# What 2026 AI Valuations Mean for Employees and Investors

In 2026, artificial intelligence mega-rounds have pushed valuations to historic highs, with frontier labs approaching trillion-dollar market caps while simultaneously absorbing over 80% of all global venture capital funding. For tech employees and public investors, this extreme concentration means that participating in the AI boom is riskier than ever, laden with punitive liquidation preferences and aggressive IPO lockup traps. Navigating this era requires looking past headline valuations to evaluate actual enterprise return on investment, defensible unit economics, and structural deal terms.

## The Dot-Com Comparison: History Rhymes, But With Cash Flow

As global artificial intelligence investments surpass $2.5 trillion [cite: 1, 2], comparisons to the late-1990s dot-com bubble have dominated financial commentary. In early 2026, the S&P 500 trades at approximately 30 times forward earnings for the technology sector, while the broader Shiller CAPE (cyclically adjusted price-to-earnings) ratio stands between 38 and 40 [cite: 3, 4]. In 155 years of recorded market history, this metric has been higher exactly once: March 2000, when it peaked at 44.19 just one month before the Nasdaq began a decline that would erase 78% of its value [cite: 3, 5].

The fundamental anatomy of the 2026 AI boom, however, diverges sharply from the internet mania of 2000. The primary differentiator is robust, undeniable cash flow.

During the dot-com era, valuations were largely built on consumer attention metrics for companies that were actively destroying capital. Retail investors funneled unprecedented capital into companies with undefined price-to-sales ratios, operating under the assumption that a paradigm shift was occurring [cite: 5, 6]. By contrast, the leaders of today's AI rally are generating historic profits. NVIDIA reached a market capitalization of approximately $4.3 trillion by early 2026, making it the most valuable company in the world [cite: 6]. This valuation is supported by a staggering $215.9 billion in FY2026 revenue with massive gross margins of 71% and net income exceeding $120 billion [cite: 3, 6]. 

The current vulnerability lies not in a lack of revenue, but in extreme market concentration. The top ten companies in the S&P 500 now account for 36% to 40% of the index's total market capitalization, a figure that is nearly 50% higher than the peak concentration seen in 2000 [cite: 3]. According to Goldman Sachs, a staggering 85% of the S&P 500's gains in the first half of 2026 came from the technology sector alone; excluding tech, the index advance drops to a mere 3% [cite: 7].

### The Infrastructure Spending Gamble

To support this continued growth, hyperscaler capital expenditure—the combined infrastructure spending of Microsoft, Google, Amazon, and Meta—is projected to reach between $660 billion and $700 billion in 2026 [cite: 3, 6]. This figure represents the largest corporate investment program in history outside of wartime mobilization [cite: 3]. Furthermore, Gartner forecasts that global end-user spending on AI-optimized infrastructure as a service (IaaS) will reach $37.5 billion in 2026, with AI-optimized servers accounting for 17% of total AI spending [cite: 8, 9]. 

The central question for investors is whether this annual infrastructure spend will generate downstream returns that justify the investment. Until the software application layer proves it can monetize this compute capacity at scale, the hardware market remains highly fragile. This fragility was starkly illustrated by a single event in early 2025: the release of a competitive, low-cost AI model by Chinese startup DeepSeek, reportedly trained for just $5.6 million, which temporarily wiped $588.8 billion from NVIDIA's market cap in a single trading day [cite: 6]. 

| Valuation Metric | Dot-Com Bubble Peak (Mar 2000) | Current AI Boom (Early 2026) | Key Difference & Context |
| :--- | :--- | :--- | :--- |
| **Top Company Market Cap** | Cisco: ~$370 Billion [cite: 6] | NVIDIA: ~$4.3 Trillion [cite: 6] | Today's leaders are more than 10x larger in nominal terms. |
| **Tech Sector Forward P/E** | ~50x to 60.1x [cite: 3, 6] | ~30x [cite: 3] | AI valuations, while high, are fundamentally supported by stronger earnings. |
| **Shiller CAPE Ratio** | 44.19 [cite: 3] | 38 - 40 [cite: 3] | Current market is the second-most stretched in 155 years. |
| **Top-10 S&P Concentration** | ~27% [cite: 3] | 36% - 40% [cite: 3] | Extreme concentration risk in a handful of mega-cap tech giants today. |
| **Profitability Profile** | Highly speculative; pure losses [cite: 3, 6] | $350B+ combined free cash flow [cite: 3] | Dot-coms traded on undefined P/S ratios; AI leaders are cash-generating machines. |

## The 2026 Funding Landscape: A Tale of Haves and Have-Nots

If the public markets are heavily concentrated, the private venture capital markets are almost completely monopolized by a handful of frontier laboratories. The first quarter of 2026 marked a historic anomaly for venture capital. Investors poured $297 billion into 6,000 startups worldwide—an increase of roughly 150% both quarter-over-quarter and year-over-year [cite: 10]. 

The driver of this explosion was unequivocally artificial intelligence. A total of $239 billion, or 81% of all global venture capital investment, flowed exclusively into AI startups [cite: 10]. To put this in perspective, in the previous record quarter of Q1 2025, that share stood at just 55% [cite: 10].

### The Frontier Lab Mega-Rounds

This capital is not being distributed evenly across the startup ecosystem. Four of the five largest venture capital rounds in history closed in the first quarter of 2026, accounting for 64% of all global venture capital deployed in that period [cite: 10]. 

The most prominent of these was OpenAI. In March 2026, OpenAI closed a record-shattering $122 billion round at an $852 billion post-money valuation [cite: 11, 12]. The round, initially anchored by a $110 billion tranche from Amazon, Nvidia, and SoftBank, was followed by an additional $12 billion that, for the first time, included participation from individual retail investors via bank channels [cite: 11]. OpenAI's revenue now reportedly sits at $2 billion per month ($24 billion annualized), driven by 900 million weekly active users and over a million enterprise customers [cite: 6, 12]. 

Anthropic, the developer of the Claude model family, followed closely behind. The company secured $65 billion in Series H funding, pushing its post-money valuation to an astonishing $965 billion [cite: 13]. Anthropic's run-rate revenue has reportedly reached $14 billion, growing 10x annually for three consecutive years [cite: 11]. Its focus on enterprise safety has paid dividends, with massive corporate rollouts including a partnership to embed Claude across KPMG's global workforce of 276,000 professionals [cite: 14]. 

Elon Musk’s xAI also capitalized on the funding frenzy, closing a $20 billion Series E round [cite: 11, 12]. This round coincided with xAI merging its strategic interests with SpaceX, creating a combined entity that positions xAI's Grok model as the primary artificial intelligence vehicle ahead of SpaceX's anticipated public offering [cite: 11].

```json
{
  "concept": "Visualizing the extreme capital concentration in the first quarter of 2026, showing how a few frontier AI labs absorbed the vast majority of global venture funding.",
  "reasoning_for_value": "Venture capital figures of this magnitude are difficult to comprehend in text alone. A bar chart comparing these mega-rounds emphasizes the 'winner-take-all' dynamic of the foundational model layer, demonstrating that AI funding is heavily skewed toward a few titans rather than broadly distributed.",
  "title": "Q1 2026 Capital Concentration: The Rise of the Mega-Round",
  "visual_type": "Horizontal Bar Chart",
  "generation_method": "CODE",
  "justification_of_choice": "A horizontal bar chart is the optimal format for comparing absolute monetary values across discrete categories (companies). Alternative 1: A pie chart would be inappropriate as these rounds do not represent 100% of the market, just the top tier. Alternative 2: A scatter plot comparing valuation to funding amount would be useful, but a simple bar chart is more direct for highlighting the sheer size of the capital raises.",
  "caption": "Four of the five largest venture capital rounds in history closed in Q1 2026, accounting for the vast majority of the $297 billion deployed globally.",
  "data_specification": {
    "source_snippets_ids": [43, 84, 85],
    "data_structure": "An array of objects containing 'Company' and 'Funding Raised (Billions USD)'.",
    "mapping": "Y-axis maps to 'Company' (OpenAI, Anthropic, xAI, Waymo). X-axis maps to 'Funding Raised (Billions USD)' (122, 65, 20, 16)."
  },
  "design_and_interaction": {
    "layout": "Standard horizontal bar chart with data labels on the end of each bar for immediate readability.",
    "aesthetics": {
      "style": "Modern & Minimalist. Clean grid lines, no excessive borders.",
      "color_palette": "Background: White (#FFFFFF). Bars: Focal Palette Google Blue (#1A73E8). Text: Near-Black (#111111).",
      "additional_details": "Ensure the X-axis is formatted with a '$' prefix and 'B' suffix."
    },
    "interactivity": "Static visual with no interactivity.",
    "animation": "No animation."
  }
}
```

### Sovereign AI and the European Push

While the United States dominates the frontier foundational model race, European startups have pivoted to a highly lucrative, specialized strategy: sovereign AI. European governments and enterprises, increasingly wary of relying entirely on American hyperscalers for critical intelligence, data infrastructure, and regulatory compliance (such as the EU AI Act), are pouring capital into domestic alternatives.

