| 200× | “200 X” | AI-driven robotic expansion could drive 200x global GDP growth in 2030s | Ryan Greenblatt | Aug 27, 2026 | — |
| 4× | “four X” | Greenblatt: Annual AI progress in 2029 will be 4x to 5x faster than in 2025 | Ryan Greenblatt | Aug 27, 2026 | — |
| 5× | “five X” | Greenblatt: Annual AI progress in 2029 will be 4x to 5x faster than in 2025 | Ryan Greenblatt | Aug 27, 2026 | — |
| 90% | “90%” | Wolf: 90% of AI Fake News Is Made by Closed-Source Models | Thomas Wolf | Aug 6, 2026 | — |
| 3M | “three million” | Trojanowski: Over Three Million Accountants Work in the US | Mitch Trojanowski | Aug 5, 2026 | Supported |
| 99% | “99%” | Samsara-connected vehicles drive 99% of U.S. roads every single day | Sanjit Biswas | Jul 29, 2026 | — |
| 10% | “10%” | About 10% of Americans do not regularly wear their seatbelts | Sanjit Biswas | Jul 29, 2026 | Supported |
| 65% | “65%” | Home Depot achieved a 65% reduction in auto claims using Samsara | Sanjit Biswas | Jul 29, 2026 | Supported |
| 3× | “three-X” | Major US utility plans to triple 125-year grid capacity in five years | Sanjit Biswas | Jul 29, 2026 | — |
| 10× | “10 X” | Cerebras cloud achieves 10x inference speedup over fast GPUs on Gemma | Sanjit Biswas | Jul 29, 2026 | Supported |
| 50% | “50%” | Physical operations industries account for 40% to 50% of global GDP | Sanjit Biswas | Jul 29, 2026 | Supported |
| $20B | “twenty billion dollars” | Feldman: Cerebras sales were 10x higher than Groq's at acquisition | Andrew Feldman | Jul 23, 2026 | Contradicted |
| 10× | “10 times” | Feldman: Cerebras sales were 10x higher than Groq's at acquisition | Andrew Feldman | Jul 23, 2026 | Contradicted |
| $20B | “twenty billion dollars” | Feldman: Cerebras signed an OpenAI compute deal worth over $20 billion | Andrew Feldman | Jul 23, 2026 | — |
| 8M | “eight million” | Feldman: Cerebras burned $8M monthly for 18 months before building successful chips | Andrew Feldman | Jul 23, 2026 | — |
| 1.5K | “a half thousand” | Feldman: Cerebras moves weights to compute ~2,500x faster than standard GPUs | Andrew Feldman | Jul 23, 2026 | Supported |
| 70% | “70%” | Feldman: Nvidia CUDA lost 70% of frontier AI model training market share | Andrew Feldman | Jul 23, 2026 | — |
| 5M | “five million” | 5 million Americans now make a living running solo companies | Emily Glassberg Sands | Jul 9, 2026 | Partly supported |
| 1M | “a million” | 5 million Americans now make a living running solo companies | Emily Glassberg Sands | Jul 9, 2026 | Partly supported |
| 300M | “three hundred million” | Stripe's Link consumer wallet now has 300 million users | Emily Glassberg Sands | Jul 9, 2026 | Partly supported |
| 2% | “two percent” | Sands: Stripe sees roughly 2% of global GDP | Emily Glassberg Sands | Jul 9, 2026 | Partly supported |
| 40% | “40%” | Sands: Agent traffic to Stripe docs grew 10x, reaching 40% of total | Emily Glassberg Sands | Jul 9, 2026 | — |
| 10× | “10 X” | Sands: Agent traffic to Stripe docs grew 10x, reaching 40% of total | Emily Glassberg Sands | Jul 9, 2026 | — |
| 70% | “70%” | 70% of Stripe CLI API requests now come from AI agents | Emily Glassberg Sands | Jul 9, 2026 | — |
| 58% | “58%” | Sands: 58% of AI startup Lovable's payment volume flows through Stripe Link | Emily Glassberg Sands | Jul 9, 2026 | — |
