Why Gen Z’s Entrepreneurship Moment is Different – Issue #31
Spotlight: Why Gen Z’s Entrepreneurship Moment is Different
This week’s Spotlight is: Why Gen Z’s Entrepreneurship Moment Is Different
Today, a new generation is doing what took my generation courage to even consider – starting a new company. And they are doing it faster, with fewer resources, and at a younger age than ever before.
Two things have changed everything.
First, AI has democratized who gets to build. For thirty years, entrepreneurship required deep expertise, technical skills, capital, and connections. Today, anyone with an idea and a conversation can build a product.
Second, for this generation, entrepreneurship is no longer an unconventional choice – it is a cultural norm. Starting a company is as natural a path as joining one, and the idea of building something from scratch is celebrated in classrooms, on campuses, and across every platform they grew up on.
My advice to every young founder thinking about taking the leap:
The best time to start is when you have the least to lose and the most energy to give. Don’t wait until you are ready. You will never feel ready. Jump and build your parachute on the way down.
Find your people first. The idea will change. The market will change. The people you surround yourself with are the only constant.
Build a painkiller, not a vitamin. Solve a problem that keeps people up at night. If customers can live without your product, they will.
And remember – company building is a marathon, not a sprint. AI may be accelerating everything around you, but the fundamentals have not changed.
This generation, armed with AI, is a combination no previous generation of founders has ever had. I am more bullish on the future of entrepreneurship than I have ever been. #SmallBizMonth
This week’s signals show why Gen Z’s entrepreneurship moment is different. OpenAI is offering every current Y Combinator startup $2 million in AI tokens, putting frontier infrastructure directly into the hands of young, early-stage founders. AWS is giving AI agents desktop access through Amazon WorkSpaces, making it easier to automate legacy enterprise workflows without waiting for APIs or modernization. And Modal Labs raised $355 million as demand grows for the inference infrastructure and developer sandboxes that AI-native teams use to build, test, and scale faster.
Full Weekend Edition below. 👇
Signals Shaping the Future of AI:
Infrastructure
Google expands its AI chip strategy through a partnership with a Blackstone-backed AI cloud provider. The deal includes a reported $5 billion investment to help bring 500MW of new data center capacity online next year. Click here
NextEra and Dominion agree to combine into the world’s largest regulated utility platform as AI data-center load reshapes U.S. power-sector scale. The transaction makes grid capacity, utility balance sheets, and large-load interconnection core AI infrastructure control points. Click here
PJM moves to accelerate data-center power procurement as AI load strains the largest U.S. grid. The change turns grid access and bilateral power contracting into near-term infrastructure bottlenecks for AI data-center expansion. Click here
Applied Digital signs a long-term AI cloud hosting agreement tied to hyperscale GPU infrastructure expansion. The deal reflects continued migration toward specialized AI infrastructure operators that provide dedicated power, cooling, and GPU capacity outside traditional hyperscaler footprints. Click here
Enterprise
NVIDIA beats quarterly revenue expectations and expands shareholder returns. The company announced an $80 billion stock buyback program, raised its dividend, and reaffirmed strong demand for AI infrastructure and next-generation Rubin chips. Click here
OpenAI and Dell partner to bring Codex into hybrid and on-prem enterprise environments. The collaboration moves coding agents closer to governed corporate data, codebases, and operational systems, expanding Codex from developer assistant into enterprise agent infrastructure. Click here
EY rolls out enterprise-scale agentic AI across its global audit platform. The deployment embeds multi-agent systems into regulated assurance workflows, moving agentic AI from productivity tooling into audit operations with large-scale data-processing requirements. Click here
Google introduces Gemini Spark, a persistent agentic assistant integrated across Gmail and Google’s broader productivity ecosystem. The launch pushes consumer AI interfaces toward continuously operating agents capable of long-horizon task execution, signaling a shift from reactive chatbots toward always-on orchestration layers embedded across personal and workplace workflows. Click here
TD launches agentic AI for end-to-end real estate secured lending. The rollout brings autonomous workflow automation into mortgage and HELOC processing, showing regulated financial institutions moving agents into core credit operations. Click here
Capital Flows
OpenAI offers every Y Combinator startup in the current batch $2 million in AI tokens. The program gives startups infrastructure credits in exchange for future equity exposure, deepening OpenAI’s position inside the next generation of AI-native companies. Click here
Hark raises more than $700 million at a $6 billion valuation to build vertically integrated personal AI models, software, and hardware. The round shows capital moving into AI-native devices and interface layers as investors search for consumer hardware platforms beyond phones, wearables, and chat surfaces. Click here
Modal Labs raises $355 million at a $4.65 billion valuation. The company, which provides AI inference infrastructure and developer sandboxes for AI-generated code, has seen rapid growth as demand for AI coding and compute capacity accelerates. Click here
DeepInfra raises $107 million for production-scale AI inference infrastructure. The financing points to inference becoming the system constraint as open models and agentic workloads drive continuous high-volume compute demand outside traditional cloud platforms. Click here
Dust raises $40 million to scale “multiplayer AI” for governed human-agent collaboration across enterprises.The round reflects demand for shared agent workspaces with permissions, auditability, memory, analytics, and integrations that turn individual prompting into organization-wide workflow infrastructure. Click here
Research
OpenAI says an internal model autonomously disproved a central conjecture in discrete geometry. The result moves AI-assisted science from benchmark performance into original research contribution, indicating that general-purpose reasoning systems can now produce expert-validated mathematical discoveries. Click here
