By Shameen Prashantham

Over the past six years, “liminality”, the unsettling in-between state, has become permanent. The three dimensions leaders need to seize opportunity must now evolve into ongoing strategic orientations, argues Shameen Prashantham.

In 2020, as the world wrestled with the pandemic, I argued that leaders were grappling with what anthropologists call “liminality” – an in-between threshold spanning a vanishing “before” and an unknowable “after.” Liminality, I suggested, represented both crisis and opportunity because it was a double-edged sword: confounding, because it strips away the familiar cognitive, social, and institutional scaffolds we depend on, yet liberating, because it opens the possibility of new ways of seeing, connecting, and growing.

Liminality on Steroids

Half a decade later, liminality is no longer a moment; it is the environment. We are leading in a world that feels persistently betwixt and between. The post-World War II geopolitical order has frayed, but its replacement remains stubbornly unclear. The digital revolution, supercharged by the rise of artificial intelligence (AI) is a prolonged, generative, and utterly unpredictable transformation. As such, the world feels like there’s liminality on steroids.

Unlike the covid era, which required leading through a liminal episode, today’s challenge is leading in a liminal world. Three dimensions of liminality – ambiguity, community and potentiality – remain the essential map. The corresponding leadership responses, however, must evolve from one-off crisis maneuvers into ongoing strategic orientations.

To see what this looks like in practice, consider two contrasting examples that bookend the liminality of our time.

Two navigators of our liminal world

In 2019, Microsoft made a $1 billion investment in OpenAI, then a small, research-driven organization. That initial bet has since evolved into a multi-layered, multi-billion-dollar partnership – at its core, a response to deep ambiguity. In the mid-2010s, no one knew which AI architectures would prevail. Rather than simply trying to forecast the future, Microsoft chose to set a series of options: investing in OpenAI, building its own internal AI research, integrating generative models across its product suite, and positioning Azure as the preferred cloud for AI workloads. It was a move made in the fog of liminality.

At the other end of the spectrum is Carl Breau, a Canadian entrepreneur. Breau has spent the last decade building and running a manufacturing SME business, specializing in LED lighting, out of China while recently acquiring an SME client of its manufacturing services in Canada. His world is one of geopolitical liminality – shifting tariffs, escalating rhetoric and politicized supply chains. Unlike Microsoft, Breu survives by reading it closely and adapting continuously. His story is one of quiet success through constant recalibration in the face of liminality.

These two cases anchor a three-part leadership agenda in a liminal world:.

1. Ambiguity: From Efficiency to Optionality

Pervasive liminality is characterized by ambiguity: the sense that old rules no longer apply and new ones haven’t yet been written. In a liminal world, the management reflex to optimize for efficiency becomes risky because efficiency presumes stability. Optionality, by contrast, preserves the ability to pivot as the landscape reveals itself.

Microsoft’s engagement with OpenAI exemplifies optionality as strategy. Rather than committing to a single technological path, it maintained multiple bets and built an ecosystem – spanning cloud infrastructure, developer tools, and applications – capable of capturing value across different possible futures.

Carl Breau’s approach reflects a grittier, entrepreneurial version of the same principle. By maintaining operations in China while establishing a North American base, he has created room to maneuver. When tariffs shift or political sentiment turns, he can reconfigure production and positioning.

Similar patterns are emerging across contexts. Firms under geopolitical pressure, such as Huawei, are being pushed to build technological and market alternatives – optionality not by choice but by necessity.

Some AI startups I have spoken to recently have very different teams operating out of different places to keep their options open of serving both the West and China. In one case, there were teams in San Francisco and Shanghai, in fact operating out of different legal entities, to take advantage of the strengths and opportunities of the respective ecosystems. In such circumstances Singapore stands out as a location that can become a platform for optionality, by deliberately maintaining openness across geopolitical divides.

Across these contexts, optionality tends to take three recurring forms: portfolios of products and business models, portfolios of places, and portfolios of partners.

