Zeitgeist #5Saturday, September 5, 2026
Underlying Desire
At the core, this trend is about human relief. Teams do not actually want more AI. They want less context switching, fewer tedious handoffs, fewer forgotten tasks, and a sense that the work is finally moving without them having to push every piece manually. Agents promise a rare thing in enterprise software: the feeling that software can take responsibility, not just display information.
Key Evidence
OpenAI says frontier firms are generating 8.3x as many output tokens per active user as typical firms, up from 2.6x in January 2026, a sign of much deeper day-to-day usage. Microsoft’s 2026 Work Trend Index says nearly one in five firms in software and technology are already using agents. McKinsey’s 2026 global AI survey found 44% of respondents say AI is already scaling across the enterprise, up from 38% the prior year.
Why Now
Three things changed at once: models got more reliable, enterprise buyers got more willing to wire AI into real work, and vendors started shipping agent platforms with enough permissions and integrations to move beyond pilots. The result is a market that is no longer asking whether agents can do work, but which workflows they should own first.
Domain-specific AI for legal and professional services that helps firms run contract analysis, due diligence, compliance, and litigation workflows.
Outcome: Raised $200 million in March 2026 at an $11 billion valuation, with more than $100 million ARR, 2,400+ customers, and 75+ AmLaw 100 firms as customers. ([harvey.ai](https://www.harvey.ai/blog/harvey-raises-at-dollar11-billion-valuation-to-scale-agents-across-law-firms-and-enterprises?utm_source=openai))
An enterprise AI agent platform for customer service and customer experience that handles real business interactions, not just conversations.
Outcome: Reached $100 million ARR in seven quarters, and in May 2026 TechCrunch reported more than 40% of the Fortune 50 as customers. The company also announced a $950 million funding round in 2026, pushing its post-money valuation above $15 billion. ([sierra.ai](https://sierra.ai/blog/100m-arr?utm_source=openai))
A SaaS layer that sits on top of existing business systems like Slack, Gmail, Jira, Salesforce, and Notion, then turns repeated multi-step work into agent-run workflows with approvals, memory, and audit trails. The target customer is ops-heavy teams in mid-market and enterprise companies: sales ops, rev ops, customer success, finance ops, and IT. It would work because the buyer does not want another chatbot, they want one place where routine work can be triggered, routed, reviewed, and measured without staff chasing tasks across five tabs.
A platform for monitoring, testing, and governing enterprise AI agents across departments. It would show which agents are running, what tools they touched, where they failed, how much they cost, and whether a human needs to step in. The target customer is any company deploying multiple agents across functions, especially regulated industries and larger software companies. It would work because once firms move past one or two pilots, they need observability and policy controls the same way they needed dashboards and logs when they adopted cloud infrastructure.
Underlying Desire
At the core, this trend is about relief. People do not just want to lose weight, they want freedom from the constant background negotiation with hunger, guilt, and willpower. GLP-1s promise something psychologically huge: fewer intrusive cravings, less self-blame, and a path toward a body that feels more manageable, which is why the products and services around them will increasingly be judged on whether they reduce friction, anxiety, and decision fatigue.
Key Evidence
PwC says 21% of U.S. households include a current GLP-1 user as of May 2026, up from 9% in January 2025, a rapid jump into mainstream consumer life (PwC, 2026). CVS Caremark expects weight-management GLP-1 use to grow about 25% in 2026, signaling the category is still expanding (CVS Caremark, 2026). CVS also projects roughly a 15% decrease in per-prescription net cost in 2026 versus 2025 for commercial formulary clients with standard utilization management, which could broaden access (CVS Caremark, 2026).
Why Now
Two things changed recently. First, usage moved from a premium, early-adopter phenomenon into the household mainstream, which means the downstream behavior changes are now visible in consumer data, not just clinic anecdotes (PwC, 2026). Second, cost and formulary dynamics are improving, with CVS Caremark forecasting lower net prescription costs in 2026, making broader employer and payer adoption more realistic (CVS Caremark, 2026).