Mistral AI has emerged as Europe's definitive champion in this space. Following a €1.7 billion Series C in late 2025 led by semiconductor giant ASML, Mistral rapidly scaled its operations [cite: 15, 16]. By mid-2026, the company was reportedly finalizing a €2 billion round led by Abu Dhabi's MGX Fund, pushing its valuation to $14 billion—making it the fourth-largest AI unicorn in the world [cite: 17, 18]. Mistral's valuation is built on tangible enterprise adoption; the company is on track to generate €1 billion in annual recurring revenue (ARR) in 2026, with more than half of that revenue originating from European clients [cite: 16, 19]. Furthermore, Mistral secured €722 million in debt financing to build NVIDIA-powered data centers across Europe, cementing its commitment to localized, sovereign compute infrastructure [cite: 19]. 

Similarly, DeepL, the German AI translation startup, has maintained significant momentum. Following a $300 million Series C round led by Index Ventures that valued the company at $2 billion, DeepL is reportedly exploring a U.S. initial public offering for late 2026 at a potential valuation of up to $5 billion [cite: 20, 21, 22, 23]. DeepL's annual revenue surged to $185.2 million by the end of 2024, driven by an expanding base of over 100,000 business and government customers [cite: 21, 22].

### The B2B Pivot: Aleph Alpha and Cohere

Not every European AI company is racing to build the biggest foundational model from scratch. Aleph Alpha, once Germany's premier large language model hope, strategically pivoted away from the pure model-training race. Realizing that competing with OpenAI's capital advantage was untenable, CEO Jonas Andrulis shifted the company's focus toward a "generative AI operating system" for B2B clients [cite: 24]. This system helps enterprises fine-tune models with highly sensitive, domain-specific data without requiring the most powerful, generalized models [cite: 24]. 

This pragmatic shift culminated in April 2026, when Canadian AI firm Cohere acquired Aleph Alpha in a merger that created a $20 billion transatlantic sovereign AI entity [cite: 14, 25]. The deal gives Cohere access to Aleph Alpha's deep German public sector relationships and complements its existing LLM capabilities with European data sovereignty infrastructure [cite: 14]. 

## The Rise of Defense as a Software Business

Perhaps the most notable shift in European and global venture capital is the explosion of defense technology. Defense, once avoided by venture capitalists due to ethical mandates and notoriously slow government procurement cycles, is now producing some of the world's most valuable private companies. AI infrastructure is no longer just about chat applications; it is viewed by nation-states as critical kinetic infrastructure.

Helsing, a Munich-based defense AI startup, perfectly illustrates this surge. Founded in 2021, Helsing develops AI software for battlefield sensor fusion, autonomous strike drones, and fighter jet systems [cite: 26]. By mid-2026, the company was in advanced talks to secure $1.2 billion in funding at a staggering $18 billion valuation, led by U.S. growth firm Dragoneer Investment Group and Lightspeed Venture Partners [cite: 27]. If closed, this would make Helsing the most valuable private tech firm in Germany [cite: 27, 28]. 

Competitors are following suit across the continent. Stark, a Berlin-based drone and autonomous defense startup founded by former Quantum Systems executives, is in discussions to raise €300 million at a €2.5 billion valuation [cite: 29]. Quantum Systems itself saw its sales triple in 2025 to €300 million on the back of battlefield intelligence gathering in Ukraine, achieving a valuation above €3 billion [cite: 26, 29]. 

Globally, the defense tech sector is breaking records. In May 2026, U.S. defense giant Anduril raised $5 billion in a Series H round at a $61 billion valuation, doubling its valuation in less than twelve months [cite: 29, 30]. Venture capital deals in defense reached $49.1 billion in 2025, nearly doubling year-over-year [cite: 12]. The strategic implications are clear: the dual-use thesis is structural, with commercial AI capabilities having immediate defense applications, and vice versa [cite: 26].

## The Asian Arms Race: China's LLM Surge

The AI ecosystem in Asia is similarly accelerating, characterized by massive investments aimed at achieving parity with Western models. In China, the landscape is intensely competitive, with a few key players absorbing the majority of venture funding. Chinese AI startups raised $16.2 billion in Q1 2026 alone, marking an 185% year-over-year increase [cite: 31].

Moonshot AI, the Beijing-based developer of the Kimi large language model, has become the poster child for Chinese generative AI. In May 2026, Moonshot closed a $2 billion funding round led by Meituan's venture arm, Long-Z Investments, pushing its valuation past the $20 billion mark [cite: 32, 33, 34]. This round cemented Moonshot as China's highest-funded LLM startup, outpacing rivals like MiniMax and Zhipu AI [cite: 33, 35]. 

Moonshot's valuation is supported by explosive growth metrics. Following the update of its Kimi K2.5 model, the company's ARR doubled from roughly $100 million in March 2026 to over $200 million in April, driven by massive API usage and a surging paid subscriber base [cite: 32, 34]. The Kimi model has gained global traction, reportedly becoming the second most-used LLM on the OpenRouter tracking platform, trailing only top-tier Western models [cite: 33]. 

## Evaluating AI Startup Equity in 2026

The influx of capital into the AI sector has fundamentally rewritten the rules of talent acquisition. For senior technology leaders and specialized machine learning engineers, the 2026 job market offers unprecedented cash compensation, but the traditional allure of startup equity has become increasingly perilous.

High-growth AI startups are now offering base salaries ranging from $170,000 to $400,000, representing a massive 25% increase in base pay since 2022 [cite: 36]. Historically, early-stage startups offered minimal salaries up front, offsetting the risk with equity stakes that served as lottery tickets. Today, to compete with Big Tech firms, startups are forced to match FAANG-level cash compensation just to get candidates in the door [cite: 36].

Data from Carta reveals that for smaller AI startups valued between $1 million and $10 million, the median equity grant for AI/ML engineers increased by 59% from January 2024 to February 2026 [cite: 37]. However, the median salaries at these AI-native companies are actually slightly *lower* than at non-AI-native startups, indicating that true specialists are commanding massive equity premiums in exchange for slightly softer cash bases [cite: 37].

### The FAANG Opportunity Cost

For employees, evaluating an AI equity package requires treating the offer as a highly illiquid venture capital investment. The opportunity cost is stark. 

Consider a mid-level software engineer (L5) weighing two offers. At a major tech company like Google, they might take home approximately $1.46 million over four years in guaranteed, liquid compensation ($190k base salary plus $700k in Restricted Stock Units) [cite: 38]. 

In contrast, joining a seed-stage startup might yield $480,000 in cash over four years, plus 0.5% to 1.0% in equity [cite: 38]. For that startup equity to simply break even with the opportunity cost of the Big Tech job (roughly a $980,000 difference), the startup must eventually exit at a massive valuation. 

After factoring in standard dilution from future funding rounds (typically 15% to 25% per round) and a 35% capital gains tax rate, an employee would need their 0.5% equity stake to be worth $1.5 million pre-tax to net the required $980k [cite: 38]. This requires the seed-stage company to exit at over $300 million [cite: 38]. Statistically, fewer than 1% of seed startups ever reach a $300 million exit valuation [cite: 38].

### The Illusion of Paper Wealth

Furthermore, paper valuations are highly deceptive in the current market. If an employee joins a startup that recently raised capital at a 50x revenue multiple, their equity is priced at peak sector hype [cite: 39]. If the broader market recalibrates to a standard 10x software multiple before the company exits or IPOs, the employee's equity may be rendered virtually worthless, even if the company's revenue continues to grow steadily [cite: 39]. 

Experts advise modeling three liquidity scenarios before accepting an offer: projecting the 409A valuation against a sector growth rate, a flat rate, and a 20% market correction [cite: 39]. If the company's exit horizon is five years out, employees are betting that the current AI valuation premium will persist through an entire cycle of hype, disillusionment, and consolidation [cite: 39]. 

## Liquidation Preferences: Navigating the Payout Waterfall

The greatest hidden risk to employee and founder equity in the 2026 AI boom lies buried in venture capital term sheets: specifically, liquidation preferences. 

A liquidation preference is a contractual provision that dictates the exact order and amount investors get paid when a company exits via acquisition, IPO, or liquidation [cite: 40, 41, 42]. Because startup investments are inherently risky, venture capitalists demand downside protection. These terms ensure that investors recover their capital (or a multiple of it) *before* any common shareholders—which includes founders, early employees, and option pool holders—receive a single cent [cite: 40, 41].

While Q1 2026 data from Cooley shows that 98.2% of venture deals utilized the standard "1x" liquidation preference, the specific structure of that preference dictates the payout [cite: 43]. 

