Earlier this year, there were signs that IBM really might pull off a turnaround. After more than five years of shrinking revenues, Big Blue reported three quarters of growth in a row. Sure, that growth wasn’t coming from new businesses, but from milking the upside in existing businesses. But it was something.

Then, in October, IBM missed its numbers. Its revenue dropped 2.1%. It fell short specifically in its most critical area–the strategic imperatives that it has touted for several years as the key to its future.

Which is why IBM’s $34 billion proposed acquisition of Red Hat is so monumental. It’s the largest software acquisition of all time, and it stands out both for the business opportunity it presents, and for the Un-IBM-y nature of IBM’s approach to an acquisition. Nearly seven years into her tenure as CEO, Ginni Rometty wants to signal that IBM can be a different kind of company–that just as it has done for the past 107 years, it can change with the times.

The problem is that changing with the times is table stakes for the tech giants. To pull off the type of growth that can carry a behemoth corporate from one decade to the next, a company must change ahead of the times–and define that change for everyone else. While the Red Hat acquisition shows Rometty can be a creative strategist and is willing to act aggressively to keep IBM healthy and cement her own legacy, the company must now help Red Hat to thrive. More important, Rometty must identify new similarly large growth areas and position IBM to own them as they emerge.

It’s easy to hang IBM’s current woes on Rometty, a career IBMer who has been trying to turn the company around since two years after she assumed the top job. But the company’s current travails emerge largely out of an awful call made by her predecessor. In 2010, former IBM CEO Sam Palmisano told investors, “You can’t do what we’re doing in a cloud.” Amazon had already been selling cloud services for four years by then, and most of the West Coast computing giants evangelized the consumerization of technology and believed that everything would one day move to the cloud. Palmisano turned his back on the future.

By the time IBM embraced the cloud, paying $2 billion for SoftLayer Technologies and attempting to advance a hybrid strategy that allowed customers to combine on-site private servers and third-party cloud computing, it was too late. Amazon and Microsoft (which, disclosure, owns my parent company, LinkedIn) had already cemented their lead in pure cloud computing, establishing themselves as the go-to services. In 2017, Amazon owned 52% of the market for public cloud-computing, according to Gartner; Microsoft held 13%. IBM’s share was 1.9%. GBH Insights analyst Daniel Ives told me, “It’s a two-horse race at this point.”

With the Red Hat acquisition, IBM can expand its reach across its hybrid cloud infrastructure. The market is expected to more than double to $130 billion by 2022. As companies become more sophisticated about pairing the benefits of cloud computing and the security of managing their own infrastructure, this hybrid approach is becoming more attractive. This, according to Ives, gives IBM a shot at capturing the growing number of customers who want to combine public and private clouds. “They have a legit path to become a very disruptive player.”

But that’s assuming both that companies are going to want an alternative to the hybrid computing offered by the three large centralized players–Microsoft, Amazon, and Google–and that they aren’t sophisticated enough to harness existing open-source software to build their own alternative. Stratechery’s Ben Thompson isn’t so sure that’s true. He says this strategy “seems more attuned to IBM’s needs than potential customers.” He asks, “If an enterprise is concerned about lock-in, is IBM really a better option? And if the answer is that ‘Red Hat is open,’ at what point do increasingly sophisticated businesses build it themselves?”

For the Red Hat acquisition to work, Rometty must also honor her promise for Red Hat to remain independent within IBM. The plan is for the company to remain a standalone business unit, in the same way that, for years, Facebook gave Instagram room to nurture a distinct culture and grow an independent business. Red Hat’s leadership team will stick around, and it will remain in its current headquarters.

At a Town Hall meeting the day after the announcement, Red Hat CEO Jim Whitehurst assured his employees “red will stay red” as Rometty looked on. Employees were skeptical, both in the meeting and on the internet, where they went so far as to eulogize their company in reddit forums. But Whitehurst made a strategy case for the approach: “We have to stay separate because otherwise the value goes away,” he said. “If we lose our Switzerland status, that’s a lot of what our value is.”

Even if everything goes according to plan, IBM’s future is far from secure. Though IBM was early to develop artificial intelligence tools, and made the world take notice when Watson won Jeopardy in 2011, it hasn’t been able to realize the commercial success of its peers. Recently, sales in the company’s business that includes Watson have declined.

But IBM is no stranger to near death experiences. No large company, and certainly no tech company, survives more than 100 years without reinventing itself many times. Rometty was at IBM when, in the early 90s, it missed the shift to personal computers and came within a quarter of going bankrupt. She watched Lou Gerstner teach the proverbial elephant to dance. A successful IBM-owned Red Hat could send a powerful signal that even if IBM’s turnaround is slow to emerge, Rometty is capable of traveling far beyond IBM’s traditional playbook to reboot the company.