| 40% | “40%” | Business registrations jump 40% in Netherlands and 80% in France | Emily Glassberg Sands | Jul 9, 2026 | Partly supported |
| 70% | “70%” | Business registrations jump 40% in Netherlands and 80% in France | Emily Glassberg Sands | Jul 9, 2026 | Partly supported |
| 80% | “80%” | Business registrations jump 40% in Netherlands and 80% in France | Emily Glassberg Sands | Jul 9, 2026 | Partly supported |
| 5× | “five times” | 2026 Stripe Atlas startups track 5x the revenue of 2025 cohort | Emily Glassberg Sands | Jul 9, 2026 | — |
| 10× | “10 times” | NVIDIA DLSS Is About 10 Times More Efficient Than Traditional Rendering | Bryan Catanzaro | Jul 2, 2026 | — |
| 31B | “a thirty billion” | NVIDIA details Nemotron 3 parameter specs from 3B to 55B active | Bryan Catanzaro | Jul 2, 2026 | Supported |
| 3B | “three billion” | NVIDIA details Nemotron 3 parameter specs from 3B to 55B active | Bryan Catanzaro | Jul 2, 2026 | Supported |
| 4× | “four times” | Catanzaro: Nemotron 3's Latent MoE quadruples experts for same inference cost | Bryan Catanzaro | Jul 2, 2026 | Supported |
| 1K | “a thousand” | Prince predicts bot traffic will be 1,000x human traffic within five years | Matthew Prince | Jun 25, 2026 | Open |
| 40× | “40 times” | Prince: One AI agent per worker requires 40x annual global CPU output | Matthew Prince | Jun 25, 2026 | — |
| 10× | “10 times” | Prince: AI will boost engineering and sales hiring while replacing corporate measurement roles | Matthew Prince | Jun 25, 2026 | — |
| 20% | “20%” | Cloudflare laid off over 20% of staff to cut middle management | Matthew Prince | Jun 25, 2026 | — |
| $3M | “three million dollars” | Balaban: Gigawatt AI clusters require up to $45B for servers alone | Stephen Balaban | Jun 18, 2026 | Supported |
| $3B | “three billion dollars” | Balaban: Gigawatt AI clusters require up to $45B for servers alone | Stephen Balaban | Jun 18, 2026 | Supported |
| $15B | “fifteen billion dollars” | Balaban: Gigawatt AI clusters require up to $45B for servers alone | Stephen Balaban | Jun 18, 2026 | Supported |
| $45B | “forty-five billion dollars” | Balaban: Gigawatt AI clusters require up to $45B for servers alone | Stephen Balaban | Jun 18, 2026 | Supported |
| 1B | “a billion” | Balaban: Lambda's cloud revenue run rate is just under $1B | Stephen Balaban | Jun 18, 2026 | — |
| $1B | “a billion dollars” | Balaban: Lambda alumni startup Positron is valued over $1B | Stephen Balaban | Jun 18, 2026 | Supported |
| $1B | “a billion dollars” | Balaban: AI agentic workflows without clear automated feedback are overhyped | Stephen Balaban | Jun 18, 2026 | — |
| $1K | “a thousand dollars” | Levie: A single coding agent task can consume $1,000 in compute | Aaron Levie | May 28, 2026 | — |
| 7% | “seven percent” | Levie: Corporate IT spending is constrained to 3% to 7% of revenue | Aaron Levie | May 28, 2026 | — |
| 7% | “seven percent” | Levie: Enterprise AI spend will shift from IT to line-of-business budgets | Aaron Levie | May 28, 2026 | — |
| 6B | “a five billion” | Levie: A $5B startup will be built around AI compute ERP | Aaron Levie | May 28, 2026 | — |
| 100× | “hundred times” | Levie: Headless AI queries will be 100x larger than interface work | Aaron Levie | May 28, 2026 | — |
