The Other Microsoft Founder: Paul Allen and the 1975 Legacy

If the question is who, officially, was the other Microsoft founder alongside Bill Gates, the short answer is: Paul Allen. But the name alone hides the best part of the story. It was Allen who saw the January 1975 cover of Popular Electronics, convinced Gates to drop out of Harvard, and wrote the code that made BASIC run in Albuquerque. The rest is worth knowing.

TL;DR

• The other Microsoft founder is Paul Gardner Allen (1953–2018), Gates's partner since their days at Lakeside School in Seattle.
• The company treats April 4, 1975 as its official founding date, in Albuquerque, New Mexico.
• The name "Micro-soft" first appears in writing in a letter from Gates to Allen dated July 29, 1975, recorded on Microsoft's own official timeline.
• Allen left the company in 1983, after being diagnosed with Hodgkin lymphoma, and died on October 15, 2018, at age 65.
• His technical legacy lives on in 2026 AI: the Allen Institute for AI (Ai2) publishes the Olmo family with open weights, data, and code.

Who was the other Microsoft founder, according to official documentation

Paul Gardner Allen. Born in Seattle in 1953, two years older than Bill Gates, his classmate in the Lakeside School computer club and partner in an earlier venture before Microsoft. When the duo sold their first program, Gates was 19 and Allen was 22 — Microsoft itself describes the episode in historical material published on Microsoft Learn.

People searching for the other Microsoft founder often run into three wrong names before finding the right one. First, Steve Ballmer: he joined in 1980, became CEO in 2000, and held the role for 14 years, but he was never a founder. Second, Ed Roberts of MITS, the maker of the Altair 8800 — he was the first customer, not a partner. Third, Monte Davidoff, the Harvard student who wrote the floating-point package for Altair BASIC: a crucial contributor, but with no equity stake.

None of the three is the other Microsoft founder in the corporate sense. The official answer is unambiguous: alongside Bill Gates, the other Microsoft founder was Paul Allen. And the division of labor between the two is what explains why the company succeeded.

1975 in order: the timeline Microsoft itself records

The chronology below comes from the company's historical documentation and shows what the other Microsoft founder did at each stage of that year. It's useful because it debunks a common myth: the company wasn't born from a business plan; it was born from a product sold before the partnership formally existed.

Date	What happened
January 1975	The MITS Altair 8800 appears on the cover of Popular Electronics. The article convinces Allen and Gates to write a BASIC for the machine.
February 1975	The two complete Altair BASIC and sell it to MITS in Albuquerque — their first customer. It's the first computer language program for a personal computer.
March 1975	Paul Allen joins MITS as director of software.
April 4, 1975	Official founding date of Microsoft, as a partnership between Gates and Allen, in Albuquerque.
April 7, 1975	"Altair BASIC – Up and Running" is the headline of the first issue of MITS Computer Notes.
July 1, 1975	The duo's BASIC is officially distributed as version 2.0, in 4K and 8K editions.
July 22, 1975	Allen and Gates sign the BASIC interpreter licensing agreement with MITS. The name "Microsoft" doesn't exist yet.
July 29, 1975	In a letter to Paul Allen, Gates uses "Micro-soft" — the earliest known written reference.
December 31, 1975	The year's revenue closes at $16,005, according to the partnership's Form 1065 filing.

Notice the sequence: product sold in February, partnership formalized in April, name defined in July, five-figure annual revenue in December. No unicorn started differently — what changes is the yardstick of who tells the story later.

The name, incidentally, is a fusion of microcomputer and software, and the credit for the idea is usually attributed to Allen. The hyphen fell away the following year. In 1979, the company moved from Albuquerque to Washington state, near home.

What the other Microsoft founder did that Bill Gates didn't

Here's the part the quiz doesn't ask, and it matters to anyone building a technical team.

The opportunity-window radar

Allen was the trend reader. It was he who brought the magazine, who identified that cheap microprocessors would create a software market for individuals, and who insisted the window was closing fast. The second founder's role in this phase was to detect timing — a skill that rarely appears on a resume and that decides a product's fate.

The engineer of the improbable

The technical detail is beautiful: the duo wrote the BASIC interpreter for a computer they had never touched, using an emulator of the Intel 8080 processor running on a PDP-10 minicomputer. Allen wrote the loader that placed the program into the Altair's memory. The demo worked on the first try, in Albuquerque, in front of the customer. Anyone who has deployed code in production without identical staging knows the size of that risk.

The partnership counterweight

Gates was the relentless negotiator of contracts and licenses. Allen was the architecture and next-product guy. The company worked as long as the two roles balanced. When the imbalance came — with illness and equity dilution — the partnership broke.

Why the second Microsoft founder left in 1983

Allen was diagnosed with Hodgkin lymphoma in 1982 and left Microsoft's operations in 1983, retaining his equity stake. He spent the next three decades doing other things: he bought the Portland Trail Blazers (NBA) and the Seattle Seahawks (NFL); funded neuroscience research with $500 million for the Allen Institute for Brain Science; and donated more than $2 billion over his lifetime.

He died on October 15, 2018, at age 65. The immediate cause was septic shock, a complication related to the non-Hodgkin lymphoma he had been treated for since 2009.

