Larry Ellison Named His First Product Version 2. The Lie Was the Strategy.

When Oracle shipped its first commercial database, it called it Version 2. There was no Version 1. There had never been a Version 1. Years later Ellison said so himself, flatly: 'There never really was an Oracle Version 1.'² The reasoning was pure Ellison — no serious enterprise buyer would risk its data on version one of anything, so the number was simply skipped. Before the product was finished, the marketing had already lied, and the lie worked. That is the entire doctrine in a single act: tell the market what it needs to believe, then build toward it.

The official story is that Ellison is a visionary technologist who built one of the great franchises in software. The truer story is that Ellison's playbook was never about invention at all. It was about seizing someone else's idea, winning the war for attention by any means available, and then buying up whatever survived. It made Oracle enormous. It also nearly killed the company — twice over the same aggression.

Three founders, one legend, and a borrowed idea

Start with what the legend quietly deletes. Oracle was co-founded in 1977 by Larry Ellison, Bob Miner, and Ed Oates — three names, not one — incorporated as the unglamorous Software Development Laboratories in Santa Clara.¹² The relational-database concept the company commercialized was not theirs; it came out of IBM research. Ellison took direct inspiration from Edgar F. Codd's 1970 paper "A Relational Model of Data for Large Shared Data Banks," and the founders discovered that no company had yet committed to commercializing the technology.⁹ Because IBM was initially unreceptive to Codd's ideas — protective of existing hierarchical products — vendors including Oracle were able to bring relational databases to market well before IBM did.¹⁰ Oracle's first real move was to take a published idea and ship it as a product faster than its originators would. That is the founding pattern, and it never changed: the doctrine begins by claiming territory someone else mapped. That is the founding pattern, and it never changed: the doctrine begins by claiming territory someone else mapped.

By March 12, 1986, the strategy had a public scorecard. Oracle went public at $15 a share — one day before Microsoft's own IPO — with Ellison holding roughly 39% of the company.³ The combative founder now had a war chest and a stock price to feed. And feeding it is exactly what went wrong next.

The aggression that built Oracle is the aggression that nearly buried it

Here is the thesis a smart friend can repeat at dinner: Ellison's doctrine isn't a strategy, it's a temperament — seize, outmarket, acquire — and the same temperament that drives the wins drives the disasters, because it has no brake. The cleanest proof is 1990. To keep the growth story intact, Oracle began booking revenue when a contract was signed rather than when the product was delivered. The court record is blunt about what that meant: second-quarter 1990 net income overstated by about 30%, roughly $15 million in improper contracts, and a forced restatement that cut fiscal-1990 revenue by more than $50 million, to $916 million.⁴ The hunger for a bigger number on the scoreboard had simply outrun the business underneath it.

The bill came due fast. The stock lost roughly 80% of its value. Oracle posted its first-ever annual loss — $12.4 million in fiscal 1991. Nervous bankers slashed its credit line from $170 million to $80 million in January 1991.⁵ This is the part popular tellings garble: Oracle never filed for bankruptcy, and the crisis had nothing to do with the 1986 IPO four years earlier. It was a self-inflicted wound from an aggression that recognized no limit. And yet the same company that booked phantom revenue secured a fresh $100 million revolving credit line in October 1991 and clawed its way back.⁵ The recklessness and the recovery were the same muscle, flexed in two directions.

When the product can't win, win the story

Watch the doctrine operate when the technical reality is inconvenient, and the mechanism becomes obvious. In the benchmark wars of the era, the test results did not cooperate. According to Roger Sippl of Informix, his company's product beat Oracle on 27 of 30 benchmarks; Oracle's response was an ad campaign titled 'Oracle wins again!' — citing the three benchmarks it had won, and saying nothing about the rest.⁶ The ads went further, depicting a jet fighter shooting down competitor biplanes and naming IBM and dBASE as targets. And when a University of Wisconsin benchmark proved unfavorable, Ellison allegedly threatened to have the professor fired.⁶ If the account is accurate, the instinct it reveals is telling: honest marketing spins results; this doctrine tries to suppress the source. When the evidence is the enemy, attack the person producing it.

The pattern held for decades. In 2002 Oracle launched the 'Unbreakable' campaign — 'Can't break it, can't break in.' Within two weeks, security researchers including David Litchfield, Alexander Kornbrust, and Cesar Cerrudo publicly demonstrated working attacks against Oracle products. The company didn't retract the claim; its security chief simply reframed it after the fact as a reference to third-party evaluations rather than a literal promise.⁸ Same move as 1990, same move as 'Version 2': make the bold assertion first, let reality catch up, then quietly redefine what you said. The marketing always leads the truth, and the truth is expected to follow.

Acquisition as a weapon, not a strategy

By the 2000s the doctrine had a final form: stop competing on product, buy the competition outright. On June 9, 2003, Oracle launched a hostile bid for PeopleSoft — striking precisely while PeopleSoft was mid-acquisition of J.D. Edwards, catching a rival with its hands full.⁷ The public framing was a war with SAP. The truer read is consolidation for its own sake. A month into the fight, Ellison reportedly told an analyst meeting the quiet part out loud: that Oracle would be interested in buying almost anything.⁷ Not a target — a posture. The takeover took 18 months and closed in January 2005, an acquisition as an act of attrition rather than a thesis about what the combined company should become.

This is where the doctrine reveals what it is. Seizing the relational-database idea required a real product to ship behind the claim. The benchmark wars only papered over a product that was, at the time, genuinely losing. And acquisition is what aggression does when it can no longer out-build a rival: it out-spends one. The throughline isn't innovation. It's the refusal to lose the fight in front of you, whatever the fight happens to be.

But it worked — doesn't that settle it?

The honest objection is the strongest one: Oracle became one of the largest software companies on earth, Ellison ran it as CEO until September 2014 and stayed on as chairman and CTO, and the man is among the wealthiest people alive.¹ If the doctrine were really broken, the scoreboard would show it. Fair. But notice what kept saving the doctrine from itself. It survived 1990 not because aggression solved the crisis but because Oracle still had a genuinely good database underneath the bad accounting — a real product cycle to recover into.⁴⁵ Strip that away and the same behavior is just the 'Unbreakable' campaign at company scale: a confident claim with nothing behind it. The doctrine doesn't produce the wins. It produces leverage, and leverage cuts both ways. The reason it keeps requiring rescue is that it is structurally incapable of providing its own.

Larry Ellison's genius was never the database; three people built that, and the underlying idea came from Codd's published research — commercialized by Oracle before IBM acted on its own researcher's work.⁹¹⁰ The genius was the willingness to declare victory before it was true and then bend the world until it caught up — to ship a Version 2 that had no Version 1, and dare anyone to check. It built an empire and it nearly torched one, using the identical reflex. That is the thing to understand about the combative founder: the aggression isn't the strategy. The product is the strategy. The aggression is only how loud the outcome gets.