In the summer of 2007, the two biggest game console makers simultaneously discovered that they were paying not to win, but for the right not to lose.
July 5, 2007 — Microsoft’s vice president of interactive entertainment, Peter Moore, published an open letter to Xbox 360 owners. The wording was gentle. The substance was not. The company acknowledged the existence of a widespread hardware defect and announced an extension of the warranty to three years for all consoles affected by what it called a “general hardware failure” — the very thing users had already known for two years as the Red Ring of Death. In its quarterly report for the period ending June 30, 2007, Microsoft recorded pre-tax charges of $1.05 to $1.15 billion. This was not some abstract accounting line: the money was going toward repairing, replacing, shipping, and servicing millions of devices already sold. Robbie Bach, Chris Liddell, Frank Brod, and Moore himself took part in the announcement — executives whose names beneath such an admission meant that, inside the company, the argument between “cover it up” and “admit it” was over.
To the industry, it looked like voluntary amputation. The Entertainment & Devices division, which had operated at a loss for years in pursuit of the console market, had only just begun to approach operating profitability — and now more than a billion had been written off in a single decision. But the alternative was worse. By the summer of 2007, forums, retailers, and the tech press had turned three red lights into a meme, a symbol, an argument against buying one. Warranty-monitoring service SquareTrade recorded a defect rate of 16.4% in its samples — and those were figures any journalist could verify. Microsoft’s internal estimates, according to engineers, pointed to numbers even higher: up to 30% for early revisions of the platform codenamed Xenon. One in every three console owners risked seeing the red ring — and one in every three prospective buyers could choose a competitor instead.
Three red segments on the Xbox 360’s ring indicator — the fourth remained dark — signaled a “general hardware failure.” The designation was deliberately vague, and that was the problem for anyone trying to explain RROD with a single cause. The red ring was not a diagnosis but a symptom — like a person running a fever, which might mean the flu or appendicitis. The console was saying: “something has gone catastrophically wrong,” without specifying what.
Engineering investigations conducted both within Microsoft and by independent specialists — Leo Del Castillo is often mentioned among the latter — uncovered a whole bouquet of interconnected problems. Under load, the graphics processor (GPU) heated up to temperatures the motherboard handled poorly. The board warped — literally bent out of shape — and the lead-free solder connecting the chips to the substrate began to fail. Contacts cracked, connections broke, the console died. The thermal design of the early models was, in essence, a time bomb: every heating-and-cooling cycle brought the moment of failure closer.
But reducing it all to “overheating” means simplifying the story in exactly the way dozens of retellings did in the years that followed. Overheating was both trigger and accelerator here. The GPU heatsink mounting design created uneven pressure on the die. The choice of thermal-interface materials, the geometry of the heat sinks, even the placement of ventilation openings in the case — all of it added up to a system in which each component might have worked on its own, but together they created a cumulative effect. I like the comparison one repair technician once used: RROD was not a bullet but a slow strangulation. Internal connectors degraded gradually, and a console could keep running for months before the accumulated microdamage reached critical mass.
Later myths — “Microsoft knew and released it anyway,” “one specific chip was to blame,” “all you had to do was wrap the console in a towel” — took hold precisely because the real picture was inconveniently complex. The truth is that the Xbox 360 was designed under a brutal deadline: Microsoft wanted to get to market before Sony, and it did — in November 2005, a full year ahead of the PlayStation 3. The cost of that head start was built into the board’s very design.
A billion dollars is a lot even for a company that, in 2007, made tens of billions from Windows and Office. The decision to extend the warranty was cold calculation, and to understand it, you have to remember the position Microsoft was in.
By the summer of 2007, the Xbox 360 had a year’s head start over the PlayStation 3 and an enormous advantage in its game library. Xbox Live was up and running, bringing in subscriptions and building an ecosystem. Developers preferred the 360 — the architecture was easier to understand, the tools more convenient, the installed base larger. All of that would turn to smoke if a widespread defect scared buyers away. And it was scaring them away: retailers reported a flood of returns, warranty centers were overwhelmed, and forums were generating a volume of negativity no marketing budget could overcome. The pressure came from all sides at once: from below (users), from the side (retailers, who had no interest in selling a product with a high return rate), and from above (the press and analysts).