| 2× | “two X” | Dubois: GPT-5.5 performs most tasks roughly two times faster | Yann Dubois | May 21, 2026 | — |
| 2× | “two times” | Dubois: Test-time compute scaling exhibits logarithmic, diminishing returns | Yann Dubois | May 21, 2026 | — |
| 1M | “one million” | Dubois: Models like Kimi and DeepSeek use ~1M RL data points | Yann Dubois | May 21, 2026 | Partly supported |
| 5M | “five million” | Burazin: Client requested 5 million concurrent AI agent sandboxes | Ivan Borzin | May 14, 2026 | — |
| 0.5% | “half percent” | Burazin: Sandboxes running over 24 hours represent 2.5% of volume, 20% of revenue | Ivan Borzin | May 14, 2026 | — |
| 10% | “10%” | Evans: 10% of population uses LLMs daily, 50% monthly | Benedict Evans | Mar 19, 2026 | — |
| 50% | “50%” | Evans: 10% of population uses LLMs daily, 50% monthly | Benedict Evans | Mar 19, 2026 | — |
| 900M | “nine hundred million” | Evans: OpenAI has 900M weekly active users, but only 5% pay | Benedict Evans | Mar 19, 2026 | Partly supported |
| 5% | “five percent” | Evans: OpenAI has 900M weekly active users, but only 5% pay | Benedict Evans | Mar 19, 2026 | Partly supported |
| 1K | “a thousand” | Evans: Submitting 1,000 prompts placed ChatGPT users in top 20% | Benedict Evans | Mar 19, 2026 | Supported |
| 20% | “20%” | Evans: Submitting 1,000 prompts placed ChatGPT users in top 20% | Benedict Evans | Mar 19, 2026 | Supported |
| $70B | “seventy billion dollars” | Evans: Nvidia hit over $70B in trailing 12-month free cash flow | Benedict Evans | Mar 19, 2026 | Supported |
| 100% | “hundred percent” | AxiomProver is the first AI to solve research conjectures end-to-end | Corinna Hong | Feb 26, 2026 | — |
| 100× | “hundred times” | Axiom's proof verifier is 100 times faster than open-source alternatives | Corinna Hong | Feb 26, 2026 | Open |
| 7M | “seven million” | Zeghidour: Kyutai trained its Moshi model on 7 million hours of speech | Neil Zeghidour | Feb 19, 2026 | — |
| 10× | “10 X” | Patel: AI workload scale enables 10x performance gains from specialized chips | Dylan Patel | Feb 5, 2026 | — |
| $60M | “sixty million dollars” | Patel: SemiAnalysis runs InferenceMax benchmark on $60M of donated GPUs | Dylan Patel | Feb 5, 2026 | — |
| 1% | “one percent” | Patel: New AI chip startups like Etched have under 1% success chance | Dylan Patel | Feb 5, 2026 | — |
| 2B | “two billion” | Patel: Nvidia is backstopping CoreWeave's land and power capital needs | Dylan Patel | Feb 5, 2026 | Supported |
| 100B | “a hundred billion” | Patel: AI software industry could hit $100 billion ARR this year | Dylan Patel | Feb 5, 2026 | Open |
| 1B | “a billion” | Patel: ChatGPT has roughly one billion users | Dylan Patel | Feb 5, 2026 | Partly supported |
| 500B | “five hundred billion” | Patel: Cumulative semiconductor CapEx in Taiwan exceeds $500 billion | Dylan Patel | Feb 5, 2026 | Supported |
| $500B | “five hundred billion dollars” | Patel: Hyperscaler capex will reach roughly $500 billion this year | Dylan Patel | Feb 5, 2026 | Open |
| 2% | “two percent” | Patel: Two percent of global GitHub commits are generated by Claude Code | Dylan Patel | Feb 5, 2026 | — |
| $2T | “Two trillion dollars” | Patel: Annual global software wages total two trillion dollars | Dylan Patel | Feb 5, 2026 | — |