### Non-Participating vs. Participating Stock

Understanding the difference between these two structures is critical for anyone holding startup equity:

1.  **Non-Participating Preference:** This is the most common and founder-friendly structure, found in 96.4% of Q1 2026 deals [cite: 43]. The investor chooses the *higher* of two options: they either take their guaranteed 1x original investment back off the top, OR they convert their preferred shares into common shares and take their pro-rata percentage of the total exit [cite: 42, 44]. 
    *   *Example:* An investor puts $5 million into a startup for 20% ownership, with a 1x non-participating preference. The startup is sold for $10 million. The investor exercises their preference to take $5 million (since 20% of $10M is only $2M). This leaves $5 million for the founders and employees. If the startup sold for $50 million, the investor would convert to common stock and take 20% ($10 million) [cite: 44]. The point at which it makes mathematical sense to convert is called the conversion threshold.
2.  **Participating Preference (The "Double Dip"):** This aggressive structure allows the investor to get their original money back *and then* participate in the remaining pool according to their ownership percentage [cite: 41, 42, 44, 45].
    *   *Example:* An investor puts $5 million into a startup for 25% ownership. The company sells for $20 million. Under a participating preference, the investor takes their $5 million off the top. The remaining $15 million is then split. The investor takes 25% of that $15 million ($3.75 million). Their total payout is $8.75 million, significantly reducing the pool left for the founders and employees [cite: 41].

### The Danger of the Preference Stack in Down Rounds

The danger of liquidation preferences compounds drastically in companies with multiple funding rounds, creating a "liquidation stack" [cite: 42, 46]. In a standard seniority structure, the last investors in (e.g., Series D) get paid back their full preference before Series C investors receive anything, and so on down to Series A and Seed investors [cite: 44, 47]. Common shareholders sit at the very bottom of this waterfall.

This becomes a catastrophic issue during a "down round"—when a company is forced to raise capital at a lower valuation than its previous financing. In 2025, startup down rounds hit a decade high, accounting for roughly 16% of all venture deals [cite: 48]. 

If a high-flying AI startup raised hundreds of millions at a $2 billion valuation, but growth stalls and the company is acquired for $300 million in a distress sale, the investors will trigger their liquidation preferences. The entire $300 million exit will likely be absorbed by the preferred investors reclaiming their initial capital. The common shareholders—the employees who worked for years believing their 0.5% equity stake was worth millions—will receive mathematically zero [cite: 40, 49].

```json
{
  "concept": "A visual schematic showing how exit proceeds flow to preferred investors first, demonstrating the severe risk to common shareholders (employees/founders) in a moderate or down-round exit.",
  "reasoning_for_value": "Liquidation preferences are a highly abstract financial concept that directly impacts the real wealth of tech workers. A visual 'waterfall' analogy makes the 'who gets paid first' hierarchy instantly understandable, clarifying why a billion-dollar valuation on paper does not guarantee a payout for employees.",
  "title": "The Liquidation Preference Payout Hierarchy",
  "visual_type": "Conceptual Infographic Diagram",
  "generation_method": "IMAGE",
  "justification_of_choice": "An IMAGE is required because this is a conceptual process flow, not a direct plotting of X/Y data. A schematic illustration of a 'waterfall' of money flowing through tiered buckets (Senior Investors -> Early Investors -> Employees/Founders) visually communicates the concept of preference stacks far better than a standard code-generated chart. Alternative 1 (CODE Pie Chart) fails because it doesn't show the chronological, conditional order of payouts. Alternative 2 (CODE Stacked Bar) is too rigid to show the 'if-then' logic of money flowing only if the tier above is satisfied.",
  "caption": "In the event of a company sale, venture capital investors with preferred stock are paid their guaranteed returns before common shareholders (founders and employees) receive any proceeds.",
  "data_specification": {
    "source_snippets_ids": [21, 23, 36],
    "data_structure": "Conceptual representation of Senior Preferred Stock, Junior Preferred Stock, and Common Stock.",
    "mapping": "Visual hierarchy flowing top-to-bottom."
  },
  "design_and_interaction": {
    "layout": "A vertical flowchart or schematic. Top level: 'Total Exit Proceeds'. Arrows flow down to a bucket labeled 'Series B/C Investors (1x Preference)'. If the bucket fills, an arrow flows down to 'Series A/Seed Investors (1x Preference)'. Only if that bucket fills does the final arrow flow down to the widest, bottom bucket: 'Founders & Employees (Common Stock)'.",
    "aesthetics": {
      "style": "Technical & Schematic. Clean lines, professional corporate look, avoiding overly cartoonish elements.",
      "color_palette": "Background: White (#FFFFFF). Investor layers: Dark Blue (#174EA6). Employee layer: Light Gray (#CCCCCC) to signify the vulnerability of the lowest tier. Accents: Green for money flow.",
      "additional_details": "Use clear typography for labels. Add a subtle 'waterfall' motif to the arrows."
    },
    "interactivity": "Static visual with no interactivity.",
    "animation": "No animation."
  }
}
```

## The 2026 Mega-IPO Wave and the "Figma Trap"

For the AI startups that do survive, scale, and navigate their liquidation stacks, 2026 is shaping up to be a historic year for Initial Public Offerings. SpaceX, OpenAI, and Anthropic are all reportedly preparing to list, representing a combined valuation pipeline of more than $3 trillion—the most concentrated tech IPO pipeline in history [cite: 50, 51]. 

However, public market retail investors and incoming employees must be incredibly wary of the lockup structures embedded in these modern tech IPOs. Historically, an IPO lockup period restricts insiders, executives, and early venture investors from selling their shares for a standard 180 days after the listing [cite: 50, 52]. This prevents the market from being flooded with supply on day one, allowing the stock price to stabilize.

In 2026, a new, highly aggressive structure has become the standard: the performance-based early release trigger, colloquially known on Wall Street as the "Figma Trap" [cite: 50, 53]. 

### The Mechanics of the Early Release

When a highly anticipated company goes public, investment banks intentionally underprice the initial offering to ensure institutional buyers get a strong return on the opening day [cite: 50]. When retail investors flood in, the stock price inevitably pops.

Under the new lockup architecture, the S-1 prospectus filings include a hidden performance clause. If the stock trades a certain percentage (e.g., 25%) above the IPO price for a consecutive number of days, a massive portion of the insider lockup is released early—sometimes in as little as 36 days rather than 180 days [cite: 50, 53]. 

This mechanism was executed flawlessly in recent AI hardware IPOs. Cerebras Systems, for example, priced its IPO at $185 [cite: 50]. The stock opened at $350, triggering circuit breakers and closing up 68% on day one at a $95 billion market cap [cite: 30, 50]. In similar historical precedents, such as the Figma IPO, the stock opened 158% above the threshold, immediately triggering the early release clause [cite: 50, 53]. By day 36, early investors and executives, protected by pre-arranged 10b5-1 trading plans, dumped millions of shares onto the open market at $80 [cite: 50, 53]. Retail investors, unaware of the early release trigger, continued to buy. Eight months later, the stock had plummeted to $22—down 81% from its peak and 33% lower than the IPO price itself [cite: 53].

### How SpaceX Utilizes Accelerated Lockups

SpaceX's anticipated listing utilizes a variation of this accelerated lockup designed to carefully manage supply dynamics. The company's structure reportedly allows insiders to sell 20% of their holdings following the Q2 earnings release, well ahead of typical timelines [cite: 54]. An additional 10% unlocks if Class A shares trade 30% above the IPO price across 5 of any 10 trading days [cite: 54]. Staggered 7% tranches then release at days 70, 90, 105, 120, and 135 [cite: 54]. While CEO Elon Musk has committed to waiting the full 180 days, other major stakeholders have the green light to exit early [cite: 54]. 

For retail investors and employees watching the OpenAI and Anthropic S-1 filings, the advice from trading desks is clear: search the prospectus for terms like "Early Release Condition" or "performance-based release." If the mechanism exists, and the company prices low enough to guarantee the trigger, the IPO is structurally designed to transfer wealth from public buyers to early-stage insiders [cite: 50, 53].

## The AI Graveyard: Understanding the ROI Crisis

Behind the headlines of trillion-dollar valuations and massive funding rounds lies a sobering reality: a vast swath of the AI startup ecosystem is quietly dying. By early 2026, an estimated 40% of AI startups that launched during the initial 2024 hype cycle have already shut down [cite: 55]. 

The failures of these startups generally fall into three distinct categories: the wrapper extinction, brutal unit economics, and an enterprise ROI crisis.

### The Wrapper Extinction and Brutal Unit Economics

Hundreds of startups raised millions in 2023 and 2024 to build narrow, vertical AI tools—coding debuggers, copywriting assistants, and legal brief generators. These were essentially thin software wrappers built on top of OpenAI or Anthropic's APIs. 

The business model was fundamentally flawed because it lacked a defensible moat. Once these startups proved a use case had traction, the frontier labs simply integrated the feature into their core models for free. For example, startups charging $20 a month for an AI debugging tool found their entire business model obliterated overnight when OpenAI released advanced debugging natively inside ChatGPT Plus [cite: 55]. 

