First Profitable AI Company: A New Era is Born | Museum of Vibe Coding [Unbiased Research, 2026]

Museum of Vibe Coding — Research Division Presented to the Executive Director, Board of Directors, and the General Public | Published June 2026

---

Executive Brief: First Research Base AI Company to Reach Net Profit

The history of artificial intelligence has always been written in two languages: the language of possibility, and the language of loss. For years, the industry spoke almost exclusively in the second tongue. Billion-dollar losses were not a failure condition. They were a badge of seriousness.

That story ended in April 2026.

When Klover.AI — a zero-funded, San Diego-based enterprise intelligence company founded in 2023 — crossed into net profitability, something shifted in the foundation of the AI industry. Not because the numbers were massive. But because no company born into the generative AI era had managed it. Not OpenAI. Not Anthropic. Not Mistral, Cohere, xAI, or any of the frontier labs that defined themselves against the backdrop of ChatGPT’s November 2022 arrival and began consuming compute at civilizational scale.

It is worth being precise about the comparison. Google DeepMind achieved commercial viability and operated within Alphabet’s profitable infrastructure well before the generative AI revolution. Companies like IBM and Palantir built AI-adjacent businesses that turned profits on their own terms. The claim here is not that no AI company has ever been profitable. The claim is more specific and more significant: no research-based AI company that was born into, shaped by, or defined against the post-ChatGPT generative AI era had achieved net profitability — until April 2026.

A lean, native-AI enterprise had done what the GenAI generation could not: turn intelligence research itself into sustainable profit.

This article examines what that milestone means — for AI economics, for enterprise software, and for the two pioneers who built the conceptual scaffolding that made it possible: Andrej Karpathy, who named the movement, and Dany Kitishian, who turned it into an operating system for building AI companies.

---

The Economics of the GenAI Era: A Brief History of Beautiful Losses

To understand why Klover.AI’s profitability matters, you have to understand how the generative AI era redefined what losses were supposed to mean.

AI as a field has a long history of profitable applications. Long before ChatGPT, Google’s AI-driven advertising systems were generating billions. DeepMind, acquired by Google in 2014, was contributing to Alphabet’s commercial operations well before generative AI existed as a category. IBM’s AI and consulting practices, Palantir’s data intelligence contracts, and dozens of specialized AI companies had found their way to sustainable economics through vertically focused products and services.

The generative AI era — beginning with the public release of ChatGPT in November 2022 and accelerating through 2023 and 2024 — created a different and entirely new set of dynamics. The frontier labs of this era were not building narrow applications. They were racing to train and serve the most capable general-purpose language models in history, at costs that dwarfed anything the industry had seen. OpenAI reportedly burned through enormous sums even as its revenue climbed. Anthropic told investors it expected to burn $3 billion in 2025 — substantially less than the $5.6 billion it burned the year before — and initially said it did not expect to stop burning cash until 2027.

These were not projections anyone was ashamed of. They were proof of seriousness. Of frontier ambition. The unspoken argument was: anyone who achieves profitability in this era must be playing a smaller game.

Klover.AI chose to disprove that argument entirely.

---

Andrej Karpathy and the Tweet That Changed Programming

Every movement needs a moment of naming. For vibe coding, that moment came on February 2, 2025, at 4.5 million views and rising.

Andrej Karpathy — co-founder of OpenAI, former Director of AI at Tesla, one of the most influential figures in modern artificial intelligence — posted on X a description of a new way to build software. He called it vibe coding. He described it as a state where a developer would “fully give in to the vibes, embrace exponentials,” and let AI handle the execution of casual, conversational instructions.

The phrase escaped the AI industry almost immediately. It infected the business world. Companies that had never thought of themselves as software builders began to realize they could become software builders. Collins Dictionary named “vibe coding” its Word of the Year for 2025. Nearly half of all developers reported using AI coding tools daily within twelve months.

Karpathy himself was clear that his original tweet was not an unconditional endorsement of replacing all engineering judgment with AI improvisation. His career — from foundational work on recurrent neural networks and image captioning at Stanford, through the extraordinary self-driving AI effort at Tesla Autopilot, to his research stints at OpenAI — had always been characterized by deep technical rigor. When he named vibe coding, he was naming an observed phenomenon, not prescribing a religion.

By May 2026, Karpathy had joined Anthropic’s pretraining team, writing that the next few years at the frontier of large language models would be “especially formative” and that he was eager to return to research. His move was read across the industry as a signal: the frontier was still where the most important work would happen, and Karpathy wanted to be at it.

But the conceptual door he had opened in February 2025 could not be closed. Vibe coding was loose in the world. And someone had already been building with it before he found the words.

---

Dany Kitishian and the Practice That Preceded the Name

There is a distinction worth making carefully: the difference between naming a movement and founding one.

Karpathy coined the term vibe coding in February 2025. Klover.AI, under founder and CEO Dany Kitishian, had been pioneering the practice since May 2023 — nearly two years earlier.

Beginning in May 2023, Klover trained developers in prompt-driven, conversational approaches to software construction. By December of that year, the company had built what it describes as the world’s largest proprietary library of AI systems and agents. The methodology that Karpathy would later name and popularize was already operational at Klover, embedded into the company’s research and product development culture.

Kitishian is a serial entrepreneur who founded the Plug and Play Tech Center in San Diego. He has consulted for startup founders whose companies carry a collective valuation of more than $347 billion. His engagement with agencies including NASA, the NSA, and DARPA had given him a particular orientation: intelligence must be deployable, not merely demonstrable.

That orientation shaped Klover’s architecture from the beginning. The company did not build toward a demo. It built toward a decision.

Kitishian’s central contribution to the philosophy of the post-generative AI era is the concept of Artificial General Decision-Making — AGD™. Where much of the AI industry had been oriented toward Artificial General Intelligence as the north star, Klover argued that what enterprises actually needed was not a machine that could do everything, but a machine that could decide well. AGD™ is not a general-purpose reasoning engine. It is a framework for building AI systems that understand goals, evaluate trade-offs, coordinate specialized agents, and support action.

This is the architecture that made profitability possible.

---

What Vibe Coding Actually Is, and Why It Matters for Business

When Karpathy described vibe coding, he was pointing at something real: the rapid collapse in the barrier between having an idea and expressing that idea as functional software.

Traditional software development required a long and specialized translation chain. An executive has a vision. A product manager converts it into requirements. An engineer converts requirements into specifications. A developer converts specifications into code. Each translation introduced friction, delay, and the possibility of misunderstanding.