Google introduces Gemini for Science, a set of experimental AI tools designed to accelerate core steps of the scientific method. The tools support hypothesis generation, computational discovery, and literature analysis, pushing AI from research assistance toward agentic scientific workflows that can help generate, test, and synthesize new ideas. Click here
Cloudflare says frontier cyber models are beginning to industrialize vulnerability discovery and exploit development at machine scale. The company’s Project Glasswing findings suggest frontier-model competition is expanding from general reasoning into offensive and defensive cybersecurity capability, where exploit discovery, triage, and infrastructure resilience become core AI deployment constraints. Click here
Researchers publish evidence that multi-agent AI systems outperform human teams on creativity tasks. The work suggests agent collectives may become a practical research and ideation substrate for product development, scientific exploration, and enterprise innovation workflows. Click here
Policy
The Trump administration prepares an executive order expanding AI oversight as national-security concerns rise around frontier systems. The effort signals growing pressure to formalize federal review, evaluation, and governance mechanisms for advanced AI models tied to security-sensitive deployment risks. Click here
California lawmakers push disclosure requirements for AI-generated political campaign media as synthetic election content proliferates. The effort expands operational compliance pressure around AI-generated advertising and synthetic communications ahead of the 2026 election cycle, increasing scrutiny of generative media in political distribution channels. Click here
The FTC begins enforcing the TAKE IT DOWN Act’s platform-removal requirements for nonconsensual intimate imagery, including synthetic abuse workflows. The enforcement start creates a federal takedown compliance regime for covered platforms, turning AI-enabled intimate-media abuse into an operational legal obligation. Click here
Millville, New Jersey votes to ban new data centers, halting a proposed 1.4GW campus. The ordinance shows AI infrastructure policy moving from federal and state debate into local land-use restrictions that can directly constrain compute siting. Click here
Global AI Strategy
Reflection AI partners with the Department of Energy on the Genesis Mission. The collaboration expands the use of open-source AI models in federally backed scientific research and national infrastructure initiatives. Click here
Mistral CEO Arthur Mensch warns Europe has a two-year window to avoid dependence on U.S.-controlled AI infrastructure. The statement reflects intensifying European pressure to build sovereign compute, energy, and model ecosystems as frontier AI increasingly concentrates around American hyperscalers and labs. Click here
Abu Dhabi launches a $13 billion initiative to build the world’s first fully AI-native government by 2027. The strategy positions sovereign AI not just as compute infrastructure but as a national operating model, integrating agentic systems, workforce transformation, and automated public services into state administration. Click here
The UAE launches a federal Agentic AI strategy to shift 50% of government services and operations to agentic systems within two years. The program positions AI-native public administration as a national operating model rather than a standalone digital-government initiative. Click here
HIVE’s BUZZ HPC announces a 320MW sovereign AI infrastructure project in the Greater Toronto Area.The project ties Canadian AI competitiveness to domestically controlled power, land, and GPU capacity, with a planned AI gigafactory designed to host more than 100,000 GPUs. Click here
Talent Signals
Each week, we spotlight key roles tied to the themes shaping this week’s AI headlines, connecting talent to the companies driving the news.
Sekaibuilds AI-native software for teams working across content, workflows, and digital experiences. As companies adopt AI tools that move beyond simple assistants into embedded productivity layers, Sekai is expanding across technical and go-to-market roles. Open roles are listed on the Mayfield job board. Click here
Modal provides cloud infrastructure for running AI workloads, including inference, training, batch jobs, and secure sandboxes for agents. As teams move from prototypes to production AI systems, Modal helps developers scale workloads without managing the underlying infrastructure. Open roles are listed on its careers page. Click here
Socket builds software supply chain security tools that help developers find, audit, and manage open-source dependencies safely. As AI coding tools increase the speed and volume of software development, Socket is helping teams secure the code and packages behind modern applications. Open roles are listed on its careers page. Click here
You can see all the opportunities at Mayfield-backed AI companies here.
Social Signals
The most important conversations in AI are unfolding across social media, where top voices are shaping the next wave of signals and strategy. Here are some of the top social signals and their takes from the past week.
Andrej Karpathy (Click here) — “I’ve joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D.” In a post viewed 26.5M+ times, Karpathy announced he is joining Anthropic’s pretraining team, which oversees the data training process behind Claude models. The move underscores how central frontier research talent remains in the race to improve model quality, reasoning, and training systems, as leading labs continue investing heavily in the small group of researchers shaping the next generation of AI capabilities.
Aaron Levie (Click here) — “Token costs will become a dominant topic in enterprises going forward with AI. Fortune 500 CIOs are debating workload prioritization, spend caps, access tiers, and how to justify AI usage by team and use case. Basically no one feels like they have the right solution.” Levie highlights how AI usage is creating a new operating expense category inside large companies, with enterprises now trying to manage model access, budgets, and compute allocation as AI adoption scales.
Zeb Evans (Click here) — “Today we reduced headcount by 22%. The business is the strongest it’s ever been. We’re restructuring around what I call the 100x organization. The great engineers are becoming 100x engineers. They’re not writing code. They’re directing agents that write code. The skill is judgment.” In a post viewed 6.4M+ times, the ClickUp CEO describes a broader restructuring around AI-native workflows, where orchestration, review, and system design replace many traditional coordination layers. Evans argues that as coding and execution accelerate, the highest leverage shifts toward small groups of highly adaptive operators who can manage agents, redesign workflows, and compound productivity across the organization.
