Happy New Year + Six Contrarian Predictions for 2026

The Future of Innovation Is Global. We discuss it here.

This post was previously published on my Forbes column.

Happy New Year!

It is prediction season again. Most forecasts focus on what will grow fastest or attract the most capital. This one takes a different approach.

These are predictions about where I think expectations are wrong. Looking ahead to 2026, here are six uncomfortable shifts I believe will matter most.

1. Cross-border expansion may slow, but cross-border replication will accelerate

AI has created the illusion that international expansion has become easier. Yes, you can easily translate your app into Mongolian.

In practice, the opposite is true.

The era of “launch everywhere” is facing more headwinds. For most business models: regulatory friction, capital discipline, and operational complexity are not bugs; they are features.

For AI companies in particular, durable moats will come not speed to launch in new markets, but ultimately from unique data or unique distribution. Thin application layers without either a unique distribution advantage (e.g. local partnerships, trustworthy brand etc) or data (e.g. on the local economics, relationships etc) will struggle to defend themselves as larger foundational models move up the stack.

Conversely, replication of proven business models will accelerate dramatically. There are now hundreds of cities with unicorn founders and effectively no barriers to launching new companies. If Uber took a decade to be replicated globally, AI-enabled business models may follow the same arc in three years.

A critical skill for both founders and investors will be saying no to shiny international expansion. Depth in a home market will often be greater than breadth across markets. Context, local execution, and restraint will become real sources of competitive advantage.

2. AI will enable “camel seed-strapping” and distort the early-stage market

AI is allowing small teams to get much farther with much less capital. Many companies will reach real revenue, real customers, and operational maturity before they ever raise a traditional Seed or Series A. I wrote about the “camel seed strapping” phenomenon recently.

Early-stage benchmarks will become harder to interpret. Later-stage investors will encounter companies that look “too early” culturally but “too late” operationally. Some will already be profitable. Others will be highly capital-efficient but mismatched to traditional venture expectations.

As a result, two phenomenon may take place in tandem: early-stage businesses will skip seed funding, but also many good businesses may never need to raise a Series B. Durability and capital efficiency may change the venture business model.

3. AI churn is coming, and those with valuation discipline will be rewarded

Yes, a small number of AI companies will continue to reach for the stratosphere.

But for many, pilot revenue won’t convert as fast as expected to long-term contracts. Switching costs may be lower than expected as competitors emerge. Foundational models will continue to converge in functionality. Customers will start to reassess AI investments and real ROI post experimentation

Many business models will work. But in the short-term, churn may increase. The companies may need more time to mature the product or the sales cycle to create enduring customer relationships.

At the next raise, if the story is not perfect, a previous round at 100x revenue may come back to bite startups with more challenging round dynamics.

In that environment, companies that managed valuation will be better positioned. Those that reach profitability and raise on their own terms will have the most leverage.

4. Venture returns and recognition are continuing to shift outside Silicon Valley

In 2026, the Midas List will continue going global. Returns, talent density, and category-defining companies are emerging outside traditional tech hubs.

What continues to lag is recognition. League tables will adjust slowly. Geography is no longer a proxy for quality, and in many cases it has become a source of edge. The next generation of enduring venture firms will be built around where insight is deepest, not where consensus is loudest.

5. M&A is returning

M&A is coming back as a deliberate strategic tool.

Companies will play offense and defense at the same time: acquiring distribution rather than users, buying teams instead of tools, and consolidating fragmented categories before pricing power erodes.

We may have seen the first such deal with Nvidia’s $20 billion dollar ‘licensing deal’ of Groq.

I expect a diverse range of active acquirers. Foundational models looking to grow vertical capabilities. Scaled technology companies looking to grow capabilities. And incumbents fearful of being left behind.

For founders, M&A may increasingly be a feature of exit strategy, and certainly not an admission of defeat.

6. The line between tech and services is blurring

The software-only ideal for startups will become more blurry.

We already see this shift in practice: forward-deployed engineers at companies like Palantir and OpenAI.

One theme I wrote about recently are AI roll-ups, which merge an M&A playbook (per above) with a technology platform. AI-enabled roll-ups borrowing from private equity playbooks, and vertical software companies embedding operations to deliver outcomes rather than tools. Some successful models are already emerging: Teamshares, a company that combines SMB M&A with employee ownership and various fintech products, will be going public via SPAC next year.

Elsewhere, some valuable AI businesses will look in part service-enabled at first. The winners will be those that use services as a wedge, not a crutch, on the path to scalable advantage.

---

As capital, talent, and technology rebalance, the biggest opportunities will belong to founders and investors willing to trade noise for nuance, and speed for depth. That’s where the 99% lives.

Happy New Year.