There's an uncomfortable lesson here for tech business partners in the U.S.: a partnership agreement without a clear clause for health-related departure, without dilution rules, and without exit provisions is a ticking time bomb. I've seen this happen in at least three agencies and software houses I've worked with in 17 years at Agathas Web. A hastily written operating agreement costs more than any tech stack.

Allen's legacy is in the AI your company uses in 2026

This isn't rhetorical homage. The other Microsoft founder's money funded one of the most relevant research institutions in the industry. The Allen Institute for AI, Ai2, based in Seattle, is today one of the few institutions that publishes fully open language models — weights, training code, checkpoints, and datasets. The Olmo family follows this policy, and the Olmo 3.1 Think 32B and Olmo 3.1 Instruct 32B versions are distributed with training and fine-tuning data available without license restrictions.

Why does this matter for a U.S. company? Because full openness is what allows you to audit bias, run the model on your own infrastructure, and avoid being held hostage by a vendor's price or policy changes. When we evaluate a model for a customer service or document classification project, the criterion isn't just the benchmark: it's what happens to the client if the API price jumps 40% or gets discontinued.

Yes, that has happened — and it's not rare. We treat third-party dependency risk as an architecture issue, not a legal detail.

Microsoft in August 2026: Copilot superapp and data centers in São Paulo

While the name of the other Microsoft founder circulates again in searches, the company he helped create is in the middle of two concrete changes.

Copilot becomes a single app. On August 13, 2026, Microsoft confirmed it will unify its assistants into a single application called Microsoft Copilot, with a new icon. The migration began on mobile and web in mid-August and reaches desktop, on Windows and macOS, in mid-September. Two features were retired on August 18, 2026: the podcast tool and Deep Research. Groups, messages, and extra content from those areas are not migrated. If your operation used Deep Research in a research routine, the deadline has passed — export what exists and review the workflow.

Infrastructure in Brazil. On February 11, 2026, at the Microsoft AI Tour, Microsoft Brazil's president, Priscyla Laham, announced that two cloud and AI data centers are already operating in the state of São Paulo, with construction completed in January. The exact location was not disclosed for security reasons. The investment is part of the R$ 14.7 billion package announced in September 2024, to be applied by 2027, along with the goal of training 5 million Brazilians in AI.

Also in 2026, Microsoft created Microsoft Frontier Company, a $2.5 billion unit with 6,000 specialists to deploy AI for corporate clients, led by Brazilian executive Rodrigo Kede Lima.

And Bill Gates, at 70, continues publishing: on August 15, 2026, he returned to the topic of AI applied to health, citing WHO data on the shortage of nearly 6 million health workers in Sub-Saharan Africa. Half a century after the Altair, the two founders continue to set the agenda — one through what he writes, the other through what he funded.

What Microsoft's founding teaches U.S. SMBs

The decisions made by Gates and the other Microsoft founder in 1975 still fit into a 2026 business plan. Five lessons that survive the 51-year test:

1. Sell before building the structure. BASIC was sold in February; the partnership came in April. Validating demand before setting up an LLC, hiring a team, and renting an office remains the right order.
2. Make little money without shame. $16,005 in the first year. Those who expect a spectacular first year usually don't reach the second.
3. Asymmetric duos win. One partner who sees the market and another who closes contracts is a better combination than two identical ones.
4. Licensing is more profitable than a one-time sale. Microsoft licensed BASIC to MITS instead of selling the code outright. Fifty years later, it's exactly the same reasoning behind any SaaS.
5. Document the partnership. Allen's exit shows that agreements need to be written down before life happens.

This fifth point has direct resonance in the U.S. in 2026, with the growth of digital operations that start informal and only formalize after generating revenue.

How we apply this logic in Agathas Web projects

I've been a full-stack developer for over 15 years and founded Agathas Web in 2008 — long enough to have seen three complete hype cycles. My reading of Microsoft's founding story is practical: technology becomes a business when someone turns technical capability into recurring contracts.

That's what we do on three fronts. In official WhatsApp API integrations, we set up the direct channel with Meta instead of going through a platform with per-message markup — the same principle of licensing rather than reselling. In Moodle and e-learning environments, we deliver Linux, Redis, and cloud infrastructure scaled for real access peaks, not for the vendor's brochure. In applied AI, we evaluate model by model, including open alternatives like those from Ai2, before tying a client to a single provider.

If you're evaluating how AI actually enters your company's operations, I've written about what changes in practice for U.S. companies and about AI agents in corporate routines. And if the topic is the infrastructure behind it, Microsoft's recent shift to Linux servers shows how platform decisions affect the final cost.

Conclusion: the right name and the right lesson

Alongside Bill Gates, the other Microsoft founder was Paul Allen — and the one-word answer is the least useful part of the story. What's worth retaining is the method: two complementary profiles, a product sold before the company existed, licensing instead of a one-time sale, and $16,005 in first-year revenue.

If your operation is at that stage of turning technical competence into predictable revenue, start with the basics: choose a product that can be licensed, document the partnership, and don't outsource the architecture decision. It's a conversation we enjoy having — and it usually yields more than any search trend.