Outcome: Raised $100M Series C in May 2026, bringing total funding to $215M. Nourish says it has 10,000+ registered dietitians, millions of completed appointments, more than 200 million covered lives, and hundreds of thousands of monthly active users. ([nourish.com](https://www.nourish.com/blog/nourish-announces-series-c?utm_source=openai))
Outcome: Berry Street says it has 15,630 LinkedIn followers, works with some of the largest health plans, and in early 2025 raised a $50M Series B. Its company page lists 294 employees, and its team said the business grew 8x in 2024 while serving over 1,000 dietitians. ([linkedin.com](https://www.linkedin.com/company/berry-street-health))
A consumer health app that helps GLP-1 users adapt their eating patterns, hydration, protein intake, side effect tracking, and appetite changes in one simple workflow. The target customer is the millions of people starting GLP-1s through employers, telehealth, or direct care who need practical guidance after the prescription, not just a clinical handoff. It works because the hardest part of GLP-1s is not getting the drug, it is successfully living on it, and current support is fragmented across doctors, dietitians, and random social media advice.
A SaaS platform for employers, benefits consultants, and health plans to manage GLP-1 programs with better adherence, education, utilization support, and outcome tracking. The buyer already pays for the medication and now needs a way to reduce waste, improve persistence, and prove ROI across a fast-growing population. It works because the cost curve and access curve are both moving, which creates demand for software that turns a blunt pharmaceutical benefit into a managed program with measurable outcomes.
Underlying Desire
At the deepest level, this trend is about reducing uncertainty. Companies want permission to keep innovating without being punished for not knowing every rule, every deadline, and every evidence requirement. AI governance and compliance software sells peace of mind: the feeling that your organization is not one audit, one regulator, or one whistleblower away from chaos.
Key Evidence
The European Parliamentary Research Service says the remaining provisions of the EU AI Act began applying on 2 August 2026, including fines for GPAI providers, which moves AI governance from optional to mandatory. The European Commission says the AI Omnibus entered into force on 27 July 2026, with a mandate to simplify compliance while keeping safeguards in place. CARB says California’s SB 253 requires large U.S.-based entities doing business in California to report Scope 1 and Scope 2 emissions beginning in 2026, with Scope 3 reporting starting in 2027.
Why Now
Three things changed at once: legal deadlines arrived, simplification did not eliminate obligations, and reporting requirements expanded beyond AI into climate disclosure. That combination creates urgent demand for software that can centralize evidence, automate workflows, and generate regulator-ready outputs. This is actionable now because firms need implementation time before the first enforcement cycles and reporting periods hit. The market is shifting from policy interpretation to operational execution, which is exactly where software beats consultants.
AI governance platform that helps regulated companies document, assess, and prove compliance for their AI systems.
Outcome: Modulos says it has raised CHF 16.4M to date, and its site highlights customer deployments and case studies across regulated sectors. ([modulos.ai](https://www.modulos.ai/company/?utm_source=openai))
Real-time compliance firewall for AI-generated communications that rewrites or blocks noncompliant outputs before they are delivered.
Outcome: ZeroDrift closed an oversubscribed $10 million seed round in June 2026, backed by a16z Speedrun, Reign Ventures, Pitchdrive, and U&I Ventures, and LinkedIn shows early traction with tier-one banks, asset managers, insurance companies, and enterprises. ([globenewswire.com](https://www.globenewswire.com/news-release/2026/06/02/3305264/0/en/ZeroDrift-Raises-10M-Seed-Round-to-Build-the-Compliance-Firewall-for-AI.html?utm_source=openai))
Policy Pilot is an AI governance SaaS for mid-market and enterprise companies that need to document model inventory, risk controls, human oversight, and approval workflows across internal AI use cases and vendor models. It would work because most teams are trying to satisfy legal, security, and engineering requirements with spreadsheets and tickets, which breaks the moment an audit request lands. The product should auto-generate evidence packs, map controls to specific regulations like the EU AI Act, and keep a living change log for every model and policy update.