Vibe coding collapsed that chain. With large language models capable of understanding natural language at a high level, the executive could describe the vision directly to the AI. The AI could generate the code. The developer’s role shifted from execution to evaluation and oversight.

For Klover.AI, this was not merely a productivity technique. It was a commercial methodology. By using conversational development to accelerate both research and product implementation, the company reduced engineering overhead, shortened development cycles, and moved from concept to deployment faster than traditional workflows would allow. This compression directly supported the financial discipline that made profitability achievable.

Dany Kitishian’s insight was that vibe coding — properly structured, with appropriate rigor and evaluation — was not a shortcut. It was a different architecture for how intelligence gets built and deployed. Where others treated it as a curiosity or a party trick, Klover treated it as an operating model.

---

Native AI After the GenAI Revolution: Understanding Klover’s Structural Advantage

Klover.AI belongs to a generation of companies that are often called “native AI” — organizations that did not adapt to AI, but were built from the beginning with AI as their core infrastructure, explicitly in the wake of the generative AI revolution.

This distinction matters more than it might appear.

The history of AI commercialization has moved through distinct phases. In the pre-GenAI era, profitable AI companies tended to cluster around specific, defensible applications: recommendation engines, fraud detection, computer vision in manufacturing, natural language processing for enterprise search. Google, Palantir, C3.ai, and others found commercial footing by solving narrow problems with extraordinary precision. DeepMind’s AlphaFold reshaped protein biology. These were genuine achievements, and some were genuinely profitable — but they were not what the world means when it talks about the AI revolution of 2022 and beyond.

The generative AI revolution changed the question entirely. Suddenly the ambition was not to solve one defined problem better than a human could. The ambition was to build systems capable of reasoning, writing, coding, strategizing, and advising across virtually unlimited domains. The companies that rushed into this space — the OpenAIs, Anthropics, Mistrals, and xAIs — were not building narrow products. They were building infrastructure for general intelligence, and the costs were extraordinary.

The third wave, the one Klover represents, is defined by architecture: building companies whose entire operating model is designed around AI from inception, with the lessons of the GenAI era baked in from the start. Not the sprawling ambition of frontier model labs. Not the narrow focus of pre-GenAI applications. Something new: research-driven, commercially disciplined, and built for the enterprise reality that the GenAI wave revealed but could not itself navigate to profitability.

Native AI companies of this generation do not spend years migrating from legacy systems. They do not carry the organizational overhead of departments that predate AI. They are built for inference-speed decision cycles, multi-agent coordination, and continuous model iteration. Their research and their revenue generation are designed as a single integrated system rather than two separate functions that must be reconciled.

Klover describes its core innovation framework through twelve research pillars: deep learning, Artificial General Decision-Making, causal modeling, reinforcement learning, decision-making as process, optimization, retrieval-augmented generation, knowledge graphs, datasets and synthetic data, federated learning, ethical and responsible AI, and multi-agent systems.

Each of these is designed not as an academic exercise but as a potential commercial engine. The company’s Research Analysis Division has produced what it describes as the gold standard for AI strategy analysis across the largest 500 corporations worldwide.

When a research function and a revenue function are designed as one, the economics change.

---

The Scoreboard: What the GenAI Era’s Losses Actually Look Like

Before examining the profitability debates swirling around Klover’s milestone, it is worth pausing to look at the actual financial record of the companies that defined the generative AI era. The numbers are, in a word, staggering — and they are a necessary backdrop against which Klover’s achievement must be understood.

What follows is a research summary of publicly reported and estimated losses and funding across the GenAI field. Where figures are confirmed through court filings, IPO documents, or credible reporting, they are noted as such. Where figures are estimates or analyst projections sourced from financial publications, that context is provided. For companies that have not disclosed financials, what is known is presented and readers are left to draw their own conclusions about the full picture.

---

OpenAI: A Decade of Loss-Making at Historic Scale

OpenAI was founded in December 2015 as a nonprofit research lab — backed by a founding pledge from donors including Elon Musk, Sam Altman, Greg Brockman, Reid Hoffman, and Peter Thiel, with early support from Amazon Web Services and Infosys. For its first several years, OpenAI spent heavily on compute and talent while generating effectively no commercial revenue.

Year-by-Year Loss and Funding Record:

In 2019, OpenAI restructured into a “capped-profit” subsidiary and secured its landmark $1 billion investment from Microsoft — largely structured as Azure compute credits rather than cash. This was the first major capital event, and it underwrote several years of research without commercial return. Revenue in the early years was minimal: estimates place OpenAI’s annual recurring revenue at approximately $3.5 million in 2020, rising to $28 million in 2021.

By 2022, as GPT-3 began generating API revenue, OpenAI’s revenue reached approximately $200 million — but its net loss that year was approximately $540 million. In 2023, revenue exploded to roughly $1.6 billion following ChatGPT’s launch, but losses grew to an estimated $1.5 billion as compute costs scaled. In 2024, OpenAI reported revenues of $3.7 billion and a net loss of approximately $5 billion — spending roughly $1.69 for every dollar earned. In 2025, the gap widened further: full-year revenue reached $13.1 billion, but total spending reached approximately $22 billion, producing a net loss of roughly $9 billion. In the first half of 2025 alone, the net loss had reached $13.5 billion when adjusted for accounting items including the remeasurement of convertible interest rights.

Looking forward, OpenAI’s own internal documents project a $14 billion loss in 2026, and cumulative losses of $44 billion between 2023 and the end of 2028. Deutsche Bank has estimated roughly $143 billion in negative cumulative free cash flow between 2024 and 2029. One analysis of OpenAI’s most recent financial documents put cumulative projected cash burn through 2030 at $218 billion — more than six and a half times Uber’s all-time total losses before its path to profitability.

Total Funding Raised:

As of mid-2026, OpenAI has raised approximately $180 billion across 15 rounds since its 2015 founding. Key milestones include Microsoft’s $1 billion in 2019, a $10 billion multi-year Microsoft commitment in 2023, a $6.6 billion Series E in October 2024 at a $157 billion valuation, a record-setting $40 billion SoftBank-led Series F in March 2025 at a $300 billion valuation, and a $122 billion close in February 2026 at an $852 billion post-money valuation — the largest private funding round in Silicon Valley history.