| 2% | “two percent” | Patel: Data center power consumption will grow from 2% to 10% of US grid | Dylan Patel | Feb 5, 2026 | Open |
| 10% | “10%” | Patel: Data center power consumption will grow from 2% to 10% of US grid | Dylan Patel | Feb 5, 2026 | Open |
| 10% | “10%” | Patel: Data centers will use under 1% of US water by 2030 | Dylan Patel | Feb 5, 2026 | Open |
| 1% | “one percent” | Patel: Data centers will use under 1% of US water by 2030 | Dylan Patel | Feb 5, 2026 | Open |
| 15% | “15%” | 50 RLVR steps boosted Qwen 3 MATH-500 score from 15% to 50% | Sebastian Raschka | Jan 29, 2026 | Partly supported |
| 5% | “five percent” | 50 RLVR steps boosted Qwen 3 MATH-500 score from 15% to 50% | Sebastian Raschka | Jan 29, 2026 | Partly supported |
| 50% | “50%” | 50 RLVR steps boosted Qwen 3 MATH-500 score from 15% to 50% | Sebastian Raschka | Jan 29, 2026 | Partly supported |
| 6M | “a five million” | DeepSeek-R1 cost roughly $300,000 to train, 10x cheaper than DeepSeek-V3 | Sebastian Raschka | Jan 29, 2026 | Supported |
| 95% | “95%” | AI evaluation will shift from single-shot answers to multi-step agentic execution | Sebastian Raschka | Jan 29, 2026 | — |
| 97% | “97%” | AI evaluation will shift from single-shot answers to multi-step agentic execution | Sebastian Raschka | Jan 29, 2026 | — |
| 2× | “two times” | GPT OSS benchmarks demonstrate 1.2x capability jump when tool calling is enabled | Sebastian Raschka | Jan 29, 2026 | Supported |
| 20% | “20%” | Dan Fu: DeepSeek-V3 was trained on ~2,000 H800s with 20% MFU | Dan Fu | Jan 22, 2026 | Partly supported |
| 5T | “five trillion” | Fu predicts increasing hardware diversity, particularly for AI model inference | Dan Fu | Jan 22, 2026 | — |
| 10× | “10 times” | Fu: AI coding tools enable expert programmers to move 10x faster | Dan Fu | Jan 22, 2026 | — |
| 100× | “hundred times” | Ai2 to release open-source coding agent with 100x cheaper training | Tim Dettmers | Jan 22, 2026 | — |
| 5% | “five percent” | Fu: Hardware utilization during AI inference is under 5% | Dan Fu | Jan 22, 2026 | — |
| 2× | “two X” | Dan Fu: Proper speculative decoding yields 2x to 3x model speedups | Dan Fu | Jan 22, 2026 | — |
| 3× | “three X” | Dan Fu: Proper speculative decoding yields 2x to 3x model speedups | Dan Fu | Jan 22, 2026 | — |
| 10T | “10 trillion” | Ai2 samples 6T tokens from 10T pool for OLMo 3 | Luca Soldaini | Nov 20, 2025 | — |
| 6T | “six trillion” | Ai2 samples 6T tokens from 10T pool for OLMo 3 | Luca Soldaini | Nov 20, 2025 | — |
| 95% | “95%” | Soldaini: 95% of web pages are under 3,000 tokens | Luca Soldaini | Nov 20, 2025 | — |
| 80% | “80%” | Lambert: 80% of a16z's open-model portfolio startups use Alibaba's Qwen | Nathan Lambert | Nov 20, 2025 | Partly supported |
| 24% | “24%” | Lambert: 80% of a16z's open-model portfolio startups use Alibaba's Qwen | Nathan Lambert | Nov 20, 2025 | Partly supported |
| 2M | “two million” | Ai2 received an initial grant of two million GPU hours from AMD | Luca Soldaini | Nov 20, 2025 | Open |
| 10× | “10 times” | Lambert: Scaling AI 10x alters post-training, not pre-training methods | Nathan Lambert | Nov 20, 2025 | — |
| 98% | “98%” | Lambert: Big tech will realize 95-98% of LLM potential by 2030 | Nathan Lambert | Nov 20, 2025 | — |