Furthermore, unlike traditional Software-as-a-Service (SaaS) where the cost to serve an additional user is near zero, AI startups face linear scaling costs. Every time a customer generates a query, the startup pays inference costs to a cloud provider. Many consumer-facing AI startups priced their subscriptions at a flat rate to acquire users rapidly, only to find that heavy users caused compute costs to outstrip subscription revenue [cite: 55]. The faster the startup grew, the more money it lost.

### Gartner's Organizational Failure Data

The most systemic threat to AI valuations is not technological, but organizational. Despite global enterprise AI spending reaching hundreds of billions, a stunning report released by Gartner in April 2026 revealed that only 28% of AI infrastructure projects fully succeed and deliver expected ROI [cite: 1, 56]. Fully 20% of enterprise AI projects fail outright, and 57% of infrastructure managers report experiencing at least one major AI initiative failure [cite: 1, 56].

Crucially, Gartner's data shows that 77% of these failures are organizational, not technical [cite: 1]. The breakdown reveals deep strategic flaws in enterprise deployment:

*   **"AI Without a Home" (41% of failures):** Systems are built and delivered, but never operationally adopted because no specific business unit takes ownership of the workflow [cite: 1].
*   **Business Misalignment (34% of failures):** The AI model performs exactly to technical specifications, but solves the wrong business problem entirely [cite: 1].
*   **Lack of Measurement (61% of failures):** Projects were approved based on projected hype, with no formal framework established to measure success metrics after deployment [cite: 1].

Compounding the issue is a severe dual bottleneck. Among organizations facing AI setbacks, 38% cited a persistent skill gap in MLOps and model governance, while another 38% blamed poor data quality and limited data availability [cite: 1]. Organizations assumed AI would immediately automate complex tasks and cut costs; instead, they discovered that an AI model is only as effective as the proprietary data it trains on and the organizational discipline guiding its integration [cite: 56]. 

If enterprises cannot demonstrate hard ROI and begin pulling back on software licenses, the downstream effect will crush the revenue projections of the application-layer startups. Without application-layer revenue, the massive capital expenditures by the hyperscalers—and the corresponding multi-trillion-dollar valuations of hardware providers—become mathematically unsustainable.

## Bottom line

The 2026 AI market represents a historic deployment of capital, driving unprecedented infrastructure development and real profitability at the foundational layer. However, the ecosystem is deeply bifurcated: while frontier labs and sovereign defense contractors command mega-rounds, specialized application startups are facing severe down rounds and brutal unit economics. For employees navigating equity offers and retail investors eyeing upcoming mega-IPOs, extreme caution is required. The mechanisms of wealth transfer—from aggressive liquidation preferences to early-release IPO lockups—are structurally designed to protect institutional capital, requiring participants to look past headline valuations and scrutinize the underlying deal terms and actual enterprise ROI.

