KR Sridhar and the problem with being early

The customer on Mars was proving difficult. KR Sridhar had spent years working on a way to extract oxygen from the Martian atmosphere, only to see the mission meant to carry the experiment cancelled. There was a practical problem with that market, too. “Martians weren’t paying good money either,” he would later joke. Earth, with its electricity bills and impatient customers, offered a rather more promising address.

That change of destination became Bloom Energy. But the interesting part of Sridhar’s story is how much had to change after the destination did. An academic who knew how to make a sophisticated experiment work had to learn how to make other people manufacture it, persuade companies to buy it, and keep it running after the visiting television crew had packed away its cameras.

Today he is Bloom’s founder, chairman and chief executive, selling on-site power into a computing boom. The route from planetary exploration to data centers took roughly a quarter-century. Along the way, he discovered that an engineering drawing can be correct and a factory can still be in trouble.

A cancelled flight, a different direction

Sridhar’s training crossed several engineering borders before his career crossed any planetary ones. He completed his undergraduate mechanical engineering degree at the institution now called NIT Tiruchirappalli in 1982. At the University of Illinois, he earned a nuclear engineering master’s in 1984 and a mechanical engineering doctorate in 1990. The combination gave him experience with energy systems well before he had a company to run.

At the University of Arizona, he became a professor of aerospace and mechanical engineering and directed the Space Technologies Laboratory. He worked with NASA on systems that could turn gases in the Martian atmosphere into useful oxygen. For an expedition, producing something at the destination could save the enormous inconvenience of carrying it there.

In 1997, an oxygen-generating system proposed by Sridhar and chemist John Finn was selected for the Mars 2001 Survey Lander. After NASA’s Mars mission failures in 1999, that mission was cancelled. The research continued, but its commercial future moved closer to home. Sridhar and Finn would build a business around reversing the electrochemical direction: use fuel and oxygen to produce electricity.

Bloom traces its beginnings to 2001; the business was incorporated in 2002. Originally called Ion America, it developed in leased space at NASA Research Park before adopting the Bloom Energy name. The origin is unusually literal. A company eventually associated with powering computer servers grew out of work on how to make another planet more usable.

The factory did not read his mind

The transition from laboratory to production supplied a less photogenic adventure. Sridhar has described how the early boxes worked, then later units failed in the field. The engineers believed the design could be handed over to manufacturing. Too much of what made it work, however, remained in the engineers’ heads.

He stopped production and asked for help. For a meeting with advisers including Intel’s Andy Grove, the team assembled stacks of binders and ordered dinner, expecting a long examination of the evidence. Grove redirected the discussion toward Sridhar himself. Had he walked the floor and asked the people building the machines what they did not understand?

Sridhar’s account makes the lesson uncomfortable in a useful way. The founder was looking for manufacturing expertise outside the company while overlooking the knowledge already inside it. A technician following instructions could expose an ambiguity that the inventor, who supplied the missing understanding automatically, would never notice.

The episode helps explain the executive he became. He describes empathy as a working practice: understand the person assembling the product, then understand the person paying for it. A founder can explain a machine beautifully and still fail to explain how to build it. Teaching turns out to be part of the production process.

Television loved the little box

In February 2010, Sridhar introduced the technology to a much larger audience on 60 Minutes. The public unveiling followed at eBay’s San Jose campus. Google, Walmart, FedEx and eBay helped give the proposition something every unfamiliar invention needs: customers whose names require no explanation.

The showmanship had an accessible object at its center. Thin ceramic cells, coatings and stacks were easier to picture than an abstract discussion of distributed generation. Sridhar could hold a piece of the system while discussing the electricity it might eventually supply. For television, a small object carrying a large promise is excellent casting.

The promise also reached beyond the customers already installing commercial systems. Sridhar talked about getting into homes within five to ten years, with a residential unit eventually costing around $3,000. That remains a revealing historical ambition. The business visible today is built around commercial and industrial power, including customers whose electricity requirements make the family toaster look modest.

