Tom Leighton and the mathematics of getting out of the way

Tom Leighton helped invent a business before he knew what a venture capitalist was. He had a doctorate, a faculty position at MIT, and years of experience with difficult mathematics. The abbreviation VC, however, had escaped his education. It is a pleasing detail in a career that would eventually involve public markets, acquisitions, and infrastructure contracts with billions of dollars attached. Everyone has to start somewhere. His somewhere happened to be rather close to a blackboard.

The company was Akamai, founded in 1998 with his graduate student Danny Lewin. Its central proposition sounded almost impolitely practical: when too many people want something on the web, arrange for the content to be available across a distributed network. Stop sending every request into the same traffic jam. The mathematics underneath required considerable invention; the desired result was something a person waiting for a web page could appreciate without a PhD.

Leighton’s career has since acquired an unusual shape. He spent fourteen years as Akamai’s chief scientist before becoming chief executive in January 2013. He has kept his MIT affiliation, where he is a professor of applied mathematics on leave. The research and the company have remained connected for decades. A proof becomes software, software becomes infrastructure, and the infrastructure presents a fresh set of problems to solve.

A childhood with access to the machines

He grew up in Arlington, Virginia. His father worked in the Navy’s nuclear power program and later continued that work as a civilian; his mother became a librarian. Leighton remembers parents who took education seriously and teachers and people in industry who made opportunities available. As a teenager, he could use large computers for number theory research. The equipment was roomier than its modern descendants. The questions were already his own.

In 1974, he took second place in the Westinghouse Science Talent Search. He had also participated in the International Science and Engineering Fair in 1972 and 1973. Those dates matter in the longer story: the executive who now supports student science competitions once stood on the student side of them. Encouragement arrived early, through adults willing to give a young person time, attention, and access to computing.

Princeton came next, with a degree in electrical engineering and computer science in 1978, followed by a doctorate in applied mathematics at MIT in 1981. He joined MIT’s mathematics faculty in 1982 and became a professor in 1989. This was a substantial academic career before a startup entered the picture. His subjects included algorithms, parallel computation, and the organization of networks - ways to make many pieces of a system cooperate without wasting their effort.

The web had a queueing problem

In 1995, Tim Berners-Lee challenged colleagues at MIT to find a better way to deliver web content. The web’s promise was obvious. Its capacity to disappoint an impatient user was becoming equally apparent. A popular destination could draw more requests than its infrastructure could comfortably handle. Success itself could make a site difficult to use. For an algorithms researcher, that was a problem with both intellectual interest and a conspicuous audience.

Lewin arrived at MIT in 1996, and his work with Leighton helped produce the ideas behind Akamai. Copies of content could be placed across distributed servers, with requests intelligently directed to suitable machines. Location, load, and the changing condition of a network all mattered. Think of a crowded reading room whose books can also be collected from neighborhood branches. The analogy is simple; making the branches coordinate under pressure is where the work begins.

The founders entered MIT’s $50K entrepreneurship competition and did not win. The exercise still forced them to think about customers and a business plan. Their initial hope of selling the technology to internet service providers met resistance. Website operators proved more receptive. A mathematically convincing idea had needed a commercially convincing audience. Akamai was incorporated on August 20, 1998; Jonathan Seelig and Randall Kaplan also joined the founding team.

Two colleagues, one small desk

Leighton and Lewin’s partnership began in research and became the daily labor of building a company. At one point, they even shared a small desk. Leighton later remembered the combination that made Lewin a trusted partner: intelligence, reliability, determination, and hard work. The physical detail is useful. Before Akamai became a vast distributed system, the founders themselves were working in rather concentrated quarters.

An early test arrived during March Madness in 1999. With ESPN and its hosting provider under pressure, Akamai was asked whether it could handle 2,000 content requests a second. The young company had been serving a request only every few minutes. It accepted the job and handled 3,000 a second. That gap between the existing workload and the requested one gives the episode its tension. The eventual customers would be buying something more persuasive than a business-plan prize.

Akamai went public in October 1999. The collapse of the technology bubble followed. Then, on September 11, 2001, Lewin was killed aboard American Airlines Flight 11. The loss remains part of the company’s history and Leighton’s. In 2017, both founders were inducted into the National Inventors Hall of Fame, recognizing the content delivery methods they had developed together.