| 80% | “80%” | Kant: AI foundation model gross margins will settle near 40% | Eiso Kant | Nov 6, 2025 | — |
| 40% | “40%” | Kant: AI foundation model gross margins will settle near 40% | Eiso Kant | Nov 6, 2025 | — |
| $1.5B | “half a billion dollars” | Kant: A 250MW AI data center costs roughly $8 billion to build | Eiso Kant | Nov 6, 2025 | — |
| $2B | “two billion dollars” | Kant: A 250MW AI data center costs roughly $8 billion to build | Eiso Kant | Nov 6, 2025 | — |
| 9B | “an eight billion” | Kant: A 250MW AI data center costs roughly $8 billion to build | Eiso Kant | Nov 6, 2025 | — |
| 40% | “40%” | Kant: Full-stack AI infrastructure ownership cuts token costs 20-40% | Eiso Kant | Nov 6, 2025 | — |
| 3× | “three times” | Kant: Poolside first compute online late 2025, 250MW by Q1 2027 | Eiso Kant | Nov 6, 2025 | Didn’t hold up |
| 1M | “a million” | Poolside scales its RL environment repository to over 1M codebases | Eiso Kant | Nov 6, 2025 | — |
| 30M | “thirty million” | Poolside scales its RL environment repository to over 1M codebases | Eiso Kant | Nov 6, 2025 | — |
| 70% | “70%” | Kant: Expert RL will lift enterprise agent quality to 90% in years | Eiso Kant | Nov 6, 2025 | — |
| 100% | “hundred percent” | Kant: Expert RL will lift enterprise agent quality to 90% in years | Eiso Kant | Nov 6, 2025 | — |
| 95% | “95%” | Kant: 95% of Poolside's researchers sit outside Silicon Valley | Eiso Kant | Nov 6, 2025 | — |
| 50% | “50%” | Corporate AI subscription retention rose from 50% to 80% since 2022 | Nathan Benaich | Oct 30, 2025 | Supported |
| 80% | “80%” | Corporate AI subscription retention rose from 50% to 80% since 2022 | Nathan Benaich | Oct 30, 2025 | Supported |
| $1.5M | “half a million dollars” | Average corporate AI spend hits $500,000 and will reach $1M | Nathan Benaich | Oct 30, 2025 | Held up |
| $1M | “a million dollars” | Average corporate AI spend hits $500,000 and will reach $1M | Nathan Benaich | Oct 30, 2025 | Held up |
| 95% | “95%” | 76% of surveyed AI practitioners pay for tools out of pocket | Nathan Benaich | Oct 30, 2025 | Supported |
| 76% | “76%” | 76% of surveyed AI practitioners pay for tools out of pocket | Nathan Benaich | Oct 30, 2025 | Supported |
| 10% | “10%” | 76% of surveyed AI practitioners pay for tools out of pocket | Nathan Benaich | Oct 30, 2025 | Supported |
| 70% | “70%” | 76% of surveyed AI practitioners pay for tools out of pocket | Nathan Benaich | Oct 30, 2025 | Supported |
| 30% | “30%” | Vertical AI application products generate gross margins around 30% | Nathan Benaich | Oct 30, 2025 | — |
| 90% | “90%” | Top AI serving companies achieve 70% to 90% gross margins | Nathan Benaich | Oct 30, 2025 | — |
| $50B | “fifty billion dollars” | A 1GW AI data center requires $50B CapEx and $8B-$11B annually | Nathan Benaich | Oct 30, 2025 | — |
| $11B | “eleven billion dollars” | A 1GW AI data center requires $50B CapEx and $8B-$11B annually | Nathan Benaich | Oct 30, 2025 | — |
| 90% | “90%” | NVIDIA chips power 90% of open-source AI research papers | Nathan Benaich | Oct 30, 2025 | Supported |
| 12× | “12 X” | NVIDIA stock outperformed private AI chip startup investments 12x to 2x | Nathan Benaich | Oct 30, 2025 | Supported |
| 2× | “two X” | NVIDIA stock outperformed private AI chip startup investments 12x to 2x | Nathan Benaich | Oct 30, 2025 | Supported |