Disclosure OS is a reporting automation platform for companies that have to produce climate, AI, and risk disclosures without hiring a small army of consultants. It would pull data from ERP, HR, cloud, procurement, and model management systems, then assemble regulator-ready reports, evidence folders, and sign-off workflows. This works because the pain is not just collecting data, it is proving consistency across teams and keeping every number traceable back to a source system.
Underlying Desire
The deeper desire is control over money movement. Businesses want payments that are faster, cheaper, and more predictable, but just as important, they want visibility, accountability, and compliance without hiring an army of specialists. Stablecoin rails promise the fantasy every finance team has quietly wanted for decades: internet-speed settlement with bank-grade trust.
Key Evidence
The GENIUS Act was enacted in 2025 to regulate payment stablecoins, according to GovInfo, turning a gray area into a defined regulatory category. The OCC’s March 2026 proposed rule says the effective date is the earlier of 18 months after enactment or 120 days after final rules, per the OCC, which creates a near-term implementation deadline. The Federal Reserve’s April 2026 note says stablecoins are increasingly tied into traditional finance through intermediation chains, wallet partnerships, and vertical integration, according to the Fed.
Why Now
This became actionable because regulation is no longer theoretical. The U.S. government has moved from debating whether stablecoins should exist to defining how they operate, and agencies are now issuing the implementation rules. At the same time, stablecoins are being integrated into traditional finance workflows, which turns them into a real business systems problem, not just a crypto market structure problem.
Bridge builds stablecoin infrastructure and APIs that let businesses accept, store, convert, issue, and spend digital dollars.
Outcome: Acquired by Stripe, with acquisition completed on February 4, 2025. Bridge also received OCC conditional approval in 2026 for a federally chartered national trust bank, and its site says it supports hundreds of teams building new money movement experiences. ([stripe.com](https://stripe.com/newsroom/news/stripe-completes-bridge-acquisition?utm_source=openai))
BVNK provides enterprise stablecoin payments infrastructure for businesses and financial institutions.
Outcome: $50 million Series B raised in December 2024. By December 2025, BVNK said it was processing $30 billion in annualized stablecoin payment volume, with 2.8 million transactions and 226 new customers added in 2025. Its about page lists 2021 as the founding year and $39 billion plus annualized volume. ([bvnk.com](https://www.bvnk.com/blog/series-b-fuel-next-era-of-stablecoin-payments?utm_source=openai))
RailOps is a compliance-first payments orchestration layer for fintechs and mid-market businesses that want to send and receive stablecoin payments without rebuilding their finance stack. It would route transactions across wallets, chains, banks, and payout partners, while handling KYC, sanctions screening, reconciliation, exception management, and audit logs in one dashboard. This works because the market is about to need boring software more than clever crypto, and the first durable winners will sell trust, visibility, and operational control to finance teams.
Treasury Bridge is a treasury management tool for companies that hold or move stablecoins across entities, geographies, and counterparties. It would give CFOs a single view of balances, settlement timing, liquidity buffers, FX exposure, and policy controls, with automated alerts for compliance and accounting exceptions. The product works because stablecoins are becoming embedded in traditional finance workflows, but most finance teams still lack software that treats them like operating cash rather than speculative assets.
Underlying Desire
At the deepest level, this trend is about control over scarcity. Data center operators, cloud buyers, and AI teams all want the same thing: reliable access to abundant power without friction, delay, or surprise. The human desire underneath is security and momentum, the ability to keep growing without being constrained by a system that feels slow, opaque, and out of sync with modern demand.