In competitive terms, the situation was even more acute. The Nintendo Wii, released almost alongside the PS3, was selling at a phenomenal pace and taking the casual audience. Sony, despite a rough start, had the PlayStation brand — synonymous with console gaming for an entire generation. Had Microsoft responded to the RROD crisis with half-measures — extending the warranty by one year rather than three, or limiting the program to certain regions — the reputational blow might have become irreversible. Losing the trust of the core audience in the middle of a console war meant writing off the next generation too: people burned by the 360 would not have bought the next Xbox.
So Robbie Bach and his team chose the painful but strategically sound path: admit the problem publicly, pay now, preserve the base for the future. The program applied to every console sold since launch — retroactively, including those whose owners had already paid for repairs out of pocket. Those people were reimbursed. It was an expensive gesture, but it worked: a community ready to revolt got an answer that was hard to call insufficient.
While Microsoft fought the consequences of haste, Sony was paying for the opposite sin — excessive ambition. The PlayStation 3 launched in November 2006 with a set of components that looked more like a workstation spec sheet than a game console. The IBM Cell Broadband Engine processor, estimated by iSuppli at $89, was architecturally unique: one main processing unit and seven of the eight physically present synergistic processing elements running in parallel. Its theoretical power was impressive. Developing for that architecture drove programmers accustomed to more traditional setups mad.
The Nvidia Reality Synthesizer graphics processor cost $129 apiece. The Blu-ray drive — $125. A Seagate hard drive — $54 in the 60 GB version and $43 in the 20 GB version. When iSuppli analysts tore the PS3 down into its components and calculated the bill of materials, or BOM, the result was disheartening: the 60 GB model, sold for $599, lost $241.35 on every unit sold. The cheaper 20 GB model at $499 was even more unprofitable: minus $306.85 per box.
A caveat is necessary here, one that is often left out. iSuppli’s figures are estimates — based on teardown analysis and market prices for components. Sony’s actual manufacturing costs may have differed: the company had its own fabs, long-term supplier contracts, different logistics. Sony never published the exact production cost of each PS3. But the company’s financial reports confirm the broad picture: the gaming division was drowning in losses.
For the fiscal year ending March 31, 2007, Sony’s game segment posted an operating loss of ¥232.3 billion — roughly $1.97 billion at the exchange rate of the time. It was not a typo. Nearly two billion dollars in losses from a division that was supposed to become the engine of the company’s future. Sony’s total operating profit for that same year fell 68.3%, to ¥71.8 billion ($608 million) — even as the company’s overall sales rose 10.5%, to ¥8.30 trillion. Put crudely, Sony was selling more and making radically less, and the main culprit was the PlayStation 3.
The Blu-ray drive deserves a conversation of its own. Sony put it in the PS3 not for games — or at least not only for games. A format war was underway: Blu-ray versus HD DVD, backed by Toshiba and, notably, Microsoft (the Xbox 360 offered HD DVD as an external accessory rather than a built-in component). For Sony, Blu-ray’s victory meant control over the physical-media standard for the next decade — licensing royalties, influence over Hollywood, a strategic advantage. The PS3 was a Trojan horse: the cheapest Blu-ray player on the market, which also happened to play games. The strategy worked — Blu-ray defeated HD DVD by early 2008 — but the cost of that victory was embedded in every console sold.
Step back and look at the situation from above, and the symmetry is striking. Microsoft and Sony made opposite mistakes that led to a similar result — financial losses in the billions of dollars.
Microsoft rushed. Its one-year head start over the PS3 gave it a genuine advantage: by the time its competitor arrived, the Xbox 360 already had an installed base, a game catalog, and a functioning online service. But the rush to design the hardware meant that the console proved unreliable at a level unacceptable for a mass-market consumer product. Thirty percent failure rates in early models — that is the kind of figure at which any other product would have been recalled. Microsoft held on only because the console kept working for most owners long enough for them to fall in love with the ecosystem before they encountered the red ring.