| $20B | “twenty billion dollars” | NVIDIA's sovereign AI business generates around $20 billion in revenue | Nathan Benaich | Oct 30, 2025 | Supported |
| 50% | “50%” | 50% of Hugging Face model derivatives are now based on Qwen | Nathan Benaich | Oct 30, 2025 | Partly supported |
| $34M | “thirty four million dollars” | Independent AI safety organizations spend an average of $134 million annually | Nathan Benaich | Oct 30, 2025 | — |
| 1.5B | “a half billion” | Anthropic agreed to a $1.5 billion training data copyright settlement | Nathan Benaich | Oct 30, 2025 | Supported |
| $100M | “hundred million dollars” | Douglas: Google LLM inference stack saved hundreds of millions in months | Sholto Douglas | Oct 2, 2025 | — |
| 10% | “10%” | Douglas: Even AI pioneer Noam Shazeer only sees 10% of ideas work | Sholto Douglas | Oct 2, 2025 | — |
| 1K | “a thousand” | Douglas: DeepMind has 1,000 on Gemini and 10,000 on foundational research | Sholto Douglas | Oct 2, 2025 | — |
| 20% | “20%” | Douglas: The entire AI industry scored under 20% on SWE-bench last year | Sholto Douglas | Oct 2, 2025 | Partly supported |
| 1K | “a thousand” | Crespo: Pigment becomes useful for companies at around 1,000 employees | Eleonore Crespo | Sep 11, 2025 | — |
| 60% | “60%” | Crespo: The US accounts for 60% of Pigment's revenue | Eleonore Crespo | Sep 11, 2025 | — |
| 25M | “twenty-five million” | Deloitte partners carry a $25 million annual quota, says Crespo | Eleonore Crespo | Sep 11, 2025 | — |
| 2M | “two million” | Deloitte partners carry a $25 million annual quota, says Crespo | Eleonore Crespo | Sep 11, 2025 | — |
| 100B | “a hundred billion” | Crespo: Pigment aims to build a $100B+ to $200B+ business | Eleonore Crespo | Sep 11, 2025 | — |
| 200B | “two hundred billion” | Crespo: Pigment aims to build a $100B+ to $200B+ business | Eleonore Crespo | Sep 11, 2025 | — |
| 70% | “70%” | Valenzuela: AI research lab gross margins currently range from 40% to 70% | Cris Valenzuela | Sep 4, 2025 | — |
| 90% | “90%” | Valenzuela: Full-stack AI companies will eventually reach 80% to 90% SaaS margins | Cris Valenzuela | Sep 4, 2025 | — |
| 50% | “50%” | Granola cut 50 percent of its product features prior to launch | Chris Pedregal | Aug 21, 2025 | — |
| 400M | “four hundred million” | Turck: The Information reports Claude Code hit $400M ARR in six months | Matt Turck | Aug 7, 2025 | Supported |
| 70% | “70%” | Misha Laskin: Engineers spend 70% of their time searching information, not coding | Misha Laskin | Jul 17, 2025 | Partly supported |
| 100× | “hundred X” | Laskin: Focused AI startups can operate with 10x less capital than frontier labs | Misha Laskin | Jul 17, 2025 | — |
| 10× | “10 X” | Laskin: Focused AI startups can operate with 10x less capital than frontier labs | Misha Laskin | Jul 17, 2025 | — |
| $1.4T | “1.4 trillion dollars” | Stripe processed $1.4 trillion in 2024, representing 1.3% of global GDP | Emily Glassberg Sands | Jul 10, 2025 | Supported |
| 1.3% | “1.3%” | Stripe processed $1.4 trillion in 2024, representing 1.3% of global GDP | Emily Glassberg Sands | Jul 10, 2025 | Supported |
| 38% | “38%” | Stripe processed $1.4 trillion in 2024, representing 1.3% of global GDP | Emily Glassberg Sands | Jul 10, 2025 | Supported |