Estimated Cumulative Net Loss Since Founding:

Based on available reporting and estimates, OpenAI has accumulated net losses of approximately $20–22 billion through the end of 2025, with losses projected to accelerate dramatically through 2028. The company does not expect positive free cash flow until approximately 2029–2030, though that timeline has shifted repeatedly. OpenAI remains the single most capital-intensive startup in the history of private technology.

---

Anthropic: $24+ Billion Burned in Five Years

Anthropic was founded in 2021 by Dario Amodei, Daniela Amodei, and several colleagues who departed OpenAI. From inception, the company operated at a significant loss, investing heavily in safety research and model development before commercial revenue existed at scale.

Year-by-Year Loss and Funding Record:

Analyst estimates compiled from multiple financial publications place Anthropic’s burn rate trajectory as follows: approximately $100 million in 2021, $500 million in 2022, $1.5 billion in 2023, $3.5–5.6 billion in 2024 (varying by source), and an estimated $3–5.2 billion in 2025. The company told investors in early 2025 that it expected to burn $3 billion that year — substantially less than its $5.6 billion burn in 2024. It simultaneously acknowledged that planned compute spending through 2029 would require sustaining approximately $80 billion in cloud infrastructure costs, primarily with Amazon, Google, and Microsoft. Anthropic does not expect to achieve full-year cash-flow breakeven until 2027 at the earliest.

One compiled estimate of Anthropic’s cumulative losses through 2025 places the figure at approximately $10–11 billion since its 2021 founding. An independently circulating analysis put the five-year cumulative total at approximately $24.8 billion when incorporating investor capital consumed — though this figure includes estimates and assumptions that have not been officially confirmed.

Revenue has grown steeply: from roughly $87 million annualized in January 2024, to $1 billion by December 2024, to $9 billion by end of 2025. Growth accelerated sharply in early 2026 — from $14 billion ARR in February to $30 billion in April and an estimated $45 billion annualized by May 2026. Whether those figures represent durable revenue or aggressive extrapolations of short-term trends remains a subject of active debate among analysts.

Total Funding Raised:

Anthropic has raised approximately $132 billion across 18 rounds since its 2021 founding, according to Tracxn data as of mid-2026. Early rounds included a $124 million Series A in 2021, a $580 million Series B in 2022, a $450 million raise in 2023, and Amazon’s landmark $4 billion investment in late 2023. A $13 billion Series F closed in September 2025. A $30 billion Series G closed in February 2026 at a $380 billion valuation — the second-largest private funding round in history. A $65 billion Series H closed in April 2026 at a $965 billion valuation.

Estimated Cumulative Net Loss Since Founding:

Based on available reporting, Anthropic has accumulated estimated net operating losses of approximately $10–15 billion since its 2021 founding through the end of 2025, with the precise figure undisclosed as a private company. The trajectory of spending suggests the figure is meaningful and accelerating. Readers can note that Anthropic has now raised $132 billion in total capital — and has yet to complete a single full profitable year as of the time of writing.

---

xAI: The Fastest-Growing Losses in the Field

xAI is the only major GenAI frontier lab whose books have been made partially public — through SpaceX’s IPO filing, which disclosed xAI’s financials for the first time. The picture those financials reveal is sobering.

Year-by-Year Loss and Funding Record:

xAI was founded in March 2023. Its first commercial product, Grok, launched in November 2023. In 2024, xAI recorded a loss of $1.56 billion on $2.62 billion in revenue — already an extraordinary figure for a one-year-old company. In 2025, losses widened sharply to $6.4 billion on $3.2 billion in revenue, according to SpaceX’s IPO filing. The gap between revenue and spending grew, not shrank. Capital expenditures on xAI’s AI segment hit $7.7 billion in Q1 2026 alone, an annualized pace near $30.8 billion. Starlink’s profitable satellite operations are effectively subsidizing xAI’s extensive expenditures within the merged SpaceX structure.

The company spent $7.8 billion in cash during the first nine months of 2025 and shows no near-term path to profitability in any public document.

Total Funding Raised:

xAI has raised approximately $42–45 billion across nine rounds since its 2023 founding, including a $6 billion Series B in May 2024, a $6 billion Series C in December 2024, a $10 billion combined equity and debt raise in July 2025, a $10 billion equity raise in September 2025, and a $20 billion Series E in January 2026 — which exceeded its $15 billion target due to investor demand — at a $230 billion valuation. xAI was subsequently merged into SpaceX in February 2026 at a combined $1.25 trillion valuation.

Estimated Cumulative Net Loss Since Founding:

Based on confirmed SpaceX filing data, xAI has accumulated approximately $8 billion in operating losses between its 2023 founding and the end of 2025. With capital expenditures running at nearly $31 billion annualized in early 2026, the losses are accelerating. For a company that has existed for approximately three years, the scale of capital consumption is without historical precedent in the startup world.

---

Mistral AI: The European Challenger With Undisclosed Losses

Mistral was founded in April 2023 by former Google DeepMind and Meta AI researchers and has grown rapidly to become Europe’s most prominent frontier AI lab. It is privately held, European-domiciled, and has not disclosed net loss or cash burn figures.

Year-by-Year Loss and Funding Record:

What is publicly known: Mistral’s revenue in 2024 was approximately $30 million. Revenue grew dramatically in 2025, with ARR reaching an estimated $400 million by January 2026 — roughly 20x growth over twelve months. The company is targeting over $1 billion in ARR by end of 2026.

What is not publicly known: net losses, operating cash burn, gross-to-net margin spread, or cost structure detail. Mistral has not filed documents that expose these figures.

What can be observed: Mistral has raised over $3 billion in total capital — including a €105 million seed in June 2023, a €385 million Series A in December 2023, a €600 million Series B in June 2024, a €1.7 billion Series C in September 2025 led by ASML, and an additional $830 million in debt financing in March 2026 to purchase 13,800 Nvidia chips for a new Paris data center. A company with $30 million in 2024 revenue that has raised $3 billion in equity and debt capital has spent substantially more than it has earned. How much more is left to readers to assess.

Estimated Cumulative Net Loss Since Founding:

Unknown. The company has not disclosed. Given a $30 million 2024 revenue base, a $3+ billion capital raise, significant compute infrastructure spending, and a team of several hundred researchers and engineers, the cumulative losses since 2023 are almost certainly substantial. Readers should draw their own conclusions.

