In short
TechCrunch Disrupt 2026 will feature Ambrosia Energy’s Ben Longmier and Bloom Energy’s Bill Thayer in a session on how AI infrastructure constraints are creating new business opportunities. The discussion will focus on power, data centers, cooling and other physical systems that are becoming central to AI scaling.
- AI scaling is increasingly limited by physical infrastructure, not just software or chips.
- Disrupt 2026 will feature a session on which AI infrastructure bottlenecks could become durable business opportunities.
- Energy, data centers, cooling and electrical systems are emerging as key areas for founders and investors.
- The next major AI winners may be companies that solve real-world power and compute delivery problems.
AI’s biggest constraint is no longer just model quality or chip supply — it is the physical infrastructure needed to keep rapid growth running. That shift is why Ambrosia Energy CEO Ben Longmier and Bloom Energy SVP Bill Thayer will appear at TechCrunch Disrupt 2026 on October 13 to discuss where the AI infrastructure boom is creating real business opportunities.
The session matters because the companies that win in the next phase of AI may be the ones solving power, cooling, grid and data center problems, not only building the most visible software products.
What began as a software revolution is becoming a hard-asset race. Every new model, agent and application needs electricity, a place to run, and a delivery system that can move power and compute together at scale. That reality is turning infrastructure from background support into one of the most competitive layers of the AI economy.
Why AI infrastructure has become the new battleground
The AI boom has pushed attention toward frontier models, specialized chips and consumer-facing applications. But none of those layers can function without a much larger physical stack underneath them. Data centers need land, grid interconnections, power generation, backup systems, cooling and electrical equipment. As AI workloads grow, each of those inputs becomes more strained.
That strain is the core issue the Disrupt 2026 session aims to explore. If demand for compute rises faster than utilities, permitting and construction can respond, shortages will not remain temporary. They can reshape who gets access to capacity, which regions attract investment and which startups can survive.
Tech leaders and investors are increasingly asking a simple but consequential question: when infrastructure becomes scarce, where does the value move?
How the infrastructure shortage changes the market
The answer, in part, is that constraints create pricing power and specialized demand. When energy, cooling or data center capacity becomes harder to secure, companies that can deliver reliable solutions gain leverage. That can produce durable businesses in areas that previously looked like ordinary industrial services.
Instead of treating infrastructure as a cost center, many in the AI industry are beginning to view it as a strategic advantage. Faster access to power or compute can determine who launches products first, whose models train on schedule and which hyperscale operators can keep expanding.
What will Ben Longmier and Bill Thayer discuss on stage?
They will focus on where the current infrastructure buildout is creating the strongest opportunities, which bottlenecks are most likely to persist and what categories founders should watch as AI becomes increasingly tied to the physical world.
Longmier, who leads Ambrosia Energy, brings a perspective centered on energy systems. Thayer, who heads datacenter solutions at Bloom Energy, offers a view from the power and onsite generation side of the market. Together, they represent two sides of the same emerging problem: AI needs far more dependable power and supporting infrastructure than legacy systems were designed to provide.
“The industry stage session will look at which constraints are creating durable business opportunities, where new categories are beginning to emerge, and what founders can build as AI moves further into the physical world,” the event description says.
That framing signals a broader conversation than a typical technology panel. The focus is not only on what AI can do in software, but on what it takes to physically deploy it at scale.
Where are the real opportunities in the AI stack?
The biggest opportunities are likely to appear in the less glamorous layers of the AI economy. Those include power generation, grid connection management, cooling systems, data center design, electrical gear, infrastructure software and other tools that help AI facilities operate efficiently.
Not every bottleneck becomes a lasting market. Some shortages are temporary and resolve as supply catches up. Others persist long enough to create entirely new categories of products and services. The challenge for founders is telling the difference early.
That is one reason the session brings together leaders across energy and datacenter solutions. A wider view can help identify which parts of the stack are likely to see sustained spending, rather than one-time bursts of interest.
What founders should watch next
Founders building around AI infrastructure are being pushed to think beyond the familiar software layer. The most promising ideas may not resemble a conventional AI startup at all. They may look more like industrial technology, power management, infrastructure orchestration or systems software built for physical assets.
For investors, the question is which infrastructure markets can support defensible companies rather than brief demand spikes. For operators, it is which systems can remove friction from a process that is becoming more capital-intensive every quarter. And for startups, the opportunity may lie in translating AI demand into new tools for a world where power is as important as code.
How AI growth is pushing companies into the physical world
AI’s expansion is increasingly exposing the limits of existing infrastructure. The industry can create models quickly in software, but scaling them requires slow, expensive and highly coordinated physical development. That mismatch is now one of the defining tensions in the market.
Data centers cannot simply be spun up overnight. Utility upgrades take time. Cooling systems must be engineered for sustained heat loads. Electrical interconnects have to be negotiated, approved and built. These realities create friction, but they also create openings for companies that can help solve them.