| 7× | “seven times” | Businesses using Stripe grew seven times faster than the S&P 500 | Emily Glassberg Sands | Jul 10, 2025 | — |
| 59% | “59%” | Stripe's AI foundation model increased card testing fraud detection to 97% | Emily Glassberg Sands | Jul 10, 2025 | — |
| 97% | “97%” | Stripe's AI foundation model increased card testing fraud detection to 97% | Emily Glassberg Sands | Jul 10, 2025 | — |
| $55B | “fifty five billion dollars” | Merchants lose approximately $55 billion annually to chargebacks | Emily Glassberg Sands | Jul 10, 2025 | Supported |
| 13% | “13%” | Vimeo and Squarespace recover 13% more disputed revenue using Stripe AI | Emily Glassberg Sands | Jul 10, 2025 | — |
| 5% | “five percent” | Turo recaptured over $100 million annually by adopting Stripe Checkout | Emily Glassberg Sands | Jul 10, 2025 | — |
| 12% | “12%” | Displaying non-card payment options increases merchant revenue by roughly 12% | Emily Glassberg Sands | Jul 10, 2025 | — |
| 7% | “seven percent” | Displaying non-card payment options increases merchant revenue by roughly 12% | Emily Glassberg Sands | Jul 10, 2025 | — |
| 65% | “65%” | Decagon's first Stripe MCP integration customer saw support costs drop 65% | Emily Glassberg Sands | Jul 10, 2025 | — |
| 78% | “78%” | 78% of the Forbes AI 50 list use Stripe for payments | Emily Glassberg Sands | Jul 10, 2025 | — |
| 100% | “100%” | 78% of the Forbes AI 50 list use Stripe for payments | Emily Glassberg Sands | Jul 10, 2025 | — |
| 30M | “thirty million” | AI startups reach $30M ARR three times faster than fast-growing SaaS startups | Emily Glassberg Sands | Jul 10, 2025 | — |
| 3× | “three X” | AI startups reach $30M ARR three times faster than fast-growing SaaS startups | Emily Glassberg Sands | Jul 10, 2025 | — |
| 50M | “fifty million” | Stockholm-based AI startup Lovable reached $50 million ARR in six months | Emily Glassberg Sands | Jul 10, 2025 | Supported |
| 300M | “three hundred million” | AI coding assistant Cursor passed $300 million ARR two years after launch | Emily Glassberg Sands | Jul 10, 2025 | Supported |
| 56% | “56%” | The median top AI startup generates 56% of revenue internationally | Emily Glassberg Sands | Jul 10, 2025 | — |
| 50M | “fifty million” | AI photo editor PhotoRoom reached $50 million ARR in three years | Emily Glassberg Sands | Jul 10, 2025 | Supported |
| 7× | “seven x” | Walsher: Global software developer population grew 7x since 2006 | David Walsher | Jul 3, 2025 | Contradicted |
| $500M | “five hundred million dollars” | Walsher: Cursor announced reaching $500M in ARR | David Walsher | Jul 3, 2025 | — |
| $60M | “sixty million dollars” | Walsher: Lovable reached $60M ARR in two quarters | David Walsher | Jul 3, 2025 | — |
| $400M | “Four hundred million dollars” | Walsher: GitHub Copilot reached $400M ARR and 15M developers | David Walsher | Jul 3, 2025 | — |
| 15M | “fifteen million” | Walsher: GitHub Copilot reached $400M ARR and 15M developers | David Walsher | Jul 3, 2025 | — |
| $100M | “a hundred million dollars” | Walsher: Windsurf recently reached $100M in ARR | David Walsher | Jul 3, 2025 | — |
| $100M | “a hundred million dollars” | Walsher: Replit grew ARR from $10M to $100M in six months | David Walsher | Jul 3, 2025 | — |
