Teahose.
SIGN IN
NEW HERE — WHAT TEAHOSE DOES
We read the entire AI & tech firehose — so you don't have to.
PODPodcastsAll-In, No Priors, Acquired…
NEWNewslettersStratechery, Newcomer…
PAPPapersPhysical AI research
PHProduct Huntdaily launches
VCInvestor ScoutSequoia, a16z, Benchmark…
CLAUDE DISTILLS →
7 reads, 30 sec each — free, 6 AM ET.
+ a live graph of the companies, people & themes underneath.
HOME/STRICTLYVC/The FTC Is Investigating OpenAI,…
NEWS
// NEWSLETTER ISSUE
STRICTLYVC

The FTC Is Investigating OpenAI, Anthropic, and Other Leading Labs

DATE October 1, 2026SOURCE STRICTLYVCPARTICIPANTS CONNIE LOIZOS
In this episode
// SUMMARY

1. Key Themes

Theme: Regulatory and safety scrutiny of frontier AI labs is intensifying, and it's already delaying capital-markets events

The FTC is moving against self-regulation

"The FTC has opened an investigation into OpenAI, Anthropic, and other frontier AI labs and plans to issue subpoena-like demands for documents and executive testimony about potential consumer risks, just as the industry is pledging to police itself under a voluntary White House safety accord."

Safety concerns are now a gating factor for IPOs

"Wall Street's expected fall IPO boom is losing steam as market volatility, rising bond yields, valuation pushback, and AI safety concerns delay offerings from Oura, SB Energy, NScale, Anthropic, OpenAI, and others."

The IPO window is deteriorating on performance, not just sentiment

"U.S. companies that went public this year are now trading less than 1% above their IPO prices on average, down from 24% at the end of June."

OpenAI's CEO is publicly tying IPO timing to safety claims

"OpenAI CEO Sam Altman says the company will not go public until it can make stronger claims about the safety of its models, arguing that an IPO could create pressure to satisfy Wall Street at a moment when the company is trying to put safety and alignment ahead of further capability gains."

Theme: Agentic AI is being applied vertically across nearly every function, and funding is following

Coding agents are the hottest, most contested vertical

"Three-year-old, San Francisco-based Factory is an agentic coding startup whose AI agents can carry out programming tasks largely on their own. The company raised $200 million at a $5 billion valuation this month from Khosla Ventures, Blackstone, Sequoia Capital, Insight Partners, and others."

"Cognition raised another $2 billion at a $48 billion valuation earlier this month, and counts Mercedes-Benz, NASA, Goldman Sachs, and Citi as customers."

Agents are being aimed at HR, recruiting, and back-office workflows

"Metaview, an eight-year-old London and San Francisco startup that uses AI agents to automate candidate sourcing, application review, interview screening, and recruiting workflows, raised a $60 million Series C round led by Insight Partners."

"Ahron, a one-year-old Munich startup that uses AI agents to automate hiring feedback, payroll audits, and employee onboarding across companies' existing HR systems, raised a $2.5 million round led by Cusp Capital."

Hardware engineering and industrial robotics are getting agent/vision treatment

"Flow Engineering, a nine-year-old San Francisco startup that develops AI agents that track and verify changes across complex hardware engineering programs, raised a $50 million Series B round at a $750 million valuation."

"Inbolt, a seven-year-old Paris startup that develops 3D vision and AI software that lets industrial robots detect changing conditions and adjust their movements, raised a $12.5 million round."

Theme: A "picks and shovels" build-out for the agent economy (trust, identity, reliability, observability)

Infrastructure for keeping agents running

"Restate, a four-year-old Berlin startup that provides infrastructure for keeping multistep software and AI-agent workflows running through crashes and network failures, raised a $20 million Series A round led by Singular, with Redpoint Ventures and Capital One Ventures also anteing up."

Identity and authority verification for agents

"Beltic, a one-year-old San Francisco startup that verifies AI agents' identities and authority before they conduct transactions on businesses' behalf, raised a $7.3 million seed round led by Norwest."

Enterprise AI spend and ROI tracking

"Ascerta, a three-year-old startup based in Bellevue, WA, that tracks enterprise AI usage, costs, adoption, and financial returns across users, teams, and applications, raised an $18 million Series A round led by Dell Technologies Capital."