Key Evidence
The IEA says data-center electricity demand rose 17% in 2025, while global electricity demand grew 3%, according to the IEA. It also says five major tech companies spent more than $400 billion on capex in 2025, with that total expected to rise another 75% in 2026, according to the IEA. The U.S. EIA says U.S. electricity demand grew about 1.7% annually from 2020 to 2025, compared with 0.1% annually from 2005 to 2019, according to the EIA.
Why Now
The shift is happening because AI infrastructure demand has outgrown legacy planning assumptions. Utilities and regulators are now dealing with faster load growth, longer interconnection queues, and more complex site selection decisions than their software stack was built for. At the same time, the DOE is openly calling for demand flexibility and grid modernization as a response to data-center growth, which creates a policy and procurement opening for software vendors. The buyer no longer wants static reports, it wants systems that can act in real time.
AI power orchestration software that gives utilities and data centers real-time visibility and control over electricity use. ([utilidata.com](https://utilidata.com/our-company?utm_source=openai))
Outcome: Raised $100 million in Series C financing as of May 2026, including a $40 million extension, and says the NexGen Cloud deployment will target up to 50% additional usable capacity. ([utilidata.com](https://utilidata.com/press-release/utilidata-raises-40-million-extension?utm_source=openai))
Software that turns AI data centers into flexible grid assets by dynamically adjusting power use in response to grid conditions. ([emeraldai.co](https://www.emeraldai.co/blog/sharing-our-strategic-expansion-round-emerald-ai-raises-25-million-to-transform-ai-data-centers-into-flexible-power-grid-assets?utm_source=openai))
Outcome: Raised $150 million in an oversubscribed Series A on August 25, 2026, at a $1.05 billion valuation, and said it had completed five global demonstrations and deployed commercially at multi-megawatt scale. ([emeraldai.co](https://www.emeraldai.co/blog/emerald-ai-raises-150-million-series-a?utm_source=openai))
GridPilot is a software platform for data center operators that forecasts power availability, automates utility coordination, and recommends the cheapest and fastest path to new load. It would ingest site data, tariff structures, interconnection status, backup generation constraints, and local grid conditions, then turn that into an operating dashboard for facilities and finance teams. It works because the biggest pain is not just power procurement, it is managing the messy coordination between growth plans and physical grid reality.
LoadShifter is a control layer for large compute customers that shifts non-urgent workloads across time, sites, and power contracts to reduce cost and avoid grid bottlenecks. The product would target AI labs, colocation customers, and cloud-heavy enterprises that want to run jobs when electricity is cheapest or when the local grid has capacity. It wins by turning power constraints into an optimization problem instead of a crisis, and by making compute scheduling financially legible to ops and finance teams.
Underlying Desire
At the core, this trend is about wanting to feel seen, chosen, and safely included by other people. Humans do not just want entertainment or information, they want evidence that they matter to a group and that someone would notice if they disappeared. Software is now being asked to serve that ancient need by lowering the friction of initiation, reducing social awkwardness, and creating repeatable rituals of belonging.
Key Evidence
The WHO Commission on Social Connection says social isolation and loneliness are widespread and linked to serious impacts on health, well-being, and society, according to WHO. The WHO report, as cited by PAHO, estimates loneliness costs employers US$154 billion annually and the broader economy US$406 billion annually. The UK government’s Community and Engagement Survey 2025/26 found loneliness levels remained broadly stable at around 5% to 6% for years, according to GOV.UK.
Why Now
Three things changed recently: loneliness is now being measured by public institutions, its costs are being translated into dollar terms, and the problem has remained stubbornly persistent instead of fading. That combination makes it easier for buyers to justify software spend and easier for founders to sell outcomes instead of vague social good. The market is actionable now because the pain is documented, recurring, and expensive.
An AI companion app for friendship, emotional support, and romantic-style conversation.
Outcome: Launched in July 2017, Replika says more than 10 million people have joined, and CEO Eugenia Kuyda said in 2024 that total users had surpassed 30 million. The company also raised at least $11 million in early funding, including a $6.5 million Series A.2 round in 2017. ([replika.com](https://replika.com/press?utm_source=openai))
A wearable AI companion that listens to your day and responds through text messages on your phone.
