Today on The Send: PitchBook's mid-year venture report reveals a historic distortion, with AI swallowing 86% of the record $412.7 billion deployed in the U.S. market during the first half of the year. We also look at Europe's sudden scramble to rewrite its flagship MiCA crypto regulations just weeks after launch, and how AI agents are attempting to solve the travel industry's booking reliability problem.
Small businesses and some large enterprises are increasingly ditching expensive SaaS contracts from vendors like Salesforce and ServiceNow, opting instead to build custom applications using AI coding tools. A report on Wednesday highlights a trend of companies achieving 40-80% reductions in software expenses by developing bespoke software tailored to their specific needs, avoiding paying for unused features in bloated SaaS suites.
Why it matters
This signals a significant disruption in the enterprise software market, flipping the traditional 'build vs. buy' calculus. For a founder, this is a powerful proof point that AI tools are making it economically viable for small, lean teams to create their own operational software, freeing up significant capital that would otherwise go to recurring SaaS fees. This trend fundamentally changes the cost structure and operational agility for new ventures.
A new analysis on Thursday details how AI is being practically applied to tourism operations to address rising costs. With tourism demand recovering, many operators face increased operational complexity and manual processes. AI is now being deployed to automate workflows, validate reservations, streamline customer communications, and provide real-time operational monitoring, leading to significant efficiency gains and improved profitability.
Why it matters
This moves the conversation about AI in travel from customer-facing discovery to back-end operations, where it can drive real margin improvements. For a founder building an outdoor travel business, this provides a concrete playbook for leveraging AI to create a lean, scalable operation. It highlights specific areas where technology can reduce overhead and enhance efficiency, which are critical for a new venture's success.
The Bureau of Land Management (BLM) is revising its grazing regulations for the first time since 1995, with proposed changes that would expand livestock grazing on 155 million acres of public land while limiting public input on permit decisions. Reports from Wednesday indicate the new rules, heavily influenced by Interior department officials, are designed to reduce regulatory burdens on ranchers and could potentially sideline non-commercial grazing, such as tribal bison herds.
Why it matters
This represents a significant policy shift that prioritizes resource extraction over conservation and public participation on Western public lands. The move could lead to increased land degradation and wildlife impacts, altering the landscape for outdoor recreation. Limiting public comment also reduces the avenues for access and conservation groups to influence land management decisions.
A new report on 2025 participation data, released Thursday, shows that while outdoor recreation hit a record 183.2 million participants, there's a troubling trend underneath. The growth is driven by a surge of newcomers, but the frequency of participation is declining, and the number of 'core' enthusiasts is decreasing. This poses a long-term challenge for the $1.3 trillion outdoor economy.
Why it matters
This highlights a fundamental shift in the outdoor market that a new business must navigate. The challenge is no longer just attracting people to the outdoors, but converting casual participants into engaged, repeat customers. For an outdoor travel founder, this data suggests a need to focus on products and services that build community and foster long-term engagement to create a sustainable business.
PitchBook has put a definitive number on the venture concentration we've been tracking: U.S. funding hit a record $412.7 billion in the first half of the year, but an overwhelming 86% ($355.9 billion) of that went straight to AI. These figures, heavily skewed by mega-rounds like Anthropic's, coincide with a 27% collapse in seed funding as institutional capital retreats to established players.
Why it matters
This starkly illustrates the 'barbell' market dynamic. For founders scouting the landscape, it confirms the immense challenge of early-stage fundraising outside the AI halo, reinforcing the strategic value of alternative capital paths or strict operational bootstrapping.
Following the investor pivot toward 'real world' AI applications we noted earlier this week, Lerer Hippeau's Eric Hippeau characterizes the market as entering its 'second generation.' While infrastructure titans still absorb the mega-rounds, early-stage investors are now prioritizing application-layer startups that apply AI to existing business workflows, demanding vertical solutions and financial discipline over foundational models.
Why it matters
For founders, this confirms a hardening directive: the venture opportunity has definitively moved from building general models to demonstrating immediate ROI through vertical, industry-specific applications.
A competition run by TestSprite challenged four leading AI coding agents to build the same application, yielding surprising results. According to the report on Wednesday, the cheapest agent achieved the highest accuracy at half the cost of the priciest contender. The CoderCup competition also found that all agents, regardless of price, experienced significant code regressions, underscoring that development speed does not equate to quality.
Why it matters
This challenges the assumption that bigger, more expensive AI models are inherently better. For a founder building with AI, this is a crucial insight: strategic implementation of robust verification loops and human oversight is more impactful than simply paying for the top-tier model. It suggests that cost-effective, high-quality AI development is accessible to startups with smaller budgets, provided they focus on process, not just the model.
A podcast released Thursday argues that as of mid-2026, 'vibe coding'—creating software through natural language prompts—has matured into a genuinely viable path for non-technical founders. The analysis credits the evolution of AI tools, particularly models like Anthropic's Claude Fable 5, for making it possible to build working software, internal tools, and web apps without writing code.
