In this world, the starlight at the treetops is not about the size of the model, but about the influence of the clients.
Article author: Purpose Investments
Source: BitpushNews
Once again, it is that time of year where we gather product heads and fund managers to review the market performance of 2025 and look ahead to the prospects for the next 12 months.
This year, two major themes run throughout: despite the ongoing volatility, the institutional viability of digital assets continues to strengthen; meanwhile, large-cap tech stocks (Big Tech) remain at the core of market conversations. With these thoughts in mind, we interviewed Paul Pincente, Vice President of Digital Assets at Purpose Investments, and fund manager Nick Mersch, to discuss the tumultuous changes of the past year and potential trends for the future.
Nick Mersch (Fund Manager) view:
In my view, the AI boom is the most dazzling decoration on the 'Christmas tree' of the 2025 market. More importantly, it is no longer just a glowing entity. Now, real revenue, genuine workloads, and tangible budget items are flowing continuously toward AI. While hyperscalers are making aggressive investments in capacity, this is increasingly matching enterprise demand—these companies are moving from pilot phases to formal production. Customer support, code assistance, advertising spending, data analytics, and internal productivity tools are beginning to show measurable benefits. This is the key to transforming capital expenditures (CapEx) from a gamble into a long-term annuity income.
Yes, the scale of investment is staggering, from hundreds of billions in capital expenditure plans to massive power procurements. But this may also be building a foundational layer of computing power and energy that the entire economy may rely on for the next decade or even longer. Every new model, agent, and application launched in 2026 will require an operating environment, and those platforms that built capacity early will be the first to capture this spending. AI's performance resembles a wave of infrastructure and productivity that lasts for years rather than a fleeting carnival.
Is there a bubble risk in individual stocks? Absolutely. Some stocks are priced not only to reflect perfect expectations but even to an absurd extent. However, that is different from saying the entire AI cycle is a mirage. In 2026, investors will have the opportunity to engage in this long-term trend while maintaining cautious exposure. We advocate focusing on companies with the following characteristics:
Have clear monetization pathways;
Can convert AI capital expenditures into recurring revenue and cash flow;
Occupy key links in computing power, networks, energy, or distribution.
In my view, the magic has not disappeared; it is simply shifting from 'storytelling' to 'execution,' which will ultimately benefit disciplined stock pickers.
Paul Pincente (Vice President of Digital Assets) view:
First, this sense of turbulence is not your imagination, and feeling anxious is not a solitary experience. Cryptocurrencies have been like that passenger: insisting that the road conditions are fine while your teeth rattle like an overactive simulation bird.
But a bumpy road does not equate to a broken carriage.
In my view, the turmoil at the end of the year is less a solemn funeral bell and more a common and somewhat dramatic 'house cleaning' in high-beta asset classes. After a strong upswing, excess leverage must be purged, bubbles must be deflated, and the boldest narratives must withstand scrutiny under the cold light of real positions. This is the season when the market forces confidence to 'show receipts.'
So, will 2026 be smooth? We should be cautious in using that term. Cryptocurrencies are not calm lakes but a turbulent and branching river. However, I do believe its turbulence may become more mature. Not becoming docile, but becoming more understandable. This may mean fewer 'system about to collapse' headlines and more identifiable risk asset rhythms.
Moreover, there is a practical reason to remain optimistic. If macro conditions allow for a more supportive policy stance, and inflation performs well enough to avoid triggering new policy-induced panic, risk appetite tends to have a habit of returning. When it does return, cryptocurrencies rarely tiptoe back into the room; instead, they burst in like a band asked to perform an encore.
So, yes, we think you can foresee drama and tension, foresee the market turning overleveraged individuals into cowards and patient ones into philosophers. But we believe you should not mistake a bumpy December for a decade destined to fail. Sometimes, the bumps are merely the cost of a better road.
Nick Mersch's view:
Tech stocks are indeed in a full 'holiday showcase' mode in 2025, but beneath the flashing lights, we believe there is more substance than pessimists acknowledge.
