Welch Told Him to Leave. The Man Who Walked Out Is Now Sitting on a $109 Billion Thermal Empire

(SeaPRwire) –

By: Christian Pierce

Jack Welch approached David Cote in a GE dining room in 1999. He told Cote he wanted him out by year-end. Cote kept asking what he had done wrong. Welch never answered. That silence speaks louder than any written performance review ever could. It is the sound of institutional power deciding your future without offering a single reason. Most executives who receive that kind of message never fully recover. The industry tells you to file the paperwork quietly and start looking elsewhere. But Cote did something that defies the standard executive playbook entirely. Three years later, he became CEO of Honeywell. He spent fifteen years engineering a corporate revival that delivered shareholder returns 150% greater than the S&P 500. Then he retired in 2018. Retirement did not last very long. He came back because the industrial cooling infrastructure underneath the AI boom had quietly become one of the most critical bottlenecks in global technology. Almost nobody on Wall Street was paying attention to it at the time. The growth deadlock is not in compute. It is not in networking or storage hardware. It sits in the thermal layer that keeps all of that expensive silicon alive. Every hyperscaler engineer knows this reality at a visceral level. Every enterprise data center operator lives it daily. But the public markets spent years treating cooling as a low-margin commodity afterthought. That blind spot is precisely what Cote found and weaponized for his own advantage. Most growth stories on Wall Street chase the shiny object at the top of the stack. The real money in this cycle is hiding in the basement.

Cote retired in 2018. He teamed up with Goldman Sachs to hunt for an acquisition target. They evaluated more than 1,000 companies before selecting Vertiv. The company was an eighty-year-old cooling business entirely new to Cote personally. His rationale was deceptively straightforward. Digital data was growing much faster than the physical data centers needed to support it. He positioned Vertiv directly in the middle of that expansion curve. The bet looked prescient on paper. Then reality arrived faster than anyone expected. COVID hit just three weeks after Vertiv’s NYSE debut. A subsequent surge in orders exposed a brutal pricing failure that went unnoticed during planning. The company had been underpricing its equipment by 20% to 30%. Profits cratered immediately and visibly across the board. By early 2023, the stock had fallen roughly 55% from its previous high. Cote stepped in heavily on day-to-day management operations. The board installed a new CEO at roughly the same time. Vertiv kept investing in direct-to-chip liquid cooling despite the margin collapse around it. It acquired cooling startup CoolTera in late 2023. Then demand from AI data centers took off in earnest. Production ramped aggressively across the entire supply chain. The market cap surged from under $11 billion to $109 billion since going public. The sequence matters far more than the headline number alone. Vertiv nearly failed in its first year of public life. It survived because Cote refused to treat the setback as a permanent verdict on the thesis. The underpricing was a execution error. The underlying demand curve was never wrong. That distinction is what separates durable infrastructure plays from speculative bubble names that evaporate on contact with scrutiny.

There is a pattern hiding inside this arc that most analysts miss entirely. Welch pushed Cote out without offering any explanation whatsoever. Vertiv nearly collapsed during its first year as a publicly traded company. In both cases, the setback looked definitive at the time. Neither one was. The cooling infrastructure market is now absorbing the overflow from an AI infrastructure buildout that outpaced its own internal capacity models. Vertiv did not win because of some sudden product breakthrough. It won because it was the right company in the right place when the thermal bottleneck finally broke wide open. Cote understood that dynamic before the broader public market did. The industry now faces a structural question that most investors have not yet begun to ask. Who owns the thermal pipes underneath the compute layer. Historically, it has been a scattered collection of mid-cap vendors with no dominant player. That era is now over. Industrial infrastructure consolidates rapidly around whoever can scale manufacturing capacity first. The capital markets have already signaled their verdict with brutal clarity. One company went from under $11 billion to $109 billion on a market cap basis since its IPO. The remaining competitors are now fighting over shrinking margins on the same demand wave. The practical takeaway for any infrastructure investor is straightforward. When you evaluate an infrastructure play, ignore the product demo entirely. Look at the thermal load curve. Look at who holds the installed base when the next demand spike arrives. That is where the capital actually flows. The next consolidation wave in data center infrastructure will not center on GPU suppliers. It will center on whatever keeps those GPUs from melting under sustained AI workloads.

Author bio: Christian Pierce, a chief financial columnist and markets commentator covering industrial infrastructure, corporate governance, and the hidden value chains beneath AI capital cycles.