| 50% | “50%” | FirstMark survey: 82% of developers use AI to write code | David Walsher | Jul 3, 2025 | — |
| 12% | “12%” | FirstMark survey: 82% of developers use AI to write code | David Walsher | Jul 3, 2025 | — |
| 17% | “17%” | FirstMark survey: 82% of developers use AI to write code | David Walsher | Jul 3, 2025 | — |
| 82% | “82%” | FirstMark survey: 82% of developers use AI to write code | David Walsher | Jul 3, 2025 | — |
| 10× | “10 X” | Turck: AI-native startups hold more buyer credibility than larger incumbents | Matt Turck | Jul 3, 2025 | — |
| 100× | “hundred X” | Turck: AI-native startups hold more buyer credibility than larger incumbents | Matt Turck | Jul 3, 2025 | — |
| 80B | “eighty billion” | Walsher: Tech buyers value Cursor's opinion over $80B legacy software incumbents | David Walsher | Jul 3, 2025 | — |
| 100M | “a hundred million” | Rauch: Vercel's v0 has generated over 100 million applications | Guillermo Rauch | Jun 26, 2025 | — |
| 1.4M | “1.4 million” | Rauch: Next.js has 1.4 million monthly active developers as top React framework | Guillermo Rauch | Jun 26, 2025 | Supported |
| 1T | “a trillion” | Rauch: Vercel's fluid compute platform reached 1 trillion function invocations | Guillermo Rauch | Jun 26, 2025 | Supported |
| 100× | “hundred times” | Rauch: MCP will replace human business development with agent-to-agent meetings | Guillermo Rauch | Jun 26, 2025 | — |
| 100× | “hundred X” | Rauch: Vercel v0 traction improved 100x following interface redesign | Guillermo Rauch | Jun 26, 2025 | — |
| 100M | “a hundred million” | Canva Code went from concept to production in just three months | Brendan Humphreys | Jun 20, 2025 | — |
| 30% | “30%” | Canva engineers using AI coding tools merge 30% more pull requests | Brendan Humphreys | Jun 20, 2025 | — |
| $200M | “two hundred million dollars” | Canva established a $200 million fund to compensate AI training contributors | Brendan Humphreys | Jun 20, 2025 | Supported |
| 1B | “a billion” | Dohmke: Visual Studio was a billion-dollar business in 2018 | Thomas Dohmke | Jun 12, 2025 | Not publicly verifiable |
| 2B | “two billion” | Dohmke: GitHub passed $2 billion in ARR in July 2024 | Thomas Dohmke | Jun 12, 2025 | Supported |
| 200M | “two hundred million” | Dohmke: GitHub grew from $200M ARR in 2017 to $2B in 2024 | Thomas Dohmke | Jun 12, 2025 | — |
| 10× | “10 X” | Dohmke: GitHub grew from $200M ARR in 2017 to $2B in 2024 | Thomas Dohmke | Jun 12, 2025 | — |
| 90% | “90%” | Dohmke: OpenAI's Codex solved over 90% of GitHub interview coding exercises | Thomas Dohmke | Jun 12, 2025 | — |
| 90% | “90%” | Dohmke: AI coding agents require 90% benchmark accuracy for broad adoption | Thomas Dohmke | Jun 12, 2025 | — |
| 70% | “70%” | Dohmke: Multilingual AI coding benchmarks hover around 20% to 30% success rate | Thomas Dohmke | Jun 12, 2025 | Supported |
| 30% | “30%” | Dohmke: Multilingual AI coding benchmarks hover around 20% to 30% success rate | Thomas Dohmke | Jun 12, 2025 | Supported |
| 5B | “five billion” | Hebbia is on track to process five billion document pages in 2025 | George Sivulka | May 29, 2025 | — |
| 50% | “50%” | Hebbia is deployed at 40% to 50% of top asset managers | George Sivulka | May 29, 2025 | — |
MADLEDGER All shows Try the podcast clipper →