The earnings revisions of key AI beneficiaries are generally trending upward, with improved margins as cloud and software scales. Many balance sheets remain in net cash or low leverage positions. I believe we are witnessing the largest capital expenditure cycle in tech history, with annual AI spending guided by hyperscalers reaching hundreds of billions of dollars. However, their starting point is a dominant market position and strong cash flow engines, which may offer greater buffers than in past cycles.
The key distinction in 2026 is whether this glow is supported by fundamentals or merely relies on the vanity of hope. I believe some software and smaller AI targets are clearly overvalued relative to current earnings. However, in core infrastructure, cloud services, and specific platform companies, valuations seem more like 'high but justified' rather than purely speculative.
Rather than framing it as a binary debate of 'bubble or not bubble,' we believe investors can view 2026 as a filtering mechanism. Those companies that demonstrate significant operational leverage on AI investments and rising marginal returns on capital will be able to maintain or even expand their valuation premiums. Meanwhile, those that only talk about gigawatts (GW) and GPUs but lack a roadmap for unit economics may be the most vulnerable when the cold wave arrives. We believe the shine can continue, provided there are earnings to support it.
Paul Pincente's view:
There's an interesting fact about markets: they often grow in the least conspicuous ways. Not through fireworks, but through pipelines; not through poetry, but through policies, filings, and the slow settling of institutional nerves.
So, when someone asks whether cryptocurrencies are ready to join the adult's table, I hear a question posed by someone in a tuxedo, but wearing practical shoes.
"The adult's table" is not reserved for the loudest believers but is prepared for assets that can be held without the need to constantly invent new rules. Investors want to know what the rules are, who enforces them, and where the exits are when the room gets crowded. This last point is crucial. Liquidity is always desirable on sunny days, but the real 'adult test' begins when the weather turns gloomy.
ETFs help not because they perform some magic to sanctify assets, but because they translate assets into the language that institutions already understand. Regulation, even if late, can alleviate the famous excuse of 'we are waiting for clarity.' Custody and operational standards, though dull, have become the dividing line between curiosity and actual allocation.
The story goes beyond Bitcoin. Stablecoins are quietly proving they are not just convenient tools for cryptocurrencies; they are becoming the rails for actual payments in the real world and real scenarios. Tokenization is moving step by step from pilot projects to the future, looking less like a marketing theater and more like infrastructure.
Will 2026 become genteel? No. Cryptocurrencies still love a grand entrance, and they might show up at the adult's table in a sequined gown. But the seating arrangements have changed. The guest list now includes more rule followers, more long-term allocators, and more capital that won't panic at the first sign of volatility.
This is how a market becomes difficult to destroy: not by becoming boring, but by becoming reliable enough to survive its own excitement.
Nick Mersch's view:
If the past three years have been about breakthroughs in models, then 2026 may evolve into a discussion about what these models can actually do in the real world. Tech companies are shifting from the race to 'build the largest systems' to the race to 'embed systems deeply into workflows.' This means agents capable of handling multi-step tasks, multimodal systems that integrate text/image/audio/video, and vertically tuned products that can understand specialized industry language. The opportunity lies not only in clever presentations but in the lasting productivity gains across white-collar work.
As we argued in July, the message for companies and investors is simple: adapt now or risk being left behind. We believe the winning organizations will not be those that release the coolest models but those that quietly replace manual processes with AI-first workflows, redesign products around intelligent assistants, and build governance, security, and data quality pipelines. This is where high stickiness income and higher switching costs will be generated.
For investors, we believe this will shift the focus to the upper tiers of the stack. Model providers remain important, but value may increasingly concentrate in three areas:
Orchestrating models and agents' platforms;
Application vendors that deliver specific business outcomes (such as increased sales productivity or reduced support costs);
Infrastructure and tools that ensure deployment is secure, observable, and compliant.
In this world, the starlight at the treetops is not the size of the models but the influence of customers. If a company can show shortened cycle times, fewer complaints, higher conversion rates, or lower unit costs due to AI, we believe the market will reward them. Otherwise, even the most dazzling decorations will eventually lose their luster.