Offensive security tooling is being augmented by AI

"RemoteThreat, a one-year-old startup based in Fulton, MD, that develops AI-assisted tools for enterprise and government teams to plan, execute, and oversee offensive cyber operations, raised a $7 million pre-seed round co-led by Osage University Partners and DataTribe."

Theme: Compute bottlenecks are shifting to interconnect, efficiency, and GPU access, attracting large rounds

Optical interconnect between accelerators

"CScale, a three-year-old Palo Alto startup that develops optical interconnect technology linking AI accelerators across large data centers, raised a $145 million Series C round. The deal was co-led by Atreides Management, Valor Equity Partners, and Premji Invest, with NVIDIA and Intel Capital as well as previous investors..."

Energy-efficient processors for edge and data center

"Efficient Computer, a four-year-old Pittsburgh startup that develops energy-efficient general-purpose processors for robots, autonomous drones, and data center workloads, raised a $97+ million Series B round at a $650 million valuation."

GPU cloud capacity and debt financing

"GMI Cloud... raised a $223 million Series B round led by ARCHIV... The company also secured a $440 million credit facility from ChinaTrust Commercial Bank."

Data center opposition is growing

"Former U.S. Vice President and renowned climate activist Al Gore is less worried about data centers than you might expect, even as opposition to them grows around the world."

Theme: Late-stage valuations are escalating fast, with employee liquidity as a key tool

ElevenLabs doubled its valuation in under a year via tender

"ElevenLabs... completed a $300 million employee tender offer at a $22 billion valuation, double its $11 billion valuation in February. The tender was led by Wellington and T. Rowe Price, with EQT and Goldman Sachs as well as previous investors Andreessen Horowitz and Lightspeed Venture Partners also taking part."


2. Contrarian Perspectives

A prominent climate voice is relatively sanguine on data centers

"Former U.S. Vice President and renowned climate activist Al Gore is less worried about data centers than you might expect, even as opposition to them grows around the world."

Why it matters: This runs against the prevailing narrative that climate-focused figures and local communities are united against AI infrastructure. It's relevant for anyone underwriting data center or energy-adjacent investments (details are in the StrictlyVC Download interview, not summarized here).

The CEO of the leading AI lab says he won't IPO, framing going public as a safety risk

"OpenAI CEO Sam Altman says the company will not go public until it can make stronger claims about the safety of its models, arguing that an IPO could create pressure to satisfy Wall Street at a moment when the company is trying to put safety and alignment ahead of further capability gains."

Why it matters: Against the consensus that mega-labs are racing to list to access public capital, Altman is positioning the IPO as something that could compromise safety priorities. This reframes OpenAI's timeline and signals that private-market capital and tenders may remain the primary funding path.

Big Tech may be gaming tax policy on AI infrastructure

"Meta has been classifying its multibillion-dollar AI data centers as experimental 'pilot models' to claim a federal research tax credit, a strategy that shaved $3.9 billion off its 2025 tax bill even as the company warns investors that the IRS could overturn billions of dollars in related tax benefits."

Why it matters: The true economics of AI capex may be flattered by aggressive tax treatment that carries IRS reversal risk, a potential hidden liability in hyperscaler earnings quality.


3. Companies Identified

FTC (Federal Trade Commission)

  • Description: U.S. consumer-protection and competition regulator.
  • Why mentioned: Opened an investigation into frontier AI labs.
  • Quote: "The FTC has opened an investigation into OpenAI, Anthropic, and other frontier AI labs and plans to issue subpoena-like demands for documents and executive testimony about potential consumer risks."

OpenAI

  • Description: Frontier AI lab behind ChatGPT.
  • Why mentioned: FTC target; IPO delayed on safety grounds; bitter rivalry with Anthropic.
  • Quote: "OpenAI CEO Sam Altman says the company will not go public until it can make stronger claims about the safety of its models."

Anthropic

  • Description: Frontier AI lab behind Claude.
  • Why mentioned: FTC target; IPO delayed; rivalry with OpenAI.
  • Quote: "AI safety concerns delay offerings from Oura, SB Energy, NScale, Anthropic, OpenAI, and others."