The underlying system converts fuel into electricity electrochemically, without combustion. Individual solid oxide cells form stacks; stacks form modules; modules combine into a larger installation. Natural gas, biogas and hydrogen are among its possible fuels. When it uses natural gas, it still produces carbon dioxide. A quiet box does not get an exemption from chemistry.

For Sridhar, the appealing feature is electricity made where it is needed. The customer can put generating capacity beside the activity it supports. That proposition has stayed recognizable even as the scale of the customer, and the urgency of the sale, have changed.

Recognition came with a reporting obligation

Engineering institutions recognized the work. In 2016, Sridhar was elected to the National Academy of Engineering for work involving transport phenomena, the thermal packaging of electrochemical systems, and power generation. Bloom went public on the New York Stock Exchange in 2018. The inventor now had an audience that expected quarterly accounts as well as working machinery.

That transition exposed another difficulty. After an interview around the IPO, Bloom filed clarifications and corrections concerning Sridhar’s remarks about profitability and shareholder sales. The episode matters because profitability has specific accounting meanings. Optimism, however sincerely held, cannot settle the difference between a financial measure and another one.

A personal profile of Sridhar has to hold both experiences together. He built a company around demanding physical technology, and the company’s public claims required scrutiny. The distance between a laboratory demonstration, a customer installation and a sustainable business is considerable. Each stage asks for a different kind of proof.

His connections span those stages. John Doerr and Kleiner Perkins were early backers; Grove became a mentor. Sridhar also serves on C3 AI’s board and on academic advisory bodies at Caltech and Illinois. He returned to Illinois as an engineering commencement speaker in 2024, the same year the University of Delaware awarded him an honorary science doctorate. The former professor has kept finding reasons to return to campus.

Then the customers needed power sooner

The recent turn in the story has a distinctly terrestrial cause: people building AI infrastructure need electricity on a schedule. A computer can be ordered, a building can be constructed, and the power supply can still determine when the investment starts doing useful work.

For Sridhar, this has made time itself a selling point. Bloom’s April 2026 expansion with Oracle covered intended procurement of up to 2.8 gigawatts, with an initial 1.2 gigawatts contracted and deploying. The announcement also described an earlier Oracle system delivered fully operational in 55 days, ahead of a planned 90-day schedule.

In June, Brookfield expanded its framework for financing power projects with Bloom from $5 billion to $25 billion. Those figures describe a financing framework, rather than money already earned by Bloom. They nevertheless show the infrastructure scale at which Sridhar is now trying to operate.

By the second quarter of 2026, Bloom reported revenue of $1.065 billion, its first quarter above $1 billion. In September it published a case for supplying AI data centers with native 800-volt direct current, arguing that fewer conversion stages could reduce equipment and costs. The financial results are reported history; the design’s projected savings remain Bloom’s estimates.

That distinction will matter as the orders grow. Manufacturing, installing and maintaining power equipment require work long after a contract announcement. Sridhar’s old factory lesson has acquired a larger setting: understanding the process matters at least as much as understanding the invention.

The professor’s unfinished assignment

There is a longer horizon behind these contracts. In his writing, Sridhar treats access to electricity as a condition for participating in a digital economy. His stated ambition extends to people and places far beyond the current race to build AI facilities. Bloom’s early appeal to household power and its present appeal to computing infrastructure belong to that same expansive ambition, even when the customers look very different.

“I look for passion, resilience, and ability to operate when there’s a lot of ambiguity.”

KR Sridhar, on choosing senior leaders

He has also kept a teacher’s interest in the next generation. His Stanford biography describes his enjoyment of encouraging schoolchildren to study mathematics and science. There is something fitting about that detail. A career that travels through engineering degrees, university laboratories and a manufacturing floor leaves plenty to explain to an inquisitive child.

The question awaiting Sridhar is practical: how widely can the electricity he talks about become available, and at what cost? His answer has occupied decades, several kinds of institution and a collection of customers that Mars never managed to supply. The box has found its audience. Keeping it working is the continuing assignment.