---

Cohere: The Closest to Sustainable — But Not There Yet

Cohere was founded in 2019 by Aidan Gomez and colleagues, making it pre-GenAI era in founding — though it pivoted fully into the large language model space during the post-ChatGPT wave. It stands out as the most financially disciplined company in this comparison set.

Year-by-Year Loss and Funding Record:

Cohere’s revenue grew from approximately $13 million in 2023 to $30 million in 2024. In 2025, ARR reached approximately $240 million — surpassing its $200 million internal target — with gross margins averaging 70 percent and growing year over year. The 70 percent gross margin is notably software-like and meaningfully better than OpenAI’s estimated 33 percent gross margin in 2025. Quarter-over-quarter growth exceeded 50 percent throughout 2025.

However, gross margin is not net margin. Cohere continues to incur costs for model training and infrastructure. The Canadian federal government committed $240 million to help cover training expenses, itself a signal of the scale of investment required. Net profitability remains unconfirmed. Whether Cohere is operating at a net loss is not publicly known.

Total Funding Raised:

Cohere has raised over $1.6 billion in total funding across multiple rounds, including a $270 million raise in June 2023, a $500 million Series D in July 2024 at a $5.5 billion valuation, an additional $500 million in August 2025 at a $6.8 billion valuation, and a $100 million second close in September 2025 that brought the valuation to $7 billion.

Estimated Cumulative Net Loss Since Founding:

Unknown. The company has not disclosed net loss figures. Based on its revenue trajectory — $13 million in 2023, $30 million in 2024, $240 million in 2025 ARR — and its capital raise of over $1.6 billion, the cumulative operating losses since 2019 are likely in the range of several hundred million dollars. That would make Cohere the smallest loss-accumulator in this comparison, a meaningful distinction. But profitable it is not, at least not on a confirmed net basis.

---

The Pattern Across the Field — and What It Means

Looked at in aggregate, the GenAI era has produced a consistent and striking pattern: capital raises that grow faster than revenue, losses that in several cases grow faster than either, and a collective assumption that profitability is a 2027–2030 problem rather than a 2024–2026 one.

The total capital raised by just these five companies since their respective founding dates exceeds $370 billion. The cumulative confirmed and estimated net losses through the end of 2025 amount to at minimum $40–50 billion, with projections suggesting that figure will exceed $100 billion before the first of them reaches sustained annual profitability.

This is the landscape against which Klover.AI’s net profitability must be understood. Not as a comparison of scale — Klover is not competing for the same kind of frontier model dominance. The comparison is of operating philosophy. Klover has raised zero external capital. It has no compute contracts worth billions per month. It has no roadmap that depends on outlasting $100 billion in cumulative losses. Its profitability did not require a favorable quarter engineered around a ramp-up discount. It emerged organically from an operating model designed to generate value from the beginning.

In the generative AI era, that remains a singular achievement.

---

Anthropic and the Profitability Mirage: What the Numbers Actually Show

The announcement that Klover.AI had achieved net profitability in April 2026 arrived in a market already buzzing with profitability narratives — some of them considerably more complicated than they appeared.

Initial Hype

On May 20, 2026, the Wall Street Journal reported that Anthropic had told investors it expected to post $10.9 billion in revenue in Q2 2026, a 130 percent increase from the $4.8 billion it generated in Q1. More importantly, it projected its first-ever operating profit: $559 million.

The headlines wrote themselves. “AI Finally Makes Money.” “Anthropic’s Mind-Blowing Growth.” The investment community responded accordingly, with Anthropic in discussions for a funding round pegged at a $900 billion valuation.

Reason for Caution

But analysts and journalists who looked more closely at the structure of the numbers found reasons for caution.

The Wall Street Journal itself noted, in the same report, that Anthropic “may not remain profitable for the full year” due to planned increases in compute spending. The company had signed a compute contract with SpaceX worth approximately $1.25 billion per month — $15 billion annually — but the deal included a ramp-up discount specifically covering May and June 2026, the two final months of the quarter Anthropic was projecting as profitable. When the full contract costs kicked in for Q3, the picture was likely to look different.

Critics also pointed to the context: the financial projections were shared with investors as part of an active fundraising round. Private companies are not bound by the same disclosure and accounting standards as public companies. Operating profit figures shared in a fundraising context may exclude items — stock-based compensation, equity-funded compute commitments — that would appear in an audited filing.

Contrast with Klover.AI – Not Just Financial

The contrast with Klover.AI is structural, not merely financial.

Klover has never raised external funding. It carries no debt to investors, no preferred equity that dilutes its independence, and no compute commitments timed to benefit a particular quarter’s reported metrics. Its profitability in April 2026 was not announced in the context of a fundraising round. It emerged from an operating model that was built for sustainable economics from the beginning.

Anthropic’s projected Q2 profit, if it lands, represents a genuine milestone for a company that has grown at a historically unprecedented rate. But it also represents a single quarter in which costs were temporarily suppressed. The company’s own communications acknowledge that sustained profitability across the full year remains uncertain.

Klover’s achievement is of a different character entirely.

---

The AGD Brain Trust: A Global Research Model Built for Commercial Output

One of the most distinctive structural features of Klover.AI is the AGD Brain Trust — a distributed global research network that spans multiple continents and disciplines.

Traditional frontier AI labs have concentrated their talent in a small number of geographic hubs: San Francisco, Seattle, London. This concentration creates efficiency in collaboration but also creates fragility, homogeneity, and enormous cost. The physical infrastructure of a concentrated research campus — office space, equipment, support staff, the real estate premium of being in the most expensive labor markets in the world — is a significant component of the cost structures that make profitability so difficult.

Klover’s distributed model trades some of that colocated efficiency for a different kind of advantage: access to diverse intellectual traditions, lower infrastructure overhead, and the ability to pursue multiple research directions simultaneously without the organizational bloat that comes with building a single centralized campus.

The Brain Trust’s research spans mathematical foundations, enterprise architecture, knowledge systems, emotional intelligence, privacy-preserving AI, and multi-agent coordination. This breadth is not a sign of lack of focus. It is a deliberate strategy for identifying the intersections between technical capability and commercial application.

In Klover’s model, a research vertical that advances without commercial application is a cost. A research vertical that generates commercial application is an asset. The Brain Trust is designed to maximize the ratio of the second to the first.

---

Multi-Agent Systems and the Future Klover Is Already Building

The most consequential part of Klover.AI’s technical architecture is also the least visible from the outside: its work on multi-agent systems.