Why it matters
This is a practical assessment of how AI is lowering the barrier to entrepreneurship. For a second-time founder, it confirms that the ability to build a product is becoming decoupled from technical coding skills. Understanding the current capabilities and limitations of these tools is critical for leveraging them to design products and manage operations with a small, agile team.
A new Omnisend survey published Wednesday reveals that nearly half (47%) of Americans are scaling back or canceling summer travel plans. The primary drivers are rising costs for essentials like gas and groceries, with many consumers redirecting their vacation budgets toward daily necessities and debt repayment. The trend favors shorter, cheaper trips, day outings, and staycations.
Why it matters
This is a clear signal of consumer belt-tightening impacting discretionary spending. For a founder in the outdoor travel space, it underscores the need for accessible, budget-friendly offerings. The market is shifting towards local and regional experiences that provide high value without requiring significant financial outlay, creating an opportunity for businesses that can cater to this cost-conscious mindset.
To tackle the 'last mile' AI travel booking problem we've been tracking, a new venture called Travel Guild has detailed a 'society of specialists' agent architecture. Single LLMs suffer a known 30% failure rate on transactional tasks; Travel Guild's system bypasses this by using eleven specialized agents and four distinct protocols to separate generative reasoning from deterministic rule enforcement, claiming 100% booking success in feasible scenarios.
Why it matters
This offers a technical blueprint for the robust architecture the travel industry has been searching for. For founders building in this space, it demonstrates that overcoming AI's inherent non-determinism requires complex, multi-agent orchestration rather than simply prompting a more powerful foundational model.
Less than two weeks after the EU's Markets in Crypto-Assets (MiCA) regulation took full effect on July 1, the European Commission is already scrambling to revise it. The framework's strict reserve requirements inadvertently forced the widespread delisting of Tether (USDT), effectively handing a dollar-denominated stablecoin monopoly to US-based Circle (USDC) and sparking immediate concerns over European monetary sovereignty.
Why it matters
This is a high-profile case study in regulatory blowback. For the fintech sector, the EU's rapid pivot acknowledges that even long-planned compliance regimes can trigger immediate, unintended market distortions, highlighting the fragility of regulating global digital assets at the regional level.
Venture Capital Consolidates Around AI Infrastructure New H1 2026 data confirms a massive structural shift in venture funding, with AI-related deals, particularly for infrastructure, capturing 86% of the record $412.7 billion invested in the US. This creates a 'K-shaped' market where mega-rounds for a few AI giants are masking a 27% collapse in seed funding for other sectors.
The 'Build vs. Buy' Equation Flips Toward AI-Built Tools A significant trend is emerging where small businesses and startups are ditching expensive SaaS contracts in favor of custom applications built with AI coding tools, reporting 40-80% cost savings. This democratization of software development is empowering small teams to create bespoke solutions, challenging the dominance of established enterprise software.
AI in Travel Moves from Discovery to Operations Following the rush to integrate AI for travel discovery, the focus is now shifting to using AI to solve core operational problems. New funding and products are aimed at automating workflows, managing complex bookings with specialized agent architectures, and driving revenue, signaling a maturation of AI's role in the travel industry.
AI Coding Enters an Era of Practical Benchmarking The conversation around AI coding tools is moving beyond hype to practical, real-world evaluation. A recent competition revealed the cheapest AI coding agent could outperform the most expensive in accuracy, while the extreme coding velocity touted by figures like Garry Tan is raising serious security concerns. This highlights a new focus on verification, security, and cost-effectiveness over raw model power.
Regulators Grapple with Unintended Consequences Both the EU's crypto framework (MiCA) and the US public lands grazing rules are facing immediate challenges. Less than two weeks after MiCA's enforcement, the EU is already consulting on a rewrite after it unintentionally created a stablecoin monopoly. Simultaneously, the BLM's proposed grazing rule overhaul, which limits public input, is drawing fire from conservationists, showing how difficult it is to get complex regulations right on the first try.
What to Expect
2026-09-01—SurfAid and Rip Curl's 'Make A Wave 2026' fundraising challenge begins.
2026-09-08—BLM plans to hold an oil and gas lease sale for 29 parcels in Colorado.
2026-10-06—The Tourism Innovation Summit 2026 kicks off in Seville, Spain.
2026-10-25—The UK's Financial Conduct Authority (FCA) new rules for cryptoassets are scheduled to take effect.
How We Built This Briefing
Every story, researched.
Every story verified across multiple sources before publication.
🔍
Scanned
Across multiple search engines and news databases
484
📖
Read in full
Every article opened, read, and evaluated
174
⭐
Published today
Ranked by importance and verified across sources
11
— The Send
🎙 Listen as a podcast
Subscribe in your favorite podcast app to get each new briefing delivered automatically as audio.
Apple Podcasts
Library tab → ••• menu → Follow a Show by URL → paste