In that sense, the next wave of AI winners may not be known primarily for AI research or consumer chat products. They may be the firms building the backbone that makes those products possible.
| Topic | What it means | Why it matters |
|---|---|---|
| Power generation | Supplying enough electricity for AI workloads | Limits how fast compute can expand |
| Grid connections | Linking facilities to utility capacity | Can delay new data center projects for months or years |
| Cooling systems | Removing heat from servers and equipment | Essential for stable, high-density AI operations |
| Data centers | Physical homes for models and compute | Core venue where AI infrastructure demand is concentrated |
| Electrical equipment | Transformers, switchgear and related hardware | Often a hidden bottleneck in expansion plans |
Why this conversation matters to founders, investors and operators
The AI infrastructure buildout affects the entire startup ecosystem, not just large cloud and chip companies. Founders need to know where capacity is available. Investors need to know which constraints are durable enough to support new businesses. Operators need to understand which inputs are becoming strategic.
When a market enters an infrastructure phase, the winners are often the companies that can manage complexity, reduce delays and make scarce resources more usable. That can lead to opportunities for startups far outside the obvious AI application layer.
There is also a geographic dimension. Regions with better access to land, energy and transmission can become more attractive for data center development. That could reshape where AI-related capital flows, where startups locate and which communities see new industrial investment.
Who benefits when infrastructure becomes scarce?
Companies that can solve bottlenecks often benefit first, but the ripple effects are broader. Utilities, energy developers, equipment makers, software vendors and construction firms can all gain from the rising need to connect, power and maintain AI systems. The most successful players will likely be those that can deliver dependable capacity at scale.
At the same time, scarcity can expose weaknesses in existing business models. If power or cooling cannot be secured quickly enough, even well-funded AI projects may face delays. That makes infrastructure not just an enabler, but a gatekeeper for the next generation of AI deployment.
What is TechCrunch Disrupt 2026 offering around the session?
The event is positioning the discussion as part of a larger program featuring more than 250 speakers across more than 200 sessions. The Ambrosia Energy and Bloom Energy conversation will take place on the Smart Systems Stage at Moscone West on October 13.
TechCrunch is also using the event to encourage early registration, noting that attendees can save up to $100 before doors open and receive a discount on a second pass. For startups, founders and investors, the appeal goes beyond the pricing. Disrupt remains one of the industry’s most visible gatherings for networking, dealmaking and market intelligence.
The broader message of this session is that AI’s next phase will not be defined by software alone. It will be shaped by the systems that make software scale in the real world.
A market where the next AI company may not look like an AI company
The emerging infrastructure opportunity is changing how the industry defines an AI business. A company solving cooling inefficiencies or grid access may prove just as important as a company building a new model or agent product. In some cases, those physical-world businesses could be the most durable firms of all.
That is because infrastructure tends to compound. Once a data center, power setup or electrical system is in place, it can support multiple generations of AI demand. Businesses that help build that backbone may benefit from recurring need, long planning cycles and high switching costs.
The session featuring Longmier and Thayer will likely explore that exact dynamic: not whether AI is growing, but how that growth is reshaping the industrial systems around it.
For the industry, the key takeaway is clear. The AI race is now an infrastructure race, and the companies that understand the physical limits of scaling may be the ones best positioned to capture the next wave of opportunity.
Event snapshot
- Event: TechCrunch Disrupt 2026
- Session: “Where the AI Infrastructure Boom Is Creating Winners”
- Speakers: Ben Longmier, CEO of Ambrosia Energy; Bill Thayer, SVP and Head of Datacenter Solutions at Bloom Energy
- Stage: Smart Systems Stage
- Date: October 13, 2026
- Location: Moscone West
Frequently asked questions
What is the TechCrunch Disrupt 2026 AI infrastructure session about?
It is about where the AI infrastructure boom is creating business opportunities. The session will examine how power, grid access, data centers, cooling and related systems are shaping the next phase of AI growth and where founders may be able to build durable companies.
Who is speaking at the Disrupt 2026 session?
Ben Longmier, CEO of Ambrosia Energy, and Bill Thayer, SVP and Head of Datacenter Solutions at Bloom Energy, will take part in the discussion. Their combined backgrounds bring perspectives from energy systems and data center power solutions.
Why is AI infrastructure such a big issue now?
AI systems need far more electricity, cooling and physical capacity than many existing facilities were built to handle. As demand for compute grows quickly, infrastructure constraints can delay deployments, raise costs and create new opportunities for companies that solve those problems.
What kinds of startups could benefit from the AI infrastructure boom?
Startups working on power generation, grid connection tools, cooling technology, electrical equipment, data center software and infrastructure management could benefit. These businesses help remove the bottlenecks that are limiting AI deployment at scale.
When and where is the session taking place?
The session will take place on October 13, 2026, at Moscone West during TechCrunch Disrupt 2026. It will be held on the Smart Systems Stage as part of the conference’s broader speaker program.