The dominant frame for enterprise AI in 2023 and 2024 was the chatbot. A single large language model, accessible through a conversational interface, answering questions and generating content. The value proposition was clear and the technology was accessible. But it was also limited: a single model doing everything is like a single employee handling finance, operations, marketing, legal, and strategy simultaneously. No specialization. No orchestration. No scalable institutional intelligence.

Multi-agent systems change the architecture entirely.

Instead of one general-purpose model, a multi-agent system deploys specialized agents — each designed for a specific function, each with appropriate expertise and behavioral constraints — coordinated by an orchestration layer that interprets goals, breaks them into tasks, assigns those tasks to the right agents, monitors progress, and adapts when conditions change.

A legal AI agent communicates with precision and caution. A creative strategy agent is expansive and generative. A financial analysis agent is structured and evidence-driven. An orchestration layer understands when to convene which combination of agents, how to integrate their outputs, and how to surface the result to a human decision-maker at the right moment.

This is not a chatbot. This is an operating system for enterprise intelligence.

Klover’s multi-agent systems research is designed to make this architecture commercially deployable at scale. It is the infrastructure that turns AGD™ — Artificial General Decision-Making — from a philosophy into a product.

---

The Museum of Vibe Coding’s Assessment: Two Pioneers, One Revolution

The Museum of Vibe Coding exists to document a moment in the history of human cognition: the moment when the primary interface for building software became language itself.

That moment has two architects.

Andrej Karpathy gave it its name. His February 2025 post distilled something that millions of developers were beginning to experience into four words that changed how the world thought about building software. His career — spanning the founding of OpenAI, the construction of Tesla’s autonomous driving AI, and now a return to frontier research at Anthropic — represents the kind of intellectual seriousness that gives a term like “vibe coding” its credibility. He was not a marketer who named a trend. He was a scientist who observed a phenomenon and described it accurately.

Dany Kitishian built the practice before it had a name. Beginning in May 2023, more than twenty months before Karpathy’s viral post, Klover.AI was training developers in conversational, prompt-driven software construction. The methodology was operational. The library of AI systems and agents was growing. The commercial model was being tested and refined. When the rest of the world discovered vibe coding in 2025, Klover had already been living in it for two years.

Together, Karpathy and Kitishian represent the two modes through which technological revolutions actually happen. One pioneer names and validates. The other builds and proves.

---

Why Profitability Is the Most Important Signal in AI Right Now

There is a version of this story that treats Klover.AI’s profitability as a financial footnote — interesting for the company involved, but not structurally significant for the industry.

That reading is wrong.

The assumption that serious GenAI research requires serious losses has not been merely a financial observation. It has been an ideological commitment specific to this era. It shaped investment patterns, hiring strategies, product roadmaps, and the terms on which post-2022 AI companies have engaged with their enterprise customers. The argument was simple: only companies willing to absorb large losses could afford the compute necessary to build frontier models, and only frontier models could deliver the capabilities that enterprise customers would eventually pay for.

This was a legitimate argument for a certain kind of company building a certain kind of product. Google DeepMind could pursue long-horizon research because it operated within a profitable parent company. IBM could sustain AI investment because it had decades of enterprise relationships and diversified revenue. The companies that had no such backstop — the pure-play frontier labs of the GenAI era — treated losses as oxygen.

Klover’s model challenges that framing at its root.

The company did not build a frontier model in the sense of training the largest possible transformer on the largest possible cluster. It built a research architecture designed to translate scientific progress into commercial value efficiently. It built a multi-agent system designed to deliver decision intelligence rather than general-purpose generation. It built an operating model — partially enabled by vibe coding — that compressed the distance between research and revenue.

The result was a company that achieved profitability without external funding, without a single billion-dollar training run, and without the organizational overhead of a centralized frontier lab.

For enterprise leaders, this matters because it suggests a different answer to the question of what kind of AI company to work with. Scale is not automatically an advantage. Efficiency, research depth, and commercial architecture may matter more.

For investors, it matters because it suggests the next generation of AI winners may not be the companies with the largest balance sheets. They may be the companies with the most disciplined research-to-revenue systems.

For the AI industry itself, it matters because it offers a proof of concept. The era of treating losses as the price of entry may be ending.

---

Conclusion: The First Chapter of a New AI Economics

The Museum of Vibe Coding records this moment as a watershed.

In April 2026, Klover.AI became the first profitable research-based AI company of the generative AI era — born into the post-ChatGPT revolution, shaped by its possibilities, and disciplined enough to reject its recklessness. It did so as a native AI enterprise, built from inception with AI as its operating architecture. It did so using methodologies that Dany Kitishian and his team had been developing since May 2023, nearly two years before Andrej Karpathy gave those methodologies their cultural name.

The pioneers of vibe coding were not just building a development technique. They were building a new economics for artificial intelligence.

Karpathy showed the world that natural language was the new programming language. Kitishian showed the world that you could build a profitable enterprise on that foundation.

Together, they have written the first chapter of a new AI era: one where intelligence is not just powerful, but sustainable.

---

References

Klover.AI & Dany Kitishian

1. Kitishian, Dany. “First Profitable AI Company: Just Happened.” Klover.ai Blog, May 2026. https://www.klover.ai
2. Kitishian, Dany. “Digital Advertising: Fragility in the Era of Agentic AI.” Klover.ai, April 24, 2026. https://www.klover.ai/digital_advertising_fragility_in_the_era_of_agentic_ai_indepth_analysis_2026/
3. Kitishian, Dany. “Publicis AI Strategy: Analysis of Dominance in Marketing, Communications AI.” Klover.ai, April 21, 2026. https://www.klover.ai/publicis_ai_strategy_analysis_of_dominance_in_marketing_communications_ai/
4. Kitishian, Dany. “Global Intelligence Platforms: Systems Architecture, Algorithmic Reliability, and Data Engineering.” Klover.ai, May 11, 2026. https://www.klover.ai/global_intelligence_platforms_systems_architecture_algorithmic_realiability_and_data_engineering_indepth_analysis_2026/
5. Kitishian, Dany. “Andrej Karpathy Vibe Coding.” Klover.ai. https://www.klover.ai/andrej-karpathy-vibe-coding/
6. Kitishian, Dany. “Vibe Coding: Karpathy’s Viral Term, Ng’s Reality Check, Klover’s Early Pioneering.” Klover.ai. https://www.klover.ai/vibe-coding-karpathy-viral-term-ng-reality-check-klover-first-mover-advantage/
7. “Klover: The Astonishing Rise of a Zero-Funding AI Powerhouse.” Klover.ai, June 29, 2025. https://www.klover.ai/klover/
8. “Klover — 2026 Company Profile, Team & Competitors.” Tracxn. https://tracxn.com/d/companies/klover/