***

## Sources

1. [Intuition Labs: AI Bubble vs. Dot-Com Comparison](https://intuitionlabs.ai/articles/ai-bubble-vs-dot-com-comparison)
2. [The Next Web: AI stocks and dot-com bubble comparison](https://thenextweb.com/news/ai-stocks-dot-com-bubble-comparison-market-outlook)
3. [Content Fulfillment: Comparing the dotcom and AI eras](https://contentfulfillment.com/2026/05/18/comparing-the-dotcom-and-ai-eras/)
4. [Investing.com: The Dot-AI Bubble](https://www.investing.com/analysis/the-dotai-bubble-why-2026-feels-like-2000-all-over-again-200668624)
5. [Pareto Investor: How the AI bubble compares to history](https://paretoinvestor.substack.com/p/how-the-ai-bubble-compares-to-history)
6. [The AI World: Mistral AI raises €722M](https://theaiworld.org/news/mistral-ai-raises-722m-to-build-europes-ai-infrastructure)
7. [Mistral AI: Series C Funding 1.7B€](https://mistral.ai/news/mistral-ai-raises-1-7-b-to-accelerate-technological-progress-with-ai)
8. [GEND: Mistral AI targets €1B revenue](https://www.gend.co/blog/mistral-ai-e1b-revenue-2026)
9. [Cloud Summit: Mistral AI $14 billion valuation](https://cloudsummit.eu/blog/mistral-ai-14-billion-valuation-europe-turning-point)
10. [FinTech Weekly: Mistral AI €2B funding](https://www.fintechweekly.com/magazine/articles/mistral-ai-2b-funding-european-ai-independence)
11. [Investing.com: DeepL considers U.S. IPO](https://www.investing.com/news/company-news/deepl-considers-us-ipo-at-potential-5-billion-valuation-93CH-4268104)
12. [Sifted: DeepL IPO 2026](https://sifted.eu/articles/deepl-ipo-2026)
13. [Tech Funding News: DeepL IPO plans](https://techfundingnews.com/deepl-plans-ipo-for-late-2025-whats-next-for-german-tech-exits/)
14. [Tracxn: DeepL Funding Rounds](https://tracxn.com/d/companies/deepl/__mTFDDSjOlFbUE5DY_w1NWY3gI5ttRRBOXtV7p6GCqJE/funding-and-investors)
15. [PitchBook: DeepL Company Profile](https://pitchbook.com/profiles/company/56947-42)
16. [The Standard: Moonshot AI hits US$20 bln valuation](https://www.thestandard.com.hk/innovation/article/331387/China-AI-startup-Moonshot-AI-hits-US20-bln-valuation-after-Meituan-led-financing-round)
17. [Tech Funding News: Moonshot AI $2B funding](https://techfundingnews.com/moonshot-ai-2b-funding-20b-valuation-kimi-llm-china/)
18. [BigGo Finance: Moonshot AI valuation](https://finance.biggo.com/news/TVo_AJ4BrAZSr0oS2gOz)
19. [Intellectia AI: Chinese AI startups Q1 2026](https://intellectia.ai/news/stock/zero2ipo-research-chinese-ai-startups-raised-162b-in-q1-2026-up-185-yoy-led-by-top-ai-labs-including-moonshot-zai-and-minimax-karen-tiansouth-china-morning-post)
20. [Pulse 2.0: Moonshot AI $2 Billion Funding](https://pulse2.com/moonshot-ai-2-billion-funding-round-values-kimi-chatbot-maker-above-20-billion/)
21. [Angel Investors Network: Liquidation Preferences](https://angelinvestorsnetwork.com/startups/liquidation-preferences-explained-for-founders)
22. [Looking for Leverage: Guide to Liquidation Preferences](https://www.lookingforleverage.com/p/your-complete-guide-to-liquidation-preferences)
23. [Eqvista (Medium): How Liquidation Preferences Affect Founder Payouts](https://eqvista.medium.com/how-liquidation-preferences-affect-founder-payouts-d9a2cbc33cbc)
24. [Carta: Liquidation preferences](https://carta.com/learn/equity/liquidity-events/liquidation-preferences/)
25. [Ashfords: Breakdown of Payout Dynamics](https://www.ashfords.co.uk/insights/blog/liquidation-preferences-breaking-down-the-payout-dynamics)
26. [IG: SpaceX, OpenAI, Anthropic 2026 IPO deals](https://www.ig.com/en/news-and-trade-ideas/spacex-openai-anthropic-2026-ipo-deals-260520)
27. [MarketWise: SpaceX IPO and the Figma Trap](https://marketwise.com/investing/spacex-ipo-2026-anthropic-openai-figma-trap/)
28. [INDmoney: SpaceX, OpenAI, Anthropic IPOs Explained](https://www.indmoney.com/blog/us-stocks/spacex-openai-anthropic-ipo-explained)
29. [El Pais: IPOs threaten market concentration](https://english.elpais.com/economy-and-business/2026-05-28/the-ipos-of-spacex-openai-and-anthropic-threaten-to-drive-wall-street-to-bubble-like-levels.html)
30. [Medium: Wall Street braces for trillion-dollar IPOs](https://medium.com/@wealthy_mindset/wall-street-braces-for-the-trillion-dollar-triplet-spacex-openai-and-anthropic-ipos-that-could-0eec2819195d)
31. [Jobs2Rely: Evaluating 2026 AI Equity Packages](https://www.jobs2rely.com/blog/evaluating-2026-ai-equity-packages-senior-tech-leaders)
32. [LastRound AI: Startup vs Big Tech Salary](https://lastroundai.com/blog/startup-vs-big-tech-salary-2026)
33. [Fonzi AI: Equity Compensation Explained](https://fonzi.ai/blog/equity-compensation)
34. [Inc.com: Startups Paying AI Talent up to $400,000](https://www.inc.com/bruce-crumley/startups-are-paying-ai-talent-up-to-400000-and-equity-is-no-longer-the-main-lure/91325618)
35. [Carta: AI Shifts in Compensation](https://carta.com/data/AI-shifts-in-compensation/)
36. [Phoenix Strategy Group: Liquidation Preferences](https://www.phoenixstrategy.group/blog/liquidation-preferences-preferred-stock-deals)
37. [Growth Equity Interview Guide: Venture Capital Term Sheets](https://growthequityinterviewguide.com/venture-capital/venture-capital-term-sheets/liquidation-preference)
38. [Ashfords: Liquidation Preferences Examples](https://www.ashfords.co.uk/insights/blog/liquidation-preferences-breaking-down-the-payout-dynamics)
39. [AngelList: Liquidation Preference Basics](https://www.angellist.com/learn/liquidation-preference)
40. [Medium: Ultimate Guide to Liquidation Preferences](https://medium.com/@CharlesYu/the-ultimate-guide-to-liquidation-preferences-478dda9f9332)
41. [Crunchbase News: Biggest Funding Rounds (Anthropic $65B)](https://news.crunchbase.com/ai/biggest-funding-rounds-ai-anthropic-65b-dominates/)
42. [Sophic Capital: AI Infrastructure Spend](https://sophiccapital.com/may-30-2026-ai-infrastructure-week/)
43. [Trending Topics: VC hits $297 Billion](https://www.trendingtopics.eu/vc-hits-297-billion-in-one-quarter-ai-swallows-81-of-funding/)
44. [Qubit Capital: AI Startup Fundraising Trends](https://qubit.capital/blog/ai-startup-fundraising-trends)
45. [Today's Startup News: Cerebras IPO & Defense Funding](https://www.todaysstartupnews.com/news/startup-news-weekly-roundup-may-10-17-2026-anduril-cerebras-ipo-cowboy-space-voltagrid-defense)
46. [Entrepreneur Loop: Helsing Funding](https://entrepreneurloop.com/europes-defense-ai-giant-helsing-funding-nears-1-2b-at-18b-valuation/)
47. [Sifted: Helsing €450m Fundraise](https://sifted.eu/articles/ai-defence-helsing-450-million-fundraise-news)
48. [Tech Funding News: Stark seeks €300M](https://techfundingnews.com/german-defence-unicorn-stark-seeks-e300m-raise-as-peter-thiel-doubles-down-on-europes-drone-arms-race/)
49. [Bits & Chips: Defense as a Software Business](https://bits-chips.com/article/defense-as-a-software-business/)
50. [Studio Global: Why Helsing's $1.2B Round Matters](https://www.studioglobal.ai/discover/answers/why-is-helsing-s-reported-1-2b-funding-round-69ffee796b0e44f35d613357)
51. [Tracxn: Aleph Alpha funding](https://tracxn.com/d/companies/alephalpha/__9w8DBslRjv4B4lTYAdG7XkrtnnWhgssw3HF9CXIPjTw)
52. [Tracxn: Aleph Alpha Investors](https://tracxn.com/d/companies/alephalpha/__9w8DBslRjv4B4lTYAdG7XkrtnnWhgssw3HF9CXIPjTw/funding-and-investors)
53. [BuildFastWithAI: Cohere Acquires Aleph Alpha](https://www.buildfastwithai.com/blogs/ai-news-today-may-28-2026)
54. [Clay: Aleph Alpha Funding Dossier](https://www.clay.com/dossier/aleph-alpha-funding)
55. [Sifted: Aleph Alpha Pivot](https://sifted.eu/articles/aleph-alpha-what-comes-next-news)
56. [Data Center News Asia: AI infrastructure spending](https://datacenternews.asia/story/ai-infrastructure-spending-to-hit-usd-37-5bn-by-2026-says-gartner)
57. [ByteIota: Gartner AI Infrastructure ROI Crisis](https://byteiota.com/ai-infrastructure-roi-crisis-why-72-fail-gartner-2026/)
58. [CFO Tech: Gartner Sees AI Spending Hit $2.59T](https://cfotech.com.au/story/gartner-sees-ai-spending-hit-usd-2-59-trillion-in-2026)
59. [Christian & Timbers: Gartner Trough of Disillusionment](https://www.christianandtimbers.com/insights/why-does-gartner-describe-2026-as-a-trough-of-disillusionment-year-for-ai)
60. [Gartner: I&O AI Projects Stall](https://www.gartner.com/en/newsroom/press-releases/2026-04-07-gartner-says-artificial-intelligence-projects-in-infrastructure-and-operations-stall-ahead-of-meaningful-roi-returns)
61. [JD Supra: Q1 2026 Venture Financing Report](https://www.jdsupra.com/legalnews/q1-2026-venture-financing-report-39-9b-4833136/)
62. [Startup Grind: Term Sheet Liquidation Preference](https://www.startupgrind.com/blog/investment-termsheet-liquidation-preference/)
63. [Writer: AI Adoption Survey 2026](https://writer.com/blog/ai-adoption-survey-2026/)
64. [Lord Abbett: 2026 Investment Outlook](https://www.lordabbett.com/en-us/financial-advisor/insights/investment-objectives/2025/2026-investment-outlook-riding-the-tailwinds.html)
65. [European Central Bank: Financial Stability Review](https://www.ecb.europa.eu/press/financial-stability-publications/fsr/special/html/ecb.fsrart202605_03~90365ff1cb.en.html)
66. [MarketWise: Lockup Performance Triggers](https://marketwise.com/investing/spacex-ipo-2026-anthropic-openai-figma-trap/)
67. [ITiger: Wall Street Banks Rethink IPO Lock-Ups](https://www.itiger.com/news/1186978332)
68. [InvestorPlace: IPO Trap (Cerebras & Figma)](https://investorplace.com/market360/2026/05/dont-fall-for-this-ipo-trap-heres-where-the-real-money-is/)
69. [Gotrade News: SpaceX IPO Lockup](https://www.heygotrade.com/en/news/spacex-ipo-filing-ai-tech-ipo-race/)
70. [Blazor Boilerplate: IPO Trading Frenzy](https://blazorboilerplate.com/first-dry/SpaceX-OpenAI-and-Anthropic-IPOs-Poised-to-Trigger-Wall-Street-Trading-Frenzy-19-8436)
71. [IdeaProof: Startup Failures 2026 Database](https://ideaproof.io/startup-failures-2026)
72. [Medium: The Real Reason AI Startups are Failing](https://medium.com/@aiempiremedia/the-real-reason-ai-startups-are-failing-in-2026-30a4cc9fd140)
73. [Reddit (r/ArtificialIntelligence): 99% of AI Startups Will Die](https://www.reddit.com/r/ArtificialInteligence/comments/1n0rj68/99_of_ai_start_ups_will_be_dead_by_2026/)
74. [IdeaProof: Failure Database Statistics](https://ideaproof.io/startup-failure-database)
75. [Medium: AI Startups and Infrastructure Reality](https://skooloflife.medium.com/99-of-ai-startups-will-be-dead-by-2026-heres-why-bfc974edd968)
76. [Reichman University: VC Term Sheet Trends 2026](https://www.runi.ac.il/whats-up-lobby/news/the-rise-of-ai-vc-term-sheet-trends-in-2026)
77. [InvestorDaily: Europe AI Boom & Unicorn Surge](https://www.investordaily.com.au/europes-ai-boom-fuels-unicorn-surge/)
78. [Crunchbase News: M&A Outlook and Down Rounds](https://news.crunchbase.com/ma/crunchbase-predicts-merger-acqusition-outlook-2026-forecast/)
79. [Tech Funding News: Down Rounds and Valuations 2026](https://techfundingnews.com/founders-guide-down-rounds-valuations-2026-term-sheets/)
80. [AlleyWatch: AI Down Rounds Forecast](https://www.alleywatch.com/2026/01/161759/)
81. [Crunchbase News: Biggest Funding Rounds Recap](https://news.crunchbase.com/ai/biggest-funding-rounds-ai-anthropic-65b-dominates/)
82. [Fundraise Insider: AI Startup Funding Highlights](https://fundraiseinsider.com/blog/ai-startups/)
83. [Wellows: Hottest AI Startups of 2026](https://wellows.com/blog/ai-startups/)
84. [Intellizence: Top Startup Funding Deals Q1 2026](https://intellizence.com/insights/startup-funding/top-startup-funding-deals-of-q1-2026-record-297-billion-raised-with-ai-dominating/)
85. [AI Funding Tracker: Top 50 Funded Startups](https://aifundingtracker.com/top-50-ai-startups/)