Outcome: Friend launched in 2024, raised $2.5 million in seed funding at a reported $50 million valuation, later expanded to about $7 million in total venture funding, and began taking preorders for a $99 necklace. TechCrunch reported shipment delays into 2025, which suggests the company is still early but already commercially visible. ([techcrunch.com](https://techcrunch.com/2024/07/30/friend-is-an-ai-companion-backed-by-founders-of-solana-perplexity-and-zfellows/?utm_source=openai))
Belong Loop is a SaaS platform for community operators, coworking spaces, alumni networks, and membership brands that turns loose audiences into repeat human relationships. It uses lightweight matching, event prompts, attendance follow-up, and engagement scoring to help organizers spot who is drifting, who should meet, and which rituals keep people coming back. It would work because most communities do not have a connection problem, they have an operations problem, and software can make those operations systematic.
Friendship CRM is a tool for HR teams, managers, and employee experience leads that tracks relationship health inside teams and nudges people toward meaningful connection before disengagement becomes attrition. It could include intro prompts, small-group matching, meeting templates, and anonymous risk signals so companies can see which teams are socially isolated. It works because employers already spend on engagement and retention, and loneliness now has a quantifiable economic cost that makes social connection a legitimate workplace KPI.
Underlying Desire
At the core, this trend is about control. Creators want freedom from algorithm roulette, brands want proof that spend converts, and audiences want a direct relationship with the people they trust. Owning the audience satisfies a very old human instinct: if something matters, keep a way to reach it without asking permission from a gatekeeper.
Key Evidence
According to IAB's 2025 Creator Economy Ad Spend Strategy Report, U.S. creator ad spend is projected to reach $37 billion in 2025, up 26% year over year, and $44 billion in 2026. IAB also says creator ad growth is nearly 4x faster than the media industry overall, a strong signal that the category is moving into mature, measurable spend. The same report points to rising demand for owned-audience infrastructure, including CRM, memberships, community, and paywalls.
Why Now
The shift is actionable now because creator budgets have become large enough for brands to demand measurement, not just reach. At the same time, creators are feeling the risk of platform dependence more acutely, which makes owned channels more valuable than ever. What changed is not that creators suddenly discovered newsletters or memberships, but that the economics finally justify the operational overhead. Once sponsored content becomes a serious business, the software stack around it has to mature too.
Outcome: $33 million Series B in April 2024, total funding around $46.5 million, and annualized revenue above $30 million in 2026. beehiiv also says its ad network has paid publishers over $1 million since launching in 2023. ([beehiiv.com](https://www.beehiiv.com/blog/the-beehiiv-story-chapter-3?utm_source=openai))
Subscription publishing platform that lets writers and creators own their audience and monetize directly.
Outcome: Substack says it has 5 million paid subscriptions and more than tens of millions of weekly readers. It raised $100 million in Series C funding in July 2025, which pushed its valuation to about $1.1 billion. ([substack.com](https://substack.com/about?utm_source=openai))
A SaaS platform for mid-tier creators and small creator teams that centralizes email, memberships, CRM, sponsorships, and retention analytics in one dashboard. It would help creators identify which followers are most likely to convert, which offers drive recurring revenue, and which sponsorships are actually worth repeating. This works because creators are moving from platform-native distribution to owned channels, but most are still juggling disconnected tools and spreadsheets.
A workflow tool for creators, talent managers, and boutique media agencies that automates sponsor intake, deal tracking, deliverable management, and post-campaign reporting. It would reduce the messy back-and-forth that kills time and margin when sponsorship volume grows. The product wins because creator ad spend is scaling, and brands now expect creator campaigns to behave more like accountable media buys than ad hoc partnerships.
Lev turns an idea into the things a company actually runs on: positioning, market and competitive research, a pitch deck, and a go-to-market plan.
Start building for free