Thank you to our Sponsor: EezyCollab

For a while, the AI industry liked a simple story about openness. Some companies positioned themselves as open and developer friendly, while others leaned into proprietary systems and tightly controlled access. That distinction is getting harder to sustain. The market is moving into a more complicated phase in which many companies want the benefits of openness without giving away their strongest strategic advantages. The result is a new divide in AI: open enough to spread, closed enough to win.

This shift matters because openness in AI is not just a philosophical choice. It is a business model decision, a competitive strategy, a distribution strategy, and a governance decision all at once. Companies want their models, tools, and ecosystems to reach developers, startups, researchers, and global users. But they also want to protect their best capabilities, defend margins, manage safety risk, and avoid handing rivals an easy way to copy expensive breakthroughs.

Key points:
• The old open versus closed distinction is breaking down
• AI openness is now a strategic and economic decision, not just an ideological one
• Companies want wide adoption without surrendering their strongest advantages
• Hybrid strategies are becoming more common across the industry

Why Full Openness Is Getting Harder

The economics of frontier AI make full openness increasingly difficult. Building leading models requires enormous spending on chips, compute, data infrastructure, engineering talent, and distribution. As those costs rise, companies have stronger incentives to keep their most powerful systems under tighter control. If a company spends billions building a frontier model, it becomes much harder to justify giving away the full product in a form that others can freely copy, fine tune, or commercialize.

This is especially true now that AI is becoming more central to enterprise software, coding tools, search, customer support, and agentic systems. The leading models are no longer just research artifacts. They are revenue engines. That pushes companies toward selective openness rather than full transparency.

Key points:
• Frontier AI is extremely expensive to build and maintain
• Companies increasingly see top models as core commercial assets
• Open releases may continue, but often not at the absolute frontier
• Economic pressure is pushing the market toward selective openness

Why Companies Still Need Openness

Even with those pressures, companies cannot simply go fully closed without consequences. Openness still creates major advantages. It helps attract developers, win mindshare, shape standards, build ecosystems, encourage experimentation, and expand global reach. A model that spreads widely can become part of the industry’s default stack, which creates long term leverage even if the company does not capture all the value immediately.

That is why so many firms are reluctant to abandon openness altogether. Many now want a strategy that gives them both ecosystem influence and competitive protection at the same time.

Key points:
• Openness helps build developer loyalty and adoption
• Open models can shape ecosystems and technical standards
• Broad distribution can still create long term strategic power
• Many firms want openness for reach, but not for their crown jewels

Meta as the Clearest Example of the New Middle Ground

Meta may now be one of the clearest examples of this hybrid position. It built a strong reputation as a prominent U.S. company willing to let others modify advanced AI models, especially through the Llama family. But the company no longer appears committed to full openness across all tiers of capability.

That matters because it shows how the competitive environment has changed. Meta still wants the ecosystem advantages that come from openness, especially in consumer scale distribution and developer reach. At the same time, it appears to recognize that the largest models may carry more strategic value if they remain closed. This is the essence of the new AI divide. The question is no longer whether a company is open or closed. The question is where it chooses to be open, where it chooses to be closed, and why.

Key points:
• Meta reflects the move toward a more explicitly hybrid model strategy
• Open versions can preserve ecosystem reach
• Closed top tier models can protect competitive advantage
• Strategy now depends on choosing which layers to expose and which to protect

China and the Open Source Dilemma

This tension is not limited to U.S. companies. AI firms in China are also balancing broad ecosystem influence with the financial and strategic pressures of proprietary frontier development. As competition intensifies and monetization becomes more urgent, openness becomes harder to sustain in pure form.

Companies may still use open releases to build influence, but the more important the model becomes to revenue, agents, enterprise control, or cloud economics, the stronger the pressure to keep the best parts proprietary. That is why even firms long associated with openness are beginning to hedge.

Key points:
• The open versus closed tension is global, not just American
• Chinese AI firms are also balancing influence against monetization
• Agents and cloud economics increase the value of proprietary control
• Openness is increasingly selective rather than absolute

Why Safety Also Pushes Companies Toward Closure

Safety is another reason the industry is becoming more selective. As models improve in coding, reasoning, tool use, and cyber related tasks, companies are under more pressure to think carefully about what they release broadly. The more capable the system, the stronger the argument that unrestricted release could create misuse or governance problems.

That does not mean openness is inherently unsafe. But it does mean the safety debate increasingly overlaps with competitive strategy. A company can argue for tighter release practices on safety grounds while also benefiting commercially from keeping the strongest systems proprietary. These motivations are not always separate. In practice, safety, control, and profit often reinforce each other.

