Vinay Nathan, 2017
That distinction remains the hinge of Altizon's business. Data collection is plumbing. A useful alert is a decision arriving on time. The distance between those two things contains integrations, context, permissions, models and the hard-won trust of the person who has to stop a line.
The factory-truth pipeline
A CEO assembled from spare parts
Nathan's route to the corner office looks less like a ladder than a well-stocked workshop. He earned a computer-engineering degree from the University of Pune, then a master's in computer science at the University of Southern California. At USC he worked as a graduate research assistant on software architectures. Early roles at Pace Soft Silicon and Encodex Technologies kept him close to engineering and product work.
Then his career crossed the membrane between building and selling. At Persistent Systems, he eventually led sales across Asia-Pacific and the Middle East. The move matters. Industrial software is a translation business: sensor to signal, signal to model, model to operator, technical capability to a line item that a chief financial officer will approve. Nathan learned both halves of the sentence.
His technical curiosity was not decorative. Nathan is named on issued US patents covering memory devices, copy protection, phone and internet services, wireless communication and secure data transmission. Those subjects sound distant from a bottling line, but the common thread is visible: devices, data and the rules governing their movement.
Selling the clock, not the cloud
Enterprise software becomes vague when nobody mentions time. Nathan does. In a 2018 discussion of Altizon's deployments, he said 65 percent of customers achieved a return within six to twelve months. In 2020, he estimated that connecting a small plant could take a couple of weeks, while a very large site might take a couple of months. Those figures are not magic. They are boundaries a buyer can argue with, plan around and eventually audit.
His go-to-market disclosures were equally concrete. Around 70 percent of Altizon's lead generation came from outbound work, and around 70 percent of its business was directly sourced. Nathan expected the second ratio to flip as the company scaled. Systems integrators would connect machines, join data silos and build custom applications; Altizon would provide the reusable product layer and share recurring software revenue.
The forecast took years to become infrastructure. In 2024, Altizon put its DFX platform in Microsoft Azure Marketplace. In July 2025, it announced a full portfolio migration to Azure and introduced APEX, a formal partner program with referral, resale and strategic tiers. The old direct-sales company had not vanished. It had built a wider door.
“Our customers wanted a faster, lower-friction way to deploy AI across factories through the Microsoft partners they already trust.”Vinay Nathan, 2025
The product changes; the plant does not
Altizon's first platform, Datonis, joined devices to cloud infrastructure and analytics. Its manufacturing suite added applications for productivity, quality, maintenance, energy and traceability. DFX now presents those capabilities as a digital-factory intelligence layer, bringing shop-floor and business records together and adding predictive and generative AI.
The progression is sensible because the factory sets the order. First connect heterogeneous equipment. Then digitise the observations that live on clipboards and in operators' heads. Create a record that can span lines and plants. Only then ask models to predict quality drift, energy waste or equipment trouble. A conversational interface is pleasant. A conversational interface attached to an unreliable production record is a confident gossip.
Nathan's recent essays return to that point with increasingly literary titles. “Before JARVIS, There's Ground Zero” considers the groundwork beneath an AI agent. “You Don't Buy Capacity. You Recover It” looks for output hiding in existing operations. “The 8% Nobody Is Counting” questions the number presented on an OEE dashboard. Even a reference to Kurosawa's Rashomon becomes a way to discuss fragmented factory knowledge: several internally consistent witnesses, no shared account.
There is a personality in those choices. A former colleague described Nathan as technically and commercially sharp, attentive to detail, humble and fun to work with. His public writing shows the same preference for a concrete scene over a cloud of abstraction: a bottling line, an incomplete cleaning cycle, an audit that has swallowed three days.
Patient capital, patient company
Altizon has raised more than $12 million in disclosed funding. Persistent Systems backed it early. Wipro Ventures led a $4 million round in 2016. A Singapore subsidiary of TVS Motor led a $7 million Series A in 2019, joined by existing investors. The company earned Gartner recognition as a Cool Vendor in 2015 and appeared in the firm's industrial IoT research in later years.
Those milestones are tidy. Building industrial software is not. Plants do not replace equipment because a startup has released a nicer slide deck. Security rules vary. Operational data is sensitive. Every site contains exceptions accumulated through decades of practical compromise. Even a successful product must coexist with the human habit of knowing which machine “always runs a little warm on Tuesdays.”
Nathan appears comfortable with that patience. Alongside Altizon, he has taken on mentoring work through TiE, including co-chairing TiE New Jersey's equity-free accelerator for early-stage startups. It fits a founder who speaks less about instantaneous disruption than about channels, implementation work and repeatable economics. He does not pretend the last mile is a rounding error. He tries to decide who should own it and how they should be paid.
When the factory speaks
The future Nathan describes is full of predictive models, digital twins and software agents. Yet the aspiration underneath is modest enough to be credible: help the people running a plant know what is happening while they can still do something about it. Save an avoidable stoppage. Catch quality drift before a batch becomes scrap. Replace a three-day audit with a record assembled as the work occurs.
Thirteen years after Altizon began, factory AI is fashionable. Nathan's advantage is that he remembers when connecting a machine to the internet was the fashion, and when a real-time dashboard was the fashion before that. He has learned to treat fashions as interfaces and factory outcomes as the durable thing beneath them.
A plant will never be as neat as its digital twin. It is louder, older and staffed by people who know things the database does not. The interesting work lies in making those versions of reality agree often enough to act. Nathan has spent a career translating between them. The machines may finally be ready to answer questions. He is still making sure they have their facts straight.