Factory

  • Description: Three-year-old San Francisco agentic coding startup.
  • Why mentioned: Raised $200M at $5B; CEO publicly accused board advisor of leaking info to competitor.
  • Quote: "Its customers include Nvidia, Blackstone, Royal Bank of Canada, Palo Alto Networks, Adobe, and T-Mobile."

Cognition

  • Description: Maker of the Devin coding agent.
  • Why mentioned: Factory's biggest competitor; raised $2B at $48B; hired Chris Degnan as CRO.
  • Quote: "Cognition raised another $2 billion at a $48 billion valuation earlier this month, and counts Mercedes-Benz, NASA, Goldman Sachs, and Citi as customers."

ElevenLabs

  • Description: Four-year-old London AI voice-generation and voice-agent company.
  • Why mentioned: $22B valuation, doubled since February.
  • Quote: "completed a $300 million employee tender offer at a $22 billion valuation, double its $11 billion valuation in February."

CScale

  • Description: Palo Alto optical interconnect startup for AI accelerators.
  • Why mentioned: $145M Series C with NVIDIA and Intel Capital participating.
  • Quote: "develops optical interconnect technology linking AI accelerators across large data centers, raised a $145 million Series C round."

Efficient Computer

  • Description: Pittsburgh energy-efficient processor startup.
  • Why mentioned: $97M+ Series B at $650M valuation.
  • Quote: "develops energy-efficient general-purpose processors for robots, autonomous drones, and data center workloads."

GMI Cloud

  • Description: Taiwan-based on-demand GPU infrastructure provider.
  • Why mentioned: $223M Series B plus $440M credit facility.
  • Quote: "The company also secured a $440 million credit facility from ChinaTrust Commercial Bank."

Hertha Metals

  • Description: Conroe, TX producer of steel and high-purity iron for rare-earth magnets.
  • Why mentioned: $133.65M Series A with U.S. government participation.
  • Quote: "produces steel and high-purity iron for rare-earth magnets using a single continuous reactor."

Jeeves

  • Description: Orlando stablecoin-based corporate cards and cross-border payments.
  • Why mentioned: $110M Series C; $570M+ raised total.
  • Quote: "uses stablecoins to provide corporate cards, cross-border payments, treasury tools, and spend management for global businesses."

Metaview

  • Description: AI recruiting agent platform.
  • Why mentioned: $60M Series C led by Insight Partners.
  • Quote: "uses AI agents to automate candidate sourcing, application review, interview screening, and recruiting workflows."

Flow Engineering

  • Description: AI agents for hardware engineering programs.
  • Why mentioned: $50M Series B at $750M valuation.
  • Quote: "develops AI agents that track and verify changes across complex hardware engineering programs."

Miter

  • Description: Construction contractor payroll/HR/field-ops platform.
  • Why mentioned: $40M Series B led by Battery Ventures.
  • Quote: "manages payroll, HR, field operations, job costing, and expenses for construction contractors."

Restate

  • Description: Berlin durable-execution infrastructure for agent workflows.
  • Why mentioned: $20M Series A.
  • Quote: "provides infrastructure for keeping multistep software and AI-agent workflows running through crashes and network failures."

Ascerta

  • Description: Enterprise AI usage and ROI tracking.
  • Why mentioned: $18M Series A led by Dell Technologies Capital.
  • Quote: "tracks enterprise AI usage, costs, adoption, and financial returns across users, teams, and applications."

Novele

  • Description: Stamford, CT energy-storage hardware plus AI software for commercial buildings.
  • Why mentioned: $17M Series A.
  • Quote: "combines energy-storage hardware and AI software to manage power use in commercial buildings."

Inbolt

  • Description: Paris 3D vision/AI for industrial robots.
  • Why mentioned: $12.5M round.
  • Quote: "lets industrial robots detect changing conditions and adjust their movements."

Kanu AI

  • Description: Seattle startup turning employee expertise into operational software.
  • Why mentioned: $11.7M seed with a16z, Accel, BMW i Ventures.
  • Quote: "uses AI to turn employees' workflows and business expertise into operational software running in customers' cloud environments."