**Sources:**
1. [byteiota.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQHBmaGXbSHfaOyk9T4750yODc8znU-YzNBwebQR0MNDLkqRtkXCSDt7mWCNmie-4XSPvi8rOowfZqGjQkXK9szY0UpDAyCOFecFZn-6tulqXRduaVriCK3RxnYMU0JfLZ_kiY5WIybV3WptupQilb9BsxPcWZQb_a7VaSGijEAIDg==)
2. [cfotech.com.au](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQHitQ_LGUwbd6DXbbmPdI5V60A4YvwcvFXx9bUgmYMZuks1B3FTL_nWsqbC4xEJ7VfZRv6h1Ev7PWFngbNQOYIETy8phqXjUm5bbrUfYntfECXBbebS3knBb9iStVvAoEMBkZhKNt2smAEnErpvs4mAFBdcaM65P2pfEGPPBmgFOVL8h40h94hf)
3. [thenextweb.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQGn7WeijSXvw_k-idbMLMz983gPm-h996SdVa_I80-cWOojSW6lJHJifoGY3rjfgkgv2IrG7_joz_KHj0h_YviQ9rVjTm3-UkyTQs75RexlHy3Hvf1gRM6A6xghNAmwn2w46i_avgaqqDvirkN5AczmCsjHbcrgCIKsvB9M51NXNh9sdA==)
4. [contentfulfillment.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQGpWuKhLZ8LkkY4jlhxMODpVkMAUoGr4PFwFyUborPWtHpsqP7nlhAgOsDZVrDH_U26As07uMw4ZkUsCiX_Iwqf1Sz8M9ub04eDbxpxMVM5FGxgb43bnGKD2i1wf4yVT19W4dwrWZP_U9qAju8kYQvI846bWwwttzxeUdv-sgeo6g==)
5. [investing.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQEqjzpt5sQUJ1kbk3JdsksZXS515rqJ46fmg788kygm20kOt4XX3S4eVO6X1N26LHWrwxmTigvHFIi8IH8ylxOchPGBPKPXT3JJPUbP7rOwZatyvGSDqYd8BfGIHSok3zBlYYYQDSnsYbF1TZ0biybvfzBu9NDYI_unJKLOuh_J6lZLW8bBBaN3VCXTJNlksFoHuv4NEW4V2auQ)
6. [intuitionlabs.ai](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQEF5sZjgg4VYiDOwvznL_tNxveLx3W7kZvNs4H0hdfzBytw8ZjmNUytYAj0O3lCIpFCalwUU6d1et814iNDS1BSjHEfVTFbWnVnBsdphK-4dMYNOQLnuQU5xYz3GIRu2srxNnNXG1f6Sih-TxihnLoxbDh7NKhw)
7. [elpais.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQEqv0yLAeWgC4ypF9w9wBjL1quDjWRO0I7QuvBddsrHAgkA0NWNZaX8dq3c6GzMlzKXFcKVK76nlzmnAjvGj6Jr1pej1GxTkD7RSr4NEJwnub3m9U-Da42M_mhtITHm0aqDbVrDg0Ukdbzm58LtCVuy4jOHgx7uKbcF1N_FfrepzBivuCB5HnxgTUTq5UPMD7VI5MEoF7ncKRI6U5PpgzGYttGZmuiiKI1XpEBK69jeXIWwtJcYpHLIe8UoIDnqT_0CnVgtn56dBje7XsxsXlz8gQ==)
8. [datacenternews.asia](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQGr18izTGjDdJivSN6qDCkFcRoKUTnSvgBrFH5X1E1HHdHYVbYxHVWROuK-PNZxYsrbl_KrX40HruUMcOzF5rhBPVghU081B-s-iD8iDqW8xuNtLIKZ00xX05hjZiiF5sYsc-djg_EFZ7tSA_yaFaVCSapRCKjge0-REZ4t12LOx_tcIHAqB5MIaNIF58xO2gVGPje14WJ2ow==)
9. [christianandtimbers.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQH17CdV6ldxOzjtxdc_zKV2BsxR08ALpueJwYFk9ITExgnLIe9kFKC4d0PfDz0o23JmcvhV_FN1jxl62trMIdmtHwvTnxFyCeJ8qDIJAXKjMieyyLT0V6LMDY6GA7e3TsGGwMB0HvzICn39zHEp9SF1xU16q-yoDLAdJYkeLxXQ48Aqq2hOGmNdTUb4x7gtD5E8allIM3L7kC04FtPrTFGRujpF6sUIHetrwWA=)
10. [trendingtopics.eu](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQGeQi6S_07kaOosqUkpdAJIyqzZScO4DtY8jSZJGS3ZFURjoKe7y55l9xkIrxxz2is2--aiL-XQ0sCdr_zJgYt8pn_GnTYFWOzE861jhOIjMmWk_Eig5kJ_oYreIs3dYUpWE8pA-kDtl-q274GCWT81aqY3dVnnWFRsVVlb-ka4P6f35bn8JAPYJpHrWPBCNqU=)
11. [intellizence.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQE0KKxNkMfAo8vpnh0t5aoDy53d1JFgv0T2f-OvY4HsWtVFD_JuxSb1U4tZhykcwVO4447wv5r87Sx6ZUGhqgxl2_jVmDc969xxnFq9Q1SoJNXc39A5a7PCpNks-r5aErnYV_0qcWPoZAqP3KWx6kzPAO7tI2ixsrfqP8pLPjO9mha-NZCiJk_EIhj_6KtYYnEwb3dNjHAshltWtPVY0TWdxG47vYl3Y7o5hDgyS6L-KUg-NPx0TJGuvJQ=)
12. [aifundingtracker.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQGFFybFeU-UU4QqjNPPBe8pI8KmTUTyztmv6jJupo2QSVNr0mORtsK4TlLDz8f9jCNjpNzG2mILODYLDmp0S-m4JEDs8PX--lxT9IGxZ2VkVuxpAYnUEGGsvTgB9ATHIontO31DVA==)
13. [crunchbase.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQE_loQKSG5NzwRArlWzULvbrnslWTJObR1CaCk0edmgDtOByFpCaWTIeih6ABM_aWCn5dzjEgMBEyqZSHk_LRvBHYxhnQogS5vsaLAsIVTVEt39yVT869Zbd7YHEJ496uwK9tPo8_aZJYui93zxR3xnOHnl7hKDAShn_RtnSmwdvAi1aQWUmw==)
14. [buildfastwithai.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFmU1Zxl4z47UPM0HMaGlBv5T6zP7rg23ejnonNEQTBk54SUWYAaT256_MFu3lC9Qx60xdVRfQBdjvGSSEA2HsaQ9Oq6gMpBLJDZ3x5nbM_PpPN6thIXJxO9JZmWJTvjLng0AGzdpU70SGj4cb6g35YWWsGwg==)
15. [mistral.ai](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQF_gYwbb3k3-NQldFBtF_TTX-6ZmoonbE4n5yXXwuCKE523O7K3Ic5DiwHdbkRTD8s1XPjh7XcvMOf08Bv_NNOIs_eL6t4Z6Tme2u_baeioFte5BpkZDo6UBdx8qqFetmIPLRNhUiQmyNp53CvcGaxGSkJO8IjXFbuBzyjwhpXrOgWXKFEf6C_2onN26x9tic41)
16. [gend.co](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQG3EfIc4LiptqCvbZ6DiYaY3vhnSGhJ4-VXuIMmQl1Xm-wcY-AMMdXmpvge-B3cwbugL-D2rEi3nFB1TC8Xt578DP84CoVz4Wm5UqWhKbX5KIsI9ptpj1Jz8Vdy-FBNp3cZ5n0Ewv7w1wA=)
17. [cloudsummit.eu](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQHtkRaWuqZtWCpa4nUgqIa-Oc1hGT8gRDVVy2Pps-wLSnf8WW86sYwwmEwQS4bBRT1y2pkkmayFFTUf8GIizI7xZBtme1X55pEKoDNgrXnPZmP8iDBeyj-QxWqlW-WuQM_x7CJelZfmICkdrwhtk1zYEeMuWR5CbvRibblz11EOG0Eq6kfS)
18. [fintechweekly.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFU84vVTnxPBagCet6oKxJ5Wsi-_DZvFtuiWYs4dESwnyYI6gyfIim2-XZ957Q77xM9DSenjpdajkusK3O3gUep7qWi6XJRCgb76F3hTrMp4ocZ5bdvJhC6CzLCZc10zx2sxKgeCnjLT98bO72VwBtT8IKIoTweczibDkCSn1Y_XIsWsxmdTmzgem6hSFV1EoGw06A=)
19. [theaiworld.org](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQHzVcCbEOmYkI7wy6YBQSN-EqqEi1IhsSV-QoYKzJAVtlVg8MfzVXus_slCQNOM-P7OjLqbKbRpEXbr1fx8wDxVBF8Pusj33dzxNyskZyGfkCYrgqWsc_mLxqdMlHQmk2Q1GMIiKXlvlryhxl9-1E7I3ukCx6BgeyznZ9HZP78hwLsuBC4FGAS-CBo=)
