SHOW DIRECTORY
20VC
The Twenty Minute VC interviews leading venture capitalists and company founders, hosted by Harry Stebbings.
SOURCE DESCRIPTION · RSS
Fractile’s Walter Goodwin on NVIDIA, Capex and Neo Clouds
Fractile is betting that stacked memory can deliver SRAM-like bandwidth with DRAM capacity, targeting faster inference for long-running agents.Goodwin says reasoning models multiply compute through longer runs and parallel agents, while land and power limits make throughput per megawatt critical.Fractile has fabricated silicon, with full-product tape-outs expected in “a few months”; reliable rack-scale operation and manufacturing volume remain key milestones.
Mistral CEO: Why OpenAI & Anthropic Will Come for Your Business | Are Chinese Models Truly Safe?
Europe’s near-term AI constraint is capital and chips, not electricity, which is about 10% of computing ownership costs; energy bottlenecks are forecast for 2029–2030 as enterprise workloads can jump from eight GPUs to a thousand.Mistral’s sovereign stack combines model training, customer-side data platforms and managed compute; reserved capacity and latency guarantees support premium inference, while committed demand must sustain expansion without excessive leverage.
20VC: Modal’s Erik Bernhardsson on inference and RL’s limits
Harry StebbingsErik Bernhardsson
Modal has reached $1 billion ARR after 15X growth in 12 months, yet capacity spending keeps cash flow negative.Its 25-neocloud pool, per-GPU-second pricing and adjacent products aim to smooth demand, retain customers and lift gross margins toward Bernhardsson’s 40–60% range.Watch whether 30–40% monthly growth can support buying roughly two to three times projected capacity as hardware-price timing remains unknowable.
20VC: Unconventional's Naveen Rao on robotics and $10TRN inference
Rao sees US spare power becoming scarce within years, tying AI growth to construction.Unconventional co-designs models and hardware for efficiency as falling token prices pressure inference margins, despite a $10 trillion market forecast within 10 years.He sees robotics 100 times larger than agentic coding, with prototypes in two to three years and a market in five to seven.
20VC: Trillion or Bust, CoreWeave’s Sell-Off and the $90B Anduril Test
Fund-scale return targets narrow venture: $50–100M seed valuations and $300–500K allocations can make an $18B winner immaterial.CoreWeave must fill capacity and fund buildout; Modal topped $1B run rate after raising $750M at $15.75B, amid unresolved inference competition.Town at $1B and Anduril at $90B highlight entry-price risk, with Anduril’s roughly 40x revenue demanding substantial execution.
20VC: Valar’s Pro-Rata Fight, Meta’s Muse and NVIDIA’s Bet
Pro rata rights emerge as economically valuable options, while Meta’s Muse Wonder signals a return to the model frontier despite unresolved usability questions.Meanwhile, rumored NVIDIA-Reflection and Jev’s $870 million Series A at a $7.5 billion valuation hinge on strategic control, hyper-accurate classification, and whether AI demand can absorb steep inference-price cuts.
Crusoe CEO: Why Everyone Gets GPU Depreciation & AI Energy Costs Wrong
Harry StebbingsChase Lochmiller
Crusoe says AI infrastructure will follow cheap, available power, targeting roughly 200 MW in Abilene within one year versus 2.5 years elsewhere.Vertical integration cut medium-voltage delivery to 28 weeks from a 100-week quote, while five-year GPU contracts and managed services diversify cash flow.Rising Hoppers usage prices challenge simple depreciation assumptions, but labor, community disruption, and uncertain IPO timing remain risks.
Positron's Sohmers: anti-data-center sentiment is a Chinese psyop
Positron co-founder Thomas Sohmers frames inference as a memory-wall problem: GPU FLOPS rose 120x from 2014–2024, versus 17x for memory bandwidth.Positron targets cutting a 500-megawatt Nvidia deployment to 100 megawatts, while cached tokens cost roughly one-thousandth of recomputation and support reported 80-point Anthropic API gross margins.Planned capacity may shift locations rather than disappear, but debt, regulation, and concentration remain risks to monitor.
Speechify's Weitzman on owning GPUs, ElevenLabs, and the AI talent war
Speechify’s owned-GPU strategy compares $35,000–50,000 annual H100 rental with roughly $30,000 to buy, potentially lowering compute costs.NVIDIA’s deal with Blackstone, BlackRock, Apollo and Goldman Sachs could create a liquid secondary market and price floor through buybacks up to 25% of GPU value.Weitzman calls missing B2B his biggest strategic mistake; Simba 3.2 costs $10 per million characters versus ElevenLabs’ $100, but execution against an “unstoppable” rival remains unresolved.
20VC: NVIDIA's record quarter & Hugging Face buy; OpenAI cuts Cursor
NVIDIA delivered a record $96.2B quarter and guided to 70% growth for the fiscal year ending January 2028, with end-user demand and share loss the key risks.The reported near-$12.9B Hugging Face deal, not finalized or confirmed, would reinforce open-source compute economics and NVIDIA’s push to win every segment.Agents are becoming software buyers, supporting the cases for Clay at $7B and Linear at $2.5B, while cyber exposure and reward hacking remain unresolved.









