There’s a quote from Steve Wozniak — the engineer who co-founded Apple and personally designed the Apple I and Apple II — that should be required reading for every leader running an AI transformation in 2026.
In an interview, Wozniak said: “The Apple III was a failure, the Lisa was a failure, and the Macintosh was a failure. It was only by modifying the Macintosh hugely and over time that we made it a good computer.”
The co-founder of Apple called the Macintosh — the product that supposedly defined the company, the one with the legendary 1984 Super Bowl ad — a failure.
And the historical record backs him up completely.
I want to walk through this history accurately, because the real story is far more useful than the myth we’ve built around it. We remember Apple as a company that strung together one triumph after another. The truth is almost the opposite. Apple’s early history is a sequence of expensive, public, near-fatal failures — and the company’s greatness was built not on avoiding those failures but on surviving them and iterating through them.
That distinction is the single most important lesson available to anyone leading an AI transformation right now.
The Apple III (1980): when committees and marketing run product
After the runaway success of the Apple II, Apple needed a follow-up aimed at the business market. The result was the Apple III, and it was the first Apple computer not designed by Steve Wozniak.
Instead, it was designed by committee — a group of Apple engineers defining specifications, with marketing executives shaping the requirements. This is the detail Wozniak returns to most pointedly. In his telling, the Apple III failed because the wrong people were in charge. “If the guys at Apple had built the machine that they would love, it would have been successful. It came instead from formulas from Apple executives. Marketing people were in charge and some very bad decisions got made.”
One of those decisions became infamous. Steve Jobs wanted the Apple III to run silently, so he insisted it have no cooling fan and no vents. Engineers tried to compensate with an aluminum case meant to dissipate heat. It didn’t work. The machine overheated. Chips worked loose from the motherboard. Floppy disks reportedly melted inside the drive.
The Apple III had, by Wozniak’s account, a 100% hardware failure rate. Apple was forced to recall and replace every one of the first 14,000 units produced.
It was Apple’s first commercial failure, and it put the company into genuine financial uncertainty. The business market Apple had wanted to capture went instead to the IBM PC, which launched the following year.
The lesson the Apple III taught Apple: when committees and marketing formulas drive product decisions instead of the people who deeply understand and love the product, you get a machine nobody loves — and sometimes one that literally melts.
The Lisa (1983): when revolutionary technology is priced out of reach
The Lisa was, in pure technology terms, a breakthrough. It was the first commercially produced computer with a graphical user interface controlled by a mouse — the paradigm that would eventually define all personal computing.
It also cost $9,995. In 1983 dollars.
That price proved prohibitive for almost everyone. The Lisa sold around 100,000 units total before being discontinued — a catastrophic number for a product that had consumed enormous development resources. And it was almost immediately eclipsed by the cheaper Macintosh, which launched a year later with much of the same conceptual DNA at a fraction of the price.
Jobs himself was blunt about it afterward. In a 1985 Playboy interview, he said: “First of all, it was too expensive — about ten grand.”
The lesson the Lisa taught Apple: revolutionary technology that the market can’t afford isn’t a triumph. It’s an expensive lesson. Being first and being right about the future means nothing if the price puts the product out of reach.
The Macintosh (1984): the failure we remember as a triumph
This is the one that matters most, because it’s the one we’ve completely mythologized.
The Macintosh launched in January 1984 with one of the most famous advertisements in history — the “1984” Super Bowl ad directed by Ridley Scott. The launch event, with Jobs unveiling the machine to a roaring crowd, is the stuff of corporate legend. We remember the Macintosh as the moment Apple changed computing forever.
What we don’t remember is what happened next.
Apple’s business plan had assumed the Macintosh would sell at enormous scale. One early plan projected 2.2 million units between 1982 and 1985 — roughly 47,000 units a month. Apple built manufacturing capacity to match that ambition.
After the initial launch frenzy faded — once the early adopters and developers had bought theirs — real sales collapsed. By late 1984, the Macintosh was selling around 5,000 units a month. The factory Apple had built for massive volume sat nearly idle. Hundreds of millions of dollars of unsold inventory accumulated across the supply chain. Dealers stopped placing orders.
The product had real problems. Its 128K of memory wasn’t enough to do meaningful work. It had almost no software — for much of 1984, MacPaint and MacWrite were essentially the only applications. At $2,495 with a single disk drive, even basic tasks meant endless, frustrating disk-swapping. The press, after the initial excitement, dubbed it an impressive but expensive toy.
The market, by the end of 1984, had figured out that the IBM PC — less elegant, less friendly — actually did the things people needed to do: spreadsheets, word processing, databases. The Mac was beautiful and could do almost none of it.
This near-collapse is what ended the first era of Apple. The sales slump, combined with internal conflict over resources diverted from the Apple II division, led to a power struggle that Jobs lost. In 1985, he was stripped of operational responsibilities and resigned from the company he co-founded. Wozniak left the same year.
And here’s the detail that should reframe everything: the Macintosh didn’t actually become the Macintosh — the reliable, capable, beloved machine — until the Macintosh Plus in 1986. Two years after the famous launch. Several brutal failures later. Long after the founders were gone.
What everyone gets wrong about this story
The myth of Apple is a story of visionary triumph. The reality is a story of serial failure survived through relentless iteration.