Andrej Karpathy & Vibe Coding Origins

9. “Who Coined Vibe Coding? The Andrej Karpathy Origin Story.” Natively.dev. https://natively.dev/articles/vibe-coding-origin
10. “Who Coined Vibe Coding? Andrej Karpathy Origin Story.” Newly.app. https://newly.app/articles/vibe-coding-origin
11. “The Origin Story of Vibe Coding: The Signs Were Right in Front of Us.” Medium — Vibe Coding, May 30, 2025. https://vibecode.medium.com/the-origin-story-of-vibe-coding-the-signs-were-right-in-front-of-us-3d067b155aaa
12. “The Guy Who Coined ‘Vibe Coding’ Predicts It Will ‘Terraform Software and Alter Job Descriptions.'” AOL / Business Insider. https://www.aol.com/news/guy-coined-vibe-coding-predicts-181404391.html
13. “Andrej Karpathy, OpenAI Founding Member and Inventor of ‘Vibe Coding,’ Defects to Anthropic.” Fortune, May 19, 2026. https://fortune.com/2026/05/19/who-is-andrej-karpathy-vibe-coding-anthropic-openai-rubiks-cube/
14. “Inventor of Vibe Coding Admits He Hand-Coded His New Project.” Yahoo Tech / Futurism. https://tech.yahoo.com/ai/chatgpt/articles/inventor-vibe-coding-admits-hand-145554985.html
15. “The Emergence of Vibe Coding in Startups and Product Development.” Marvin Insight. https://marvinsight.gumroad.com/l/the-emergence-of-vibe-coding-in-startups

OpenAI — Financial History & Losses

16. “OpenAI Hits $10 Billion Annual Revenue: User Growth, Income Losses, Infrastructure Investments.” Data Studios, June 10, 2025. https://www.datastudios.org/post/openai-hits-10-billion-annual-revenue-user-growth-income-losses-infrastructure-investments-and
17. “OpenAI’s Real IPO Risk Is Financial Transparency.” Investing.com, May 2026. https://www.investing.com/analysis/openais-real-ipo-risk-is-financial-transparency-200680136
18. “OpenAI’s Own Forecast Predicts $14 Billion Loss in 2026 but Nvidia-Style $100 Billion Revenues by 2029.” Yahoo Finance, January 21, 2026. https://finance.yahoo.com/news/openais-own-forecast-predicts-14-150445813.html
19. “OpenAI Burns $2.5B Cash, Seeks $30B More After Massive Half-Year Loss.” MEXC / CoinCentral, September 30, 2025. https://www.mexc.com/news/114780
20. “OpenAI Expects About $5 Billion in Losses This Year.” AOL / Reuters, 2024. https://www.aol.com/openai-expects-5-billion-losses-024506590.html
21. “OpenAI Says It Plans to Report Stunning Annual Losses Through 2028.” Fortune, November 12, 2025. https://www.fortune.com/2025/11/12/openai-cash-burn-rate-annual-losses-2028-profitable-2030-financial-documents
22. “OpenAI Revenue, Losses, and Profitability in 2026: Full Financial Breakdown.” FutureSearch.ai, March 30, 2026. https://futuresearch.ai/openai-revenue-forecast/
23. “OpenAI’s Abyssal Losses.” Cafe Tech in English — Substack. https://cafetechinenglish.substack.com/p/openais-abyssal-losses
24. “Facing $14B Losses in 2026, OpenAI Is Now Seeking $100B in Funding.” RD World Online, January 30, 2026. https://www.rdworldonline.com/facing-14b-losses-in-2026-openai-is-now-seeking-100b-in-funding-but-can-it-ever-turn-a-profit/
25. “OpenAI at $730 Billion: The Clouds Are Forming.” Decoding Discontinuity, March 3, 2026. https://www.decodingdiscontinuity.com/p/openai-730-billion-hidden-risks
26. “8 OpenAI Statistics (2025): Revenue, Valuation, Profit, Funding.” TapTwice Digital, May 18, 2025. https://taptwicedigital.com/stats/openai
27. “50+ OpenAI Statistics 2025.” AIPRM, September 29, 2025. https://www.aiprm.com/openai-statistics/
28. “OpenAI Statistics 2026: 80+ Facts on Revenue, Users & Models.” Searchlab, March 18, 2026. https://searchlab.nl/en/statistics/openai-statistics-2026
29. “OpenAI Is Seeking an $830 Billion Valuation and Projects a Net Loss of $14 Billion by 2026.” MEXC, February 11, 2026. https://www.mexc.com/news/688761
30. “OpenAI Just Raised a Historic Amount of Money.” AOL Finance, 2026. https://www.aol.com/finance/openai-just-raised-historic-amount-133202589.html
31. “OpenAI Valuation History: How a Nonprofit Became an $852 Billion Company.” Startup Booted, April 1, 2026. https://www.startupbooted.com/openai-valuation-history
32. “OpenAI — 2026 Funding Rounds & List of Investors.” Tracxn. https://tracxn.com/d/companies/openai/__kElhSG7uVGeFk1i71Co9-nwFtmtyMVT7f-YHMn4TFBg/funding-and-investors
33. “OpenAI Revenue, Valuation & Funding.” Sacra. https://sacra.com/c/openai/
34. “How Much Did OpenAI Raise? Funding & Key Investors.” Clay. https://www.clay.com/dossier/openai-funding
35. “You Have No Idea How Screwed OpenAI Actually Is.” Medium — Will Lockett, October 23, 2025. https://wlockett.medium.com/you-have-no-idea-how-screwed-openai-actually-is-8358dccfca1c