20. [investing.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQE6s7D7V44oFYNQnN_PJI592QzFEfgbRccm3DdXnmu6fxVQk0-aAds9kRe_ok48XZI_Q1KI2M9WKBslSTPMuBAwhmjfCS-huxUH7iL_iW4hIUTjSFCAX729sMoWcIgsGtOUv7nTdXSg1Cch1eqVx0ZQLCTgnREd9eDdZIq15fNDN1cAr_0BlD6MPPmyx_zn_obOe70nCcmmAmlmBsLGxS4asnVuPu4=)
21. [sifted.eu](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQGEiBrdksE8HhEsSQpj3PNbmqv7xKHVBwrxfqYERdTeriqGjUoMyoTqL8vZsPCBCI1KtgKSIn0dmdhcBh-4GZXkyEQXd2mYF2MYw7X3FDkwFhoYDvFdbZWjZmPbpAa0)
22. [techfundingnews.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQHVHxeVjEoec-BDP6eN_eEc_6gIGJIBts_GUKU8ulWku_vJP1hMOIn7erOz_Si-QO01wuG8C7fBPKqnSmxwgg2fhhldNycTxaU1UC--hMiEGA5hhmDTlEdLTVf13942WUCaC7wzvqDq297O0_jn1zX4CF96UvmvWLy2xghOEmuCxLU___-z3-WvD8e_QVycF2I=)
23. [tracxn.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFgJokAB3X8-0IHfPcW5FtDnAXjCOTOZA4r444SB9XcwnNlfgG5JKUx3VceowFkvUA1zU5d_GY9pLGF-Cop7K460D3LlGWqRCeGyN9Q4LakMWA0lHoSMPHiigIDIP1td5e9vB-qL-Jr1WKyZmvdFZieg4b9HfCQkKx2sCUVUSwpSYA62Lt5C8cRk4gjYn33oBpTShCM8EJsC1vyarFT)
24. [sifted.eu](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQE5kISezV9Lsvv9CR-rYuAeO2ubkwa9BijjYdmnKE6l2WcGgKqzwAlGg2yN81RwX9uEXBNwzIYlsHE9nL4C80k0cN7r2DkkThjt8QrfM2NdSGDzepr_7iAxcauYiX6nDF1ZlhY6Pqd3mSsEv8nxEukO)
25. [tracxn.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFyw5DpVzVUoXr_4_YCSRw6t57CvrS0mKWy8zDqkmZRcxzZf0G3OXnPMC3Jp_eTuu4pH0FK_T-ozY6Z5uyO9L0W8f5YFJ60kaqGtbNRPAgiynoV-c2l4hVwWgDDAmYorF-2o_O73dBkMW4n7Pd05Io0ColYfBCI8QjuFKY3rVadmc8XATK-HRShivWk9Q==)
26. [bits-chips.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQGc8grebPMGFw_uhVJbR3UveXJ5am3NvLBxSYyI3AbNxjkElftWUasCjKhjiIv9x2CfLv1kIBRI9tU7kFOvlXZZXk3fJM6ZyT-eB4GHTlQa8JIDXiMPyFulZrz8y2q3cQjmGm0aO7Wh3JGW_g6iP-_-0-xL)
27. [entrepreneurloop.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQF39fLa0EvLqZR1Irjr4HPioRq5mZyXGW97K2QttdF90pyJ-nJOZOM35eWabLwFe4xROuTqpqVLHSdUppAjXSf6Y-DTJNhzvDyDyRVkZfcZxHXEYjokJ5k_RnQYYHzG3yJaQZiw9V7WpKz2kLsN_eBUvPNjdR5_dp8P1Ix6SeBFxjP4lnh2U1b7rjdtIXfcN3SpCJe_0qyu)
28. [studioglobal.ai](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQEcDE-12Ys0K4tLp71ekfBtk0H7ddPx6JxOl0GKzc0hqcITQ9IQ9Z3Wnh1TUpQ9ritatdUpDyalMTuzTcK_2SjLwDMRIn52hxYI1eXFYsMcMFii3CDQUxqcGgS2nheqt3hCVsIwflgl_xupH8oLQ795U_l3_f-IFzCl2wcRjR4QS97VPd-gmBeb0D4dmT6ZTMLpnIuYbSCYULTVevablPL93fVRjiqy8g==)
29. [techfundingnews.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFrUiOA10QgPcf2HPzEUdcZZER2xcm295bfhi4wrwfpbiJ1-0KaHxXxeg1umzAyiw6TeQIcz7V3e-JLFmXYmYPuMJAfyXayev41Wja6toLr-g2tQYTZdE-xA0xKo9YxmtCZMVhZId6n5Hz2g4LtSmDVgil9W9I0PyKK2voJDbYg0GhNGHUDl3W5dJ8OXYnLaYdIP5WcHfDCHUHjvrZx2jQxJIHWshKzIy55Rxi7RvGCK879Fog4VGY0)
30. [todaysstartupnews.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQHpqJn5zGYzQufcXEVJl6ThiUnnPT9r6ksQeIzbL9Q37RIDZF2wwdDo1TR67phdlk6KiD3cgIh8q1O0AC4GWuhydvYAbKetosnMxMYBnFJZdFhGIXHjZHPS0llYIBc10p3o8BjiPZ1VSXQUyzyCL6KX0Xmp5QqEUPF4MpP3CA2kBb9RBlIfO4tLJ2_TNWcbhaDJ8ROFeXO0THIUEfta0ZHRC11sKJADGMgL8WzgatiSBUqdvtZc6GS0gXf0)
31. [intellectia.ai](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQF-a8XM3LTDz_4sd6WN3pjNhvaDnTuQME2opjPMiUdrsM-oG6Bhb3TW2NdKe93J5PtDjK3cYWZK9zFzBstWE6SNkkpCCsP6_GzleHScKz49E1AXwunpeFIiBghNwuqg8QOXznWvgPKj0mCg5KPZf5nX-iV2P8gQwA_9O5FscMC0DuV99hqbnHVNpwHhpM1qZHZfm5FeFfg65tbh5UAGJoWaPn31od9uukNwgXgZ0gg0IDlhMhKSxlLlVhX6YpHv95DfRPO4Eq9Jt7Pnq7mhgWMzAIl96HBE8NIfwAEYQjNzTb0gtjJwZ3J-fuqrxo7eNggV9m8SgztmqA==)
32. [thestandard.com.hk](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFz6nLsB_VcFc0d9OzK_6GCUM7XnHw6uQFzK_Yc0jtTepqbYycXR6iJ4FmKUN7w4_4f_uYVCmOckMKJsfAE8CW-AI2hamUSWnPt1w4_v4grfM1GILyRc-4ar4wZY3xFs0Ay_lukRDIpWG8J0dYbPqxJCvUU_2LtGihiqpTd4P_AU2j0HFoQTnoE-QKBrjhoQOBOtCTfBReUrwWp2Z8cTqP0zm93E8Z8JK1I4WHbYIW9KYaprVGHaDBtGW0HtYuT-BJiPapFxg==)
33. [techfundingnews.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQHz38vmuzJJOlsCHukIVCBTJGRBLnmXMkx_Ng_eI1rmAq7g47S6mAabF6PkHsmC2QCnKiiHd1l4Y5qBmG_1ATRoUR3FcZI8EFGIrXIGXFu81Z5GyyR0tAuSN-1aB0jq2DgcV7_Le6k0CN9tAVd4ZvuniTzjatDH8sLyszBMp136PT4dXfsD)
34. [pulse2.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFyQe7QrZGSW4cWa2UaJVyFctYcm73smfUKQQD3J-htlrA-JnE17rqeN5oARXeo3VO8oEQ2qNiioeYxUVcgp1MOA1VadjUpum27vRj1aT74gPRQ3qJvd3W2_6xZKNsOylw8llClM_c52vS7JLDHn1osXzEAYgl75KDIYpEzhDpOT5bdUnXzXBI3jKjy9hz2w8iaiyGV0Ij3)
35. [biggo.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQHohKX43wnancYRn7QGrZfCOrhg0mPMKfyUjzqHs1dUcJSYn9g6-x2CbpUGa6kdDsNHlFZTiL8_FyKFCxZtljez8d6Svruiip4jy8KZ4G1XNnNBq-hrOWDl2NBNKnkyzFq5TE28VB2OcA==)
36. [inc.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQEVduXb5Vv_y_BbCqxbYgCF1uE3dtVCJq0PuFjFWrU5gihNXy7S-v5ILxp3yctAoPtwTt9b3dwNw2YYyxqk4a3BmtIXaYjM1u_Cfx_9CuIyfpLpPpWm8FtHskHSVL4s1AGreK-e4REkn5APLagYHSt3wMj1nt3v9ncnWlUNv1I2ocg_1uMLHcAFaROzACCk8qMb2RmMXpgTMckTErMjgnCWAhwZtEL0z8rtU34n1ebViA==)
37. [carta.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQH59c1sKvT_90weopxOGMcmEZMk6LW6tVFrN-dNy9bPTB3Xb3_wmAyeG-n4CbtYLpwsmN9Aeyj0YoZzEfAC80XYcJtwf2qtLEaJz8oIaIccCaf07BGTXWlhyURJZzr6LBPOa4xviDY=)