Key points:
• More capable models raise harder misuse and governance questions
• Safety concerns can justify more restricted releases
• Commercial and safety incentives often point in the same direction
• Closure is increasingly framed as both prudent and strategic

Thank you to our Sponsor: Partnerly

Why Hybrid Will Probably Become the Dominant Model

The hybrid approach is likely to become the dominant model because it solves several problems at once. It allows companies to keep feeding the developer ecosystem with tools, smaller models, or open weight releases while still preserving exclusivity where the competitive stakes are highest. It also lets firms present themselves as contributing to openness without fully exposing their leading edge.

That hybrid logic is already visible across the market. Even where companies still speak positively about openness, they are increasingly careful about which parts of the stack they open and which parts remain closed.

Key points:
• Hybrid strategies combine ecosystem benefits with competitive protection
• Companies can release useful models without surrendering the frontier
• This approach is more flexible than pure openness or full closure
• It is becoming the most practical compromise in a high-pressure market

What This Means for Developers and the Market

For developers, this new divide means more access than a fully closed market would provide, but less freedom than the early open source AI movement promised. Useful models will still spread widely. Open ecosystems will still matter. But the most powerful systems may increasingly sit behind APIs, enterprise contracts, managed platforms, or tightly controlled licenses.

For the broader market, this means competition will happen across multiple layers. Companies will compete on model quality, distribution, cloud integration, enterprise trust, developer loyalty, safety posture, and consumer reach. Being “open” will no longer be a simple identity. It will be a calibrated market position.

Key points:
• Developers will likely still get broad access to strong models
• The absolute frontier may increasingly remain controlled
• Competition will move beyond model release alone
• Openness will become a spectrum rather than a binary label

Why This Strategy Makes Sense

From a business perspective, the hybrid model is rational. Pure openness can weaken a company’s ability to capture value from its most expensive work. Pure closure can limit adoption, reduce ecosystem energy, and weaken developer loyalty. A mixed model offers a middle path.

It lets companies seed the market, shape the tools developers build on, and influence technical habits, while still reserving the most commercially important capabilities for direct monetization. In many ways, this resembles earlier technology markets, where firms opened some layers to drive adoption but kept the most valuable layers proprietary.

Key points:
• Hybrid strategies help balance adoption with monetization
• Companies want both influence and control
• Mixed openness can strengthen ecosystem reach without sacrificing margins
• The model mirrors earlier platform competition in tech

The AI industry is moving past the old symbolism of open versus closed. What is emerging instead is a more strategic balance in which companies want enough openness to spread their tools, influence the ecosystem, and attract developers, but enough closure to protect margins, control risk, and keep their strongest advantages from becoming commodities.

That is why the phrase open enough to spread, closed enough to win captures the moment so well. It describes an industry that still values openness, but values competitive leverage even more. It explains why companies that once championed broad sharing are now drawing new lines around their most important systems. And it suggests that the future of AI competition will not be defined by who is simply open or closed, but by who manages this balance most effectively.

Looking to sponsor our Newsletter and Scoble’s X audience?

By sponsoring our newsletter, your company gains exposure to a curated group of AI-focused subscribers which is an audience already engaged in the latest developments and opportunities within the industry. This creates a cost-effective and impactful way to grow awareness, build trust, and position your brand as a leader in AI.

Sponsorship packages include:

• Dedicated ad placements in the Unaligned newsletter

• Product highlights shared with Scoble’s 500,000+ X followers

• Curated video features and exclusive content opportunities

• Flexible formats for creative brand storytelling

📩 Interested? Contact [email protected], @samlevin on X, +1-415-827-3870

Just Three Things

According to Scoble and Cronin, the top three relevant and recent happenings

​​​​​​​​​​Meta Moves Toward a Hybrid Open Source AI Strategy

Meta plans to release new AI models developed under Alexandr Wang and eventually open source some versions, though its biggest and most advanced models may stay proprietary. The strategy suggests Meta is trying to balance broad developer access and consumer reach with stronger control over its most competitive systems. Axios

Jamie Dimon Sees an AI Future of Longer Lives and Less Work

Jamie Dimon says AI will eliminate some jobs, but he believes it will ultimately be a major positive force by raising productivity, improving health outcomes, and possibly reducing the workweek over time. He argues that AI could help people live longer and safer lives, even as current evidence suggests it may also increase pressure and workload for many workers. Fast Company

AI Is Reshaping Cancer Care

AI is already helping reshape cancer care by identifying patients for clinical trials earlier, improving drug discovery, and speeding up research documentation. This can expand access to treatment options, make oncology research more precise, and free scientists to focus more on developing better therapies for patients. Pfizer

Scoble’s Top Five X Posts

 Five X Posts