2. Community: Partner Proactively, Without Passivity 

Liminality introduces fluidity in community, reshaping the social fabric of business – some bonds weaken but new ones may be possible, including partnerships among actors dissimilar to each other in geography, capabilities, or institutional logic. The key is to build such connections intentionally rather than retreat into familiar networks.

Microsoft’s relationship with OpenAI is again instructive. As I document in my book Gorillas can Dance, Microsoft had spent years building a startup partnering capability, but even they needed a different playbook to navigate the AI world. In 2019, Microsoft took the unconventional step of investing in an AI-focused organization that had been initially created as a non-profit. This was an intentional bridging of a highly dissimilar pair: a platform giant and a scrappy, mission-driven research lab. It was co-evolutionary in the sense that both actors changed in parallel, meaning that the partnership was itself a liminal space for them.

On a different scale, Carl Breau’s partnerships are more improvised but no less intentional. Operating a China-based business as a Westerner, and now straddling two national bases, he has had to become a professional insider-outsider, weaving trust in contexts where suspicion is the default. Breau’s partnerships with dissimilar others are forged with Chinese factory managers and North American clients. His approach is active: leaning into, rather than avoiding, geopolitical tension.

One of the new “gorillas” of the AI era, Nvidia’s Inception program for startups has a clear “building block” synergy as it equips thousands of startups with compute, tools, and technical support, embedding its architecture at the core of diverse ventures. The one-to-many interface is essentially that of a funnel (not a cohort), allowing startups to graduate to greater levels of support around go-to-market access and introductions to venture capital networks.

Additionally, third-party specialists continue to be prominent. One of the companies I had previously studied, Plug and Play, continues to operate as an ecosystem orchestrator, structuring interactions across a global network of startups, corporations, and institutions through accelerators, curated dealflows, and AI Centers of Excellence.

In each case, the emphasis is the same: community does not simply evolve – it must be constructed through deliberate, and often uncomfortable, engagement across boundaries.

3. Potentiality: Building an Institutional Capability

The third dimension of liminality is potentiality – the untapped human capability that can emerge when old competencies are no longer sufficient.

Capability learning now translates into a requirement for a new institutional capability: the capacity to continuously make sense of the macroenvironment, in particular the interplay of geopolitics and AI-driven digital disruption. Microsoft’s journey into AI is a story not just of a smart bet, but of an institutional sensing capability that developed over many years.

For Carl Breau, this capability is intensely personal and equally vital. He must be his own institutional sensor: tracking political discourse in China and North America, interpreting subtle shifts in customs enforcement, reading local business sentiment, and constantly asking, “What does this mean for my physical manufacturing capacity, my logistics, my brand?”

At the same time, capability building in a liminal world is not only about high-level sensing; it also unfolds in more grounded ways. C Park, a Chinese firm I encountered in Johannesburg, exemplifies this by positioning itself as a China-South Africa Digital Innovation Hub: it equips young South Africans with e-commerce and livestreaming skills while gaining access to local talent.

The African context also highlights another dimension of capability: coping with constraint. For instance, MTN Ghana’s recently announced ambition to support AI development, while very welcome, must realistically contend with limited access to 5G and reliable power, implying that innovation will often need to be frugal.

During a research visit to Cambridge I encountered Fellows at the Judge Business School working on a Centre for Frugal AI. Their premise is simple: in much of the Global South, the priority is not frontier performance but doing more with less. Capability building in a liminal world, therefore, is not only about advancing the technological frontier, but also about adapting it creatively under constraint.

Leading Without Closure

Liminality is disorienting, but it is also where novelty emerges. The shifts outlined – from efficiency to optionality, from passive positioning to proactive partnership, and from static expertise to continuous sense-making – form an operating logic for a world that resists closure.

Microsoft and Carl Breau differ vastly in scale and context, yet they share a common instinct: they do not wait for uncertainty to resolve. They move forward with options in hand, partners alongside them, and a disciplined attentiveness to emerging patterns.

About the Author

Shameen Prashantham is a Professor of International Business & Strategy, and Associate Dean, at China Europe International Business School. He is the author of Gorillas can Dance: Lessons from Microsoft and Other Corporations on Partnering with Startups.