When the chief scientist took the chair

Leighton succeeded Paul Sagan as CEO in 2013. Moving from chief scientist to chief executive changed the range of decisions attached to his name. A network has to function; a company must also decide where to invest, what customers need next, and which capabilities to build or buy. The technical foundation remained valuable, but its commercial uses could keep changing.

Cybersecurity became a central example. Akamai was already handling traffic for customers’ websites and applications. It expanded into protecting those properties, including through a web application firewall delivered as a cloud service. In his own account of the company’s development, Leighton emphasized listening to customers. The delivery network gave Akamai a place to observe and intercept threats. A position earned by moving content could support a business in protecting it.

By 2025, security revenue was $2.2 billion, the majority of Akamai’s roughly $4.2 billion annual revenue. In 2012, before Leighton became CEO, security revenue had been below $25 million. These figures describe a considerable change in the work customers were paying Akamai to do. They also explain why describing him only as the co-founder of a content delivery company leaves a large part of his operating career outside the frame.

The network learns another job

Cloud computing added another direction. Akamai announced its agreement to acquire Linode in February 2022. Linode brought developer-oriented computing services; Akamai brought an established distribution network and security business. The ambition was to let customers build, run, deliver, and protect applications across a broader platform. For Leighton, this extended a familiar question: where should the work happen, and how should the parts connect?

His 2024 explanation of the shift included an unusually cheerful piece of executive arithmetic. Moving Akamai’s own applications onto its cloud was expected to save more than $100 million in annual operating expenses. He plainly enjoyed the prospect of sending less money to competitors. The professor had become fluent in a second kind of optimization, one whose results could be appreciated by the finance department.

On September 24, 2026, Akamai announced an expanded agreement with Anthropic: $11.6 billion in contractual commitments over seven years for infrastructure supporting CPU workloads. Akamai estimated about $5.5 billion in associated capital expenditures. These are commitments and planned investment across years. The scale gives Leighton a substantial execution task: buy capacity, build it, and operate it reliably enough to support a demanding customer.

Three days earlier, he had published an argument for keeping AI under human control. He called for checks and balances and for stronger defenses around systems that AI agents might attack. The proximity of the two announcements is revealing. His company is selling infrastructure to the AI industry while he is publicly arguing that the industry needs safeguards. Growth brings additional engineering work, including the work of limiting damage when systems are misused.

“For that reason, AI must always be kept under human control.”

Tom Leighton · September 2026

A television prize for the work behind the picture

In December 2025, Leighton received a Technology & Engineering Emmy Lifetime Achievement Award. It is a pleasantly unexpected addition to the biography of an algorithms professor. Television prizes usually suggest people on a screen. His contribution sits in the delivery system that helps the screen keep showing pictures. Streaming entertainment depends on engineering whose absence is often noticed more quickly than its presence.

The Emmy joined the 2018 Marconi Prize and the 2023 IEEE John von Neumann Medal. MIT lists more than 100 research papers across his career. His academic work also includes textbooks: Mathematics for Computer Science, written with Eric Lehman and Albert Meyer, and a text on parallel algorithms and architectures. His life has produced both explanations for students and systems that students use without ever seeing his name.

The teaching connection reaches beyond books. Leighton helped design MIT’s Mathematics for Computer Science subject and, with Michael Sipser, contributed to the Mathematics with Computer Science major. In 2016, their work on that major earned the Irwin Sizer Award for improvements to MIT education. Giving a useful idea a workable institutional home has been a recurring occupation.

Back to the student side of the table

Leighton is married to MIT mathematician and computer scientist Bonnie Berger. Their family’s support for mathematics includes the F. Thomson Leighton and Bonnie Berger Family Prize for STEM Excellence, a $50,000 annual scholarship award at ISEF. It recognizes students applying advanced mathematics and technology to significant challenges. The description could almost be an invitation to a younger version of its benefactor.

He has also supported the Akamai Foundation’s mathematics education work and the Akamai Technical Academy, which trains people from nontechnical backgrounds for technical careers. These efforts widen the places from which a future colleague might arrive. A teenager with access to a computer, a student with a good adviser, an adult learning a new technical skill: each needs someone prepared to make room.

The thread through Leighton’s story is the journey of an idea into use. He has spent years helping mathematics travel through classrooms, software, and commercial networks. Most people meet the result while doing something else: loading a page, watching a match, opening an application. For someone whose working life began with questions about algorithms, that ordinary moment is a rather satisfying place for an answer to arrive.