Sony did not rush — and that became a problem too. The PS3 arrived a year later, priced at $499–599 in an era when the Wii sold for $249 and the Xbox 360 was already offering models for $299. The Cell architecture, for all its computing power, frightened developers away: porting games from Xbox 360 to PS3 took substantial effort, and many studios either made inferior PS3 versions or did not release them at all. The result: a slow buildup of the library, slow sales, slow reductions in losses. Sony was essentially asking buyers to pay a premium for technology whose advantages were not obvious in the console’s first years.
What surprises me is how precisely these two cases illustrate the classic “fast, cheap, good — pick two” dilemma. Microsoft chose “fast” and “cheap for the consumer,” sacrificing quality. Sony chose “good” — in the sense of technological ambition — and “fast” — in the sense of deploying Blu-ray — sacrificing an affordable price. Both paid in billions.
Both companies spent the next several years patching the holes — each its own.
Microsoft moved through iterations. The motherboard revision codenamed Falcon, which appeared in late 2007, shifted the hardware to a finer 65 nm process node, replacing Xenon’s original 90 nm process. A smaller process meant lower power consumption, less heat, and consequently less stress on the solder and the board. Failure rates fell to roughly 10% — still high by consumer-electronics standards, but no longer catastrophic. Subsequent revisions — Jasper (2008) and especially Valhalla/Kronos, which combined the CPU and GPU on one die in the Xbox 360 S model (2010) — brought reliability down to acceptable levels. The 2010 Slim model physically could not display a red ring: Microsoft removed the ring indicator and replaced it with a single light. A gesture containing as much engineering logic as symbolic desire to close the chapter.
Sony took another route — cost-cutting. The company gradually stripped features from the PS3 that made it expensive but that few people used. Hardware backward compatibility with the PS2 — the Emotion Engine chip — disappeared from the European model as early as its March 2007 launch, replaced by software emulation, then removed altogether. The number of USB ports fell from four to two. The card reader disappeared. Every such simplification shaved several dollars off the manufacturing cost. At the same time, the Cell itself became cheaper: production at 65 nm, then 45 nm, lowered die costs. Blu-ray drives, an exotic component in 2006, became mainstream. By 2009, Sony released the PS3 Slim at $299 — almost half its launch price — and finally began approaching break-even on the hardware itself.
Sales curves reflected these changes. The PS3, which started painfully slowly, gradually picked up momentum as its price dropped and strong exclusives appeared. The Xbox 360, having survived its reputational crisis, held on to its core audience thanks to the warranty program and continued to dominate the North American market so crucial to Microsoft. By the end of the generation, in 2013–2014, both consoles finished with comparable worldwide sales: the Xbox 360 at around 84 million, the PS3 reaching 87.4 million by 2017 — while the Wii surpassed both with more than 100 million, albeit under an entirely different business model.
The console race of the mid-2000s taught a lesson both companies paid for with interest: in this industry, you can win on specs, win on timing, win on sales — and still lose money. Microsoft entered the generation with a time advantage and paid more than a billion for a defect that nearly wiped out that advantage. Sony entered with technological superiority and a Blu-ray strategy, lost nearly two billion in its gaming division in a single fiscal year, and spent three years chasing its own price point.
Both companies survived. Both released the next generation of consoles. Both learned lessons — the PS4 got a standard x86 architecture instead of the exotic Cell, while the Xbox One went through far more extensive reliability testing. But here is what seems truly remarkable to me: in 2007, at the point of greatest strain, neither Microsoft nor Sony could afford to leave the race. Leaving would have meant losing the ecosystem, losing developers, losing future licensing and subscription revenue. The billions in losses were, in essence, the price of admission — or, more precisely, the price of not losing the ticket they had bought a generation earlier.
Nintendo, during that same period, was selling the Wii at a profit on every console, using previous-generation components and staying out of the specs war. Which of the three made the right choice depends on what horizon you are looking at, and on what you consider victory.