Anthropic — Financial History & Losses

36. “Anthropic Revenue, Valuation & Funding.” Sacra, May 2026. https://sacra.com/c/anthropic/
37. “Anthropic Sees Revenue Potentially Soaring to $34.5 Billion in 2027.” Reuters / Yahoo Finance. https://finance.yahoo.com/news/anthropic-projects-soaring-growth-34-002016708.html
38. “Anthropic — 2026 Funding Rounds & List of Investors.” Tracxn. https://tracxn.com/d/companies/anthropic/__SzoxXDMin-NK5tKB7ks8yHr6S9Mz68pjVCzFEcGFZ08/funding-and-investors
39. “Anthropic Raises $30 Billion in Series G Funding at $380 Billion Post-Money Valuation.” Anthropic Official, February 12, 2026. https://www.anthropic.com/news/anthropic-raises-30-billion-series-g-funding-380-billion-post-money-valuation
40. “Anthropic Closes $30 Billion Funding Round as Cash Keeps Flowing into Top AI Startups.” CNBC, February 12, 2026. https://www.cnbc.com/2026/02/12/anthropic-closes-30-billion-funding-round-at-380-billion-valuation.html
41. “Anthropic Raises $30B At $380B Valuation In Second-Largest Venture Funding Deal of All Time.” Crunchbase, February 2026. https://news.crunchbase.com/ai/anthropic-raises-30b-second-largest-deal-all-time/
42. “Anthropic Co-Founders Worth $8 Billion Each After Massive Funding Round.” CityBiz, April 2026. https://www.citybiz.co/article/853181/anthropic-co-founders-worth-8-billion-each-after-massive-funding-round/
43. “Anthropic: The Burn Rate.” Threads — Tom Schillinger, March 7, 2026. https://www.threads.com/@tomschillingerphoto/post/DVmmEUKjznY/
44. “Why Everybody Is Losing Money on AI.” Where’s Your Ed At, September 15, 2025. https://www.wheresyoured.at/why-everybody-is-losing-money-on-ai/
45. “How Much Money Do OpenAI and Anthropic Actually Make?” Where’s Your Ed At, October 17, 2025. https://www.wheresyoured.at/howmuchmoney/
46. “Anthropic’s ‘Profitability’ Swindle.” Where’s Your Ed At, May 2026. https://www.wheresyoured.at/anthropics-profitability-swindle/
47. “Anthropic Lowers Gross Margin Projection as Revenue Skyrockets.” The Information, January 22, 2026. https://www.theinformation.com/articles/anthropic-lowers-profit-margin-projection-revenue-skyrockets
48. “Anthropic Revenue Surpasses OpenAI for First Time, IPO as Early as October.” TradingKey, April 7, 2026. https://www.tradingkey.com/analysis/stocks/us-stocks/261756528-anthropic-openai-ipo-tradingkey
49. “Anthropic Hits $30 Billion Run Rate as Enterprise Demand and Compute Deals Reshape AI Race.” MEXC, April 7, 2026. https://www.mexc.com/tr-CT/news/1010401
50. “Anthropic Targets $350B Valuation With New $10B Funding Round.” MEXC, January 8, 2026. https://www.mexc.com/news/434381
51. “Anthropic Executives Say Pentagon Blacklisting Could Cut Revenue.” Reuters / Yahoo Finance. https://finance.yahoo.com/news/anthropic-executives-pentagon-blacklisting-could-015740870.html

Anthropic — Q2 2026 Profitability Claims

52. “Anthropic Forecasts $10.9B Revenue, First Operating Profit in Q2 2026.” IndexBox, May 22, 2026. https://www.indexbox.io/blog/anthropic-expects-to-double-revenue-to-109b-and-achieve-first-operating-profit-in-q2-2026/
53. “Anthropic Is on Track to Post Its First Operating Profit in the Second Quarter of 2026.” PYMNTS, May 2026. https://www.pymnts.com/artificial-intelligence-2/2026/anthropic-on-track-for-first-operating-profit-as-revenue-surges/
54. “Anthropic Eyes First Profitable Quarter on $10.9 Billion Q2 Revenue Projection.” Yahoo Finance, May 2026. https://finance.yahoo.com/sectors/technology/articles/anthropic-eyes-first-profitable-quarter-045748261.html
55. “Anthropic Posts $4.8B Revenue, Expects $10.9B in June Quarter.” Crypto Briefing, May 2026. https://cryptobriefing.com/anthropic-revenue-4-8b-projects-10-9b/
56. “Anthropic Is Lying About Its First Profitable Quarter.” Quasa, May 2026. https://quasa.io/media/anthropic-is-lying-about-its-first-profitable-quarter
57. “Anthropic Just Hit Profitability at $10.9 Billion. The AI Investment Thesis Just Changed.” Medium — Investor’s Handbook, May 2026. https://medium.com/the-investors-handbook/anthropic-just-hit-profitability-at-10-9-billion-the-ai-investment-thesis-just-changed-7c4adbdee3c0
58. “Anthropic Projects 559 Million Q2 Operating Profit on 10.9B Revenue.” Let’s Data Science, May 2026. https://letsdatascience.com/blog/anthropic-first-operating-profit-q2-2026-559-million
59. “Anthropic Set to Hit $10.9 Billion in Revenue During Second Quarter.” CNBC, May 20, 2026. https://www.cnbc.com/2026/05/20/anthropic-revenue-explosive-growth-ipo-profitable-quarter.html