The Apple III failed. The Lisa failed. The original Macintosh failed. By the count of the man who co-founded the company, Apple’s three major products following the Apple II were all failures.
The greatness we attribute to Apple wasn’t the absence of failure. It was the capacity to keep going through failure — to treat each expensive, public, near-fatal mistake as an input to the next attempt rather than as a reason to stop.
The Apple III taught them what happens when committees and marketing run product. The Lisa taught them what happens when you price revolutionary technology beyond reach. The original Macintosh taught them that a beautiful interface means nothing without memory, software, and a genuine reason to buy. Each failure was tuition. Each lesson got paid forward into the next product.
This is not a detour from the path to greatness. This is the path to greatness, and almost every great company’s real history looks like this once you strip away the mythology.
Why this is the most important lesson in AI transformation today
Every leader running an AI transformation in 2026 is operating in an environment defined by failure. AI pilots fail constantly. Use cases that looked promising turn out not to work. Models hallucinate in front of executives. Deployments get pulled. Vendors underdeliver. The failure rate of enterprise AI initiatives is, by every available measure, extremely high — McKinsey’s 2025 research found only 5% of companies capturing substantial value at scale, with 60% generating no material value despite real spending.
The question isn’t whether your AI program will produce failures. It will. The question is what you do with them.
Most leaders treat each AI failure as evidence that the strategy is wrong. As something to hide from the board, to assign blame for, to avoid repeating. The failed pilot becomes a source of organizational shame, the program quietly defunded, the people associated with it subtly sidelined. The lesson the organization absorbs is: don’t try things that might fail.
Apple’s actual history says this instinct is catastrophic.
If Apple had treated the Apple III the way most enterprises treat a failed AI pilot — as a reason to fire the team, retreat from the business market, and stop experimenting — there would be no Lisa, no Macintosh, no iMac, no iPhone. The willingness to keep failing forward, in public, at enormous cost, was the greatness. Not a flaw in the greatness. The substance of it.
The companies that win at AI over the next decade won’t be the ones that avoid failed experiments. That’s not possible; the technology is moving too fast and is too poorly understood for anyone to bat a thousand. The winners will be the ones who survive enough failures to iterate into something that works — who treat the failed pilot the way Apple eventually learned to treat the Apple III: as expensive, valuable, necessary tuition.
The kind of leadership this requires
This demands a specific kind of leadership that’s rarer than it should be.
It requires a leader who treats a failed AI pilot as tuition, not as evidence of incompetence — who can extract the lesson without needing a scapegoat.
It requires a leader who can stand in front of the board after a public failure and say “here is what we learned, and here is the next version,” instead of quietly killing the program to protect their own reputation. The instinct to bury failures to preserve career capital is precisely the instinct that would have killed Apple in 1981.
It requires a leader who understands, deeply, that the products and programs we hold up as triumphs almost never arrived as triumphs. They arrived as failures that someone had the conviction and the institutional support to iterate through.
And it requires an organization that rewards intelligent failure rather than punishing it — because in an environment where failure is punished, people stop running the experiments that the breakthroughs depend on. The organization optimizes for the appearance of competence and stops producing the actual competence that only comes from iterating through real failure.
The deeper coaching lesson
There’s a personal dimension to this that matters for any senior leader.
The Apple story is usually told as a story about Steve Jobs’s genius. But the more honest reading is that Jobs’s early career was a sequence of expensive failures — the Apple III’s fan, the Lisa’s price, the Macintosh’s near-fatal launch — that got him removed from his own company.
The genius we celebrate came later. It came after the failures. It came, by most accounts, partly because of them. The Jobs who returned to Apple in 1997 and built the iMac, the iPod, and the iPhone was a different leader than the one who got pushed out in 1985 — chastened, matured, and far better at the balance between vision and the practical reasons people actually buy things.
The failures didn’t interrupt his path to greatness. They produced the version of him capable of greatness.
For senior leaders navigating AI right now, this is the most important thing to internalize. The failures you’re experiencing in your AI program aren’t evidence that you’re not cut out for this moment. They’re the mechanism by which you become the leader this moment requires. Provided — and this is the entire condition — you stay in the game long enough to let the lessons compound.
The closing thought
We tell ourselves stories about great companies and great leaders that edit out the failures. We remember the Macintosh launch and forget the near-bankruptcy that followed. We remember the visionary Jobs and forget that his early vision produced a computer that melted, a computer nobody could afford, and a computer that nearly sank the company and got him fired.
The edited version of the story is comforting and useless. It suggests greatness arrives fully formed in the hands of people who simply see further than the rest of us.
The real version is uncomfortable and essential. Greatness is what’s left after you survive a string of failures that would have stopped most people and most companies. The Apple III. The Lisa. The original Macintosh. Each one a failure. Together, the foundation of everything that came after.
In 2026, your AI transformation is going to produce its own Apple IIIs and Lisas. Pilots that overheat and melt. Initiatives priced beyond what the organization will bear. Beautiful capabilities that lack the substance to be useful yet.
The question isn’t whether you can avoid them. You can’t.
The question is whether you’ll treat them the way Apple eventually learned to — as tuition, as inputs, as the road itself — or whether you’ll treat them the way most organizations do, as reasons to stop.
Wozniak was right. The Macintosh was a failure.
And it became one of the most important products in history anyway — because the people building it kept going.
The world has changed. The leaders who notice will be the ones the next decade is built around.