38. [lastroundai.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQGiCsX0LnvVeGOlHmZt6iw4TC1z_cv_xPZO26JGmp2tNBxT6ShHetrfVwBi9opSxHMRmfoK8b9w1sNIa4t1Dm2KXA1Xz6PoXJoxUJ2SnRBYLHt_6uHdVlIG9o26eO2zVV609DmfEEzGTn3XXZLRR2zhmw==)
39. [jobs2rely.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFyaJFrrsSOVHzL4BY3H6rdqToerf3o2wa9wzea5YmEbw_XvpZJSo6BLcTwBdhOzr5ahpTWspsuJ1y1JyLftByjhYnLrWn0NducUMoU8ZfOoUVeUHIX0SG501Bn3Pd2NRFXDTmfYPScLg7WIr_-fx6ZIvEHxK9WbuZN7rNGa82sFeHbtoeW2TxziCI=)
40. [angelinvestorsnetwork.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFyzzysvgubXGjzSKOKd62IY85aFambTdVn2UjoGVukQcA-tn8BI43EGIz2CVNUPXZnKQgJcUVJIu6qvSwSiya6pGRoVTVTmUBtv-u3fi_mrTCRwIYIdYu_8_ckcLD-U2ui97nVRr3h-WoD2pBbPL7v2Cxl9se_74Xsepo8ko-8z3hSWUxZTiI2U2ugwNWk)
41. [medium.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQHvbpcaOUbhBxPpldHA9BpOXm-D06dzyb0SpepSsrJ4Cv39141D_OdmdhVw-7vg9Rvy-jxNH4ZTh6_pRQNCOdaCB9FzOg5XeU6Hf1gARTmnkxGH0RSXmhjvz0i8iXHWJP4pKB905scIOIU48OiopeI_hjiyCMTB1Qrd-uq0cBqIiNcR0dIkR87RlIUdA82cLQ==)
42. [ashfords.co.uk](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFZqUL0Fy5oPcY-6cDYU0tlejORWaJLo6qIE0hm2IeKXhbF1hF22ONHKsd8AfQa1ozbGO0czvOMVdy_vL7eQYymukjSTTC_d2fMgxCJspUIps-5Nm2f3dvLUdVM6pkQqrOL39BP7lhRkzyX-u73rzQo325qsCjcxC22E-OK9BUxY3CPrkrfc4BWgQoH8l-LF7U77DMsEr5w)
43. [jdsupra.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFMkj_F4QHlL4D_JYExgpITPPivJR41XUeh73D6WNUFtLBU9Z0bJNW7CUgcAwRseJI8PbV0Cw6HEYHRGHzrVag83AF0K0hyhQ4nAlPgtJU7rpwrcslLrU8e4m7r4uXG38LMcdAMCTHWo821GYCoK3Uc6b9G7BVbZEaeZ78jxdD6YIr6SG8bBw==)
44. [medium.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFFmn8ELPuyipT2QRUAtMJmECJ0v8hTs7LliDSosiJWVyKRzGQDgBwDa0NXkgwbn5gPuy2U_VZbKS1Y_FSbyDodhSz2gqPFS-kr5JwE_IHV0AXUY3MSOJHBF15gVtPQQMnE0BnV0CPEiuLtXFlTEfhoMPK7dRBmIzPQ0oPV2mngpfNncWDPy-tQthX0p6Q=)
45. [lookingforleverage.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQF4MYxoHADzSZe_iFKLlxYbiWotFfPcc0KkyQrTtQbmxby3kb31W0Lv3MDN-IapyTTwAquAN1OXZKfnqecTaJwSowGIDVVaL_OawGbM_2zwPSFdaUjUZI3_8QL7b8sHY_PIx_zC3lSG-su7hc26RzpJ7HjEDQlFG7SI3pXzCDMgAO1Cn59CbQfe)
46. [phoenixstrategy.group](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFhmIPUay6qNue-EjvcVNVswSdrmEhH5uhV0leVm8OGR79xn4Bs0OI9Nmpa15JFynaCAEh6MO_jeU-Iu-12ThM24aOhMyB5ZRUI68CJ0176DIrKQYB3bVH_yMoJfvV6Yopr8GEXLARqoIln0UqKYeUiJvMiSgbiVhTwq85TPM8xvf7DcsQHz8ia-A==)
47. [angellist.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFovHpJINF68y_9BSoDV-LlePSw_-JMeCv_zvtdl7xJGVaDnkwCASkvsop_SpywxP3OQ4agAD4civ9F0fIrFzPcPdt-dRYB8x_0XJ7-f-weqRLl9xFJiw0GqsBgWM9NMChoGP9_igIDwAP2sw==)
48. [crunchbase.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQEwTnrs4Ayc5QdTcm_lIvZZ4i0FgFzMX5FcjaNj-PJk6n6oWDn7MR9qwlo4L20In_ayD22baFFRV3CsVFTmm-6CtH7Siv2NzyxXqLJhmU0zqImONoDOwjDtZwOOuCQdtNeoAOHzgAMh22IsE4U8voRQRndnb5AarMu0Vm5powOxfmuxKJ8wRqyG-F8BD1IFb88=)
49. [carta.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQHM95mcSRTwJ2RrcCQ8glM3ekQQlhKlYaBlF3X5YDcTXdea23Djia7xfdT5YKnNJbKN302Jp-AHLyUfnpANYD5hbuDzuKuy_TSHzc2nV64gNFScEg42r2k-CJBZHX2FCcKZe1smYYiYlRPeomKP0RSu14OK2ry1uQ9hLwfXXA==)
50. [marketwise.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQE7RX35R9R8dT4pd8mP5woQdWFVkwiMehZKoujPPq1SuaDm0ob0mQ-uEDOIIy8eIbDefLj6l-FX7Hp6sbP-9e2g2xylMSt0lcwr7invMDCX1G7qubw7iWup2NB0-Z_a_ehHYbyA4M4J7WnigcvosY9q2jOsLAI7w0h16H1hcoWZia6P)
51. [medium.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQEdeubHc4-ayQvbBFErLSbjvWt8GmVB0mmyB3vW01WI8fVJTq1o806xdzDFK2SeyoNOZNWFJwhQG_rV2LQaKiod0bqhlQnC2K3txXfk_8V_2399Wtk7j23QJQdDRIRlz_fQ89Kr4n8niY0sNZs1x7Pvsl6ruhi-BQ9vzyxy461pgu2xWDI1KJ5W2dnapU-G-jIm3fafG7WJw_Z5i1cjja37X_w6IAYNg-iHWI_e1OWCoB_GnSOscNXq9nJc_XkAdLD0w_hj0g==)
52. [ig.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQG59Blbt3wvA2NAmZoqzmD8Oqac41IPR8J-IowjDR3_sRN36mU2ltZ0bJTzHY19qhQX9WkEWqfASCyEl7S668vQc3t6Q9u_n7zvT6IWiH_6pLpo-FybGh_-LtR5QWCdknJHBh81xDq-WkEkOM8l0u7eSY_iLtDkB7VMBfCx04qykWqu3_Wp064g06STlqw=)
53. [investorplace.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQEiLFMAjs2g6fVPPrhasOyZPAjtnP_9aP8s9E3v67daNCspgO07PTK5qfPHQVAtNRGDXqcPvcvf8HJnrHMTb_4pbr6UGDcNEuO733l58jWHt-P4lzDP7UQtpjonvZl_iMiV9DnBknpDdqNEQl9Hnjezx9cipB9RQrXSUJsXMmz6tf_kXP6SvfZjZU75fHWtlHeBhIqqW6CF92V72w==)
54. [heygotrade.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQE-jZSZvKDkZsbrztbUBQ4hoCdGi4Z-LjDjS7J79sHFSEgJlJR7rUzsAb2hiPny2O0GR2vrjKBErmHChaZDaoD9qa96PuF94HCCJ72dFso4pL2xX3yNNiWblGSHNY_XWXta0-PLoOwtvP-_3qH5I825Ojuz343t7y2ncyM=)
55. [medium.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFFbSg_nNDKxYF7KBpnUvtWupaThB8_mQgYZ4zVGH25qER28GnNBOxB8ole6979BIL5oSN4opwgo96aH_lCn3Hh1BQ7zPyi9cTt1So4VgRMi0-yJKdV0nnk0oOsWjmRpiTKhEu8oWsVdDTIEeCYNqWky5ipEg7K4B2qBoxucgBgRvKgMJEn7M2enl_a7fZ2uwwxdfQ=)
56. [gartner.com](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQGc91l3Wb4Lq1OuA92aM_pm0SZ-MoHHNVzcn28KnOcTa3VyNNLRWr1_TaKE9SjLPSmPRiKFiJ5JOmiH0ss-LHeIHhh1yeZAQ6qT3AidwhFvRXQpV4TOIoQ20jm6btPa_4-Oy7969-_wciLR8DB3vjmgCaoB7BALi5TmtUVbhGcva0t480ZUlOCcugB_Z62kkMjrn5gurM66kbMssKFKIYC5DNlSYLd1TqCbwsM5IvLYcrgmVUtGzAIj7zTTuRGJUnkdJ4sArAELxvQ1yGraSmqPSe1_Wj6WGB-Wg2YI62HDMu_JmjZK6m74)