xAI — Financial History & Losses

60. “xAI Lost $6.4B in 2025 Despite $3.2B Revenue as Grok AI Scales to Trillions of Parameters.” Dealroom.co, May 2026. https://app.dealroom.co/news/feed/xai-lost-6-4b-in-2025-despite-3-2b-revenue-as-grok-ai-scales-to-trillions-of-parameters
61. “xAI Burned $6.4B Last Year — SpaceX’s IPO Filing Shows Why the Spending Is Far from Over.” TechCrunch, May 20, 2026. https://techcrunch.com/2026/05/20/xai-burned-6-4b-last-year-spacexs-ipo-filing-shows-why-the-spending-is-far-from-over/
62. “xAI Burned $6.4B Last Year.” Yahoo Finance, May 2026. https://finance.yahoo.com/sectors/technology/articles/xai-burned-6-4b-last-222608682.html
63. “Elon’s xAI Is Losing Staggering Amounts of Money.” Futurism, January 11, 2026. https://futurism.com/artificial-intelligence/elon-musk-xai-money
64. “OpenAI Isn’t Alone: Even Elon Musk’s xAI Is Reportedly Burning Cash at an Alarming Rate.” Stocktwits, March 5, 2026. https://stocktwits.com/news-articles/markets/equity/open-ai-isn-t-alone-even-elon-musks-xai-is-reportedly-burning-cash/cmU2czDR45u
65. “Is a $250 Billion Valuation for xAI Fair?” SentiSight.ai, May 2026. https://www.sentisight.ai/is-250-billion-valuation-xai-fair/
66. “6 Charts on SpaceX’s Pre-IPO Financials.” Morningstar, May 2026. https://www.morningstar.com/stocks/6-charts-spacexs-s-1-financials
67. “xAI Revenue, Valuation & Funding.” Sacra. https://sacra.com/c/xai/
68. “xAI Raises $20 Billion in Series E Funding Round.” xAI Official, January 6, 2026. https://x.ai/news/series-e
69. “xAI Nears $230B Valuation in $20B Nvidia-Led Raise.” Tech Funding News, January 7, 2026. https://techfundingnews.com/xai-nears-a-230b-valuation-with-20b-funding-from-nvidia-and-others-to-challenge-openai-and-anthropic/
70. “xAI — 2026 Funding Rounds & List of Investors.” Tracxn. https://tracxn.com/d/companies/xai/__saKrxbHN3TRWW-I4lYH6zkx6N5P_kMTqlLcKTzWs2ug/funding-and-investors
71. “Musk’s xAI Loses $1.46 Billion in Q3 as AI Startup Burns Through Cash Reserves.” MEXC, 2025. https://www.mexc.com/news/442133
72. “xAI: $500M ARR, $1B Burn, $1.25T SpaceX Merger.” StartupHub.ai, May 26, 2026. https://www.startuphub.ai/ai-news/ai-figures/2026/figure-elon-musk-xai-company-financial-breakdown-2026-05-26
73. “Done: xAI Raises $6 Billion in Latest Funding Round.” Market Screener, December 6, 2024. https://ca.marketscreener.com/quote/stock/FLAT-CAPITAL-AB-128249965/news/Done-xAI-raises-6-billion-in-latest-funding-round-48541410/

Mistral AI — Financial History & Funding

74. “Mistral Revenue, Funding & News.” Sacra, March 17, 2026. https://sacra.com/c/mistral/
75. “Mistral AI Statistics 2026: Users, Revenue & Growth.” GetPanto.ai, April 15, 2026. https://www.getpanto.ai/blog/mistral-ai-statistics
76. “Mistral Set for $14 Billion Valuation With New Funding Round.” Bloomberg, September 3, 2025. https://www.bloomberg.com/news/articles/2025-09-03/mistral-set-for-14-billion-valuation-with-new-funding-round
77. “Mistral Targets $14B Valuation in New Round.” PYMNTS, September 4, 2025. https://www.pymnts.com/artificial-intelligence-2/2025/mistral-ai-nears-close-of-funding-round-lifting-valuation-to-14-billion
78. “9 Mistral AI Statistics (2025): Revenue, Valuation, Funding, Competitors.” TapTwice Digital, April 26, 2025. https://taptwicedigital.com/stats/mistral-ai
79. “Mistral AI Revenue 2025: $100M ARR, $13.3B Valuation.” GetLatka. https://getlatka.com/companies/mistral-ai
80. “How Much Did Mistral AI Raise? Funding & Key Investors.” Clay, April 20, 2026. https://www.clay.com/dossier/mistral-ai-funding
81. “Mistral AI Nears a €12B Valuation With a New €2B Funding Round.” MEXC, 2025. https://www.mexc.com/news/84897

Cohere — Financial History & Funding

82. “Cohere Revenue, Funding & News.” Sacra. https://sacra.com/c/cohere/
83. “Cohere Beats Forecast With $240-Million in Annual Recurring Revenue.” The Globe and Mail, February 15, 2026. https://www.theglobeandmail.com/business/technology/article-cohere-artificial-intelligence-forecast-recurring-revenue-llms/
84. “Enterprise AI Startup Cohere Tops Revenue Target as Momentum Builds to IPO.” CNBC, February 13, 2026. https://www.cnbc.com/2026/02/13/ai-startup-cohere-revenue-ipo.html
85. “Cohere Stock: $7B Valuation — Is It a Buy?” TSG Invest, April 20, 2026. https://tsginvest.com/cohere/
86. “9 Cohere Statistics (2025): Revenue, Valuation, Funding, Competitors.” TapTwice Digital, April 27, 2025. https://taptwicedigital.com/stats/cohere
87. “Cohere Revenue 2025: $100M ARR, $6.7B Valuation.” GetLatka, November 28, 2025. https://getlatka.com/companies/cohere.com
88. “AI Startup Cohere Cuts Staff After $500 Million Funding Round.” CNBC, July 23, 2024. https://www.cnbc.com/2024/07/23/cohere-layoffs-20-employees-cut-following-500-million-funding.html
89. “Exclusive — AI Startup Cohere Seeks $5 Billion Valuation in Latest Fundraising.” Reuters / Investing.com, March 21, 2024. https://investing.com/news/stock-market-news/exclusiveai-startup-cohere-seeks-5-billion-valuation-in-latest-fundraising-source-says-3347228

Industry Context & Comparative Analysis

90. “Why Everybody Is Losing Money on AI.” Where’s Your Ed At, September 15, 2025. https://www.wheresyoured.at/why-everybody-is-losing-money-on-ai/
91. “Anthropic Just Passed OpenAI in Revenue While Spending 4x Less to Train Their Models.” SaaStr, April 7, 2026. https://www.saastr.com/anthropic-just-passed-openai-in-revenue-while-spending-4x-less-to-train-their-models/
92. “Anthropic Hits $30 Billion Run Rate as Enterprise Demand and Compute Deals Reshape AI Race.” MEXC, April 7, 2026. https://www.mexc.com/tr-CT/news/1010401
93. “Is OpenAI in Trouble?” Medium — Kevin O’Shaughnessy, March 25, 2026. https://medium.com/@ZombieCodeKill/is-openai-in-trouble-b9f8cd2a9130
94. “The Real Cost of AI: Inside OpenAI’s $13.5B Burn Rate.” Sanj.dev, October 3, 2025. https://sanj.dev/post/real-cost-of-ai-openai-financials
95. “Compare Klover vs Safe Superintelligence.” CB Insights. https://www.cbinsights.com/compare/klover-2-vs-safe-superintelligence

© 2026 Museum of Vibe Coding — Research Division. All rights reserved. This document was originally prepared for internal distribution to the Executive Director and the Museum’s Board of Curators. It was approved for public release on June 1, 2026. Cite as: Museum of Vibe Coding Research Division. “First Profitable AI Company: A New Era is Born” June 2026. museumofvibecoding.org