Hop Aero

  • Description: Autonomous suborbital cargo rockets.
  • Why mentioned: $11M SAFE round led by Initialized.
  • Quote: "develops autonomous suborbital rockets for delivering cargo hundreds of miles in minutes."

Zenithon

  • Description: London AI simulation for fusion, rockets, semiconductors.
  • Why mentioned: $10M seed.
  • Quote: "develops AI models that rapidly simulate designs for fusion reactors, rockets, and semiconductor manufacturing."

Destro

  • Description: Brooklyn warehouse robot/human coordination software.
  • Why mentioned: $8M seed.
  • Quote: "develops software for coordinating warehouse robots and human workers across logistics workflows."

Beltic

  • Description: AI agent identity verification.
  • Why mentioned: $7.3M seed led by Norwest.
  • Quote: "verifies AI agents' identities and authority before they conduct transactions on businesses' behalf."

RemoteThreat

  • Description: AI-assisted offensive cyber operations tools.
  • Why mentioned: $7M pre-seed.
  • Quote: "develops AI-assisted tools for enterprise and government teams to plan, execute, and oversee offensive cyber operations."

Elio Mortgage

  • Description: AI-driven mortgage brokerage founded this year.
  • Why mentioned: $5.1M pre-seed.
  • Quote: "operates a mortgage brokerage that uses AI to automate home-loan origination and provide embedded mortgage services."

Champion

  • Description: Minneapolis AI sales lead-generation startup.
  • Why mentioned: $5M Series A.
  • Quote: "uses AI to help sales teams identify enthusiastic customers and generate higher-quality leads."

Charter Space

  • Description: Insurance brokerage for satellite operators and space companies.
  • Why mentioned: $5M seed.
  • Quote: "provides insurance brokerage services for satellite operators and other space companies."

Cotierra

  • Description: Zurich decentralized biochar and carbon-removal tracking.
  • Why mentioned: $3M round.
  • Quote: "develops decentralized biochar systems that convert agricultural waste into soil-improving biochar and track carbon removal."

Ahron

  • Description: Munich AI HR automation startup.
  • Why mentioned: $2.5M round led by Cusp Capital.
  • Quote: "uses AI agents to automate hiring feedback, payroll audits, and employee onboarding across companies' existing HR systems."

PurposeMed (Freddie)

  • Description: Calgary telehealth startup and HIV-prevention provider.
  • Why mentioned: Acquired by Grindr for $250M.
  • Quote: "Grindr agreed to acquire PurposeMed... for $250 million in cash and stock, its first major acquisition as it expands beyond dating into healthcare."

Grindr

  • Description: Dating platform expanding into healthcare.
  • Why mentioned: Acquirer of PurposeMed.
  • Quote: "its first major acquisition as it expands beyond dating into healthcare."

BAG Ventures

  • Description: New San Francisco VC fund by ex-Google executives.
  • Why mentioned: Closed $11.3M debut fund for early-stage AI.
  • Quote: "closed an $11.3 million debut fund to back early-stage AI startups, writing checks of $100,000 to $500,000."

Meta

  • Description: Big Tech platform company.
  • Why mentioned: Tax strategy around AI data centers.
  • Quote: "classifying its multibillion-dollar AI data centers as experimental 'pilot models' to claim a federal research tax credit."

Amazon

  • Description: E-commerce and logistics giant.
  • Why mentioned: Smart glasses for delivery drivers raising privacy issues.
  • Quote: "Amazon plans to equip 5,000 delivery drivers with smart glasses by year-end that can capture thousands of images during a shift, including of customers' private property."

Google

  • Description: Big Tech company.
  • Why mentioned: Pushing into Hollywood via 100 Zeros.
  • Quote: "Google is pushing deeper into Hollywood through 100 Zeros, a year-old venture with Range Media Partners that finances films and short-form shows."

Nike

  • Description: Athletic footwear company.
  • Why mentioned: New $350 Apex marathon shoe.
  • Quote: "Nike says the combination delivers up to 40% more energy return than the Alphafly 3."

4. People Identified

Matan Grinberg

  • Description: Co-founder and CEO of Factory.
  • Why mentioned: Publicly accused a board adviser of leaking confidential information to a competitor.
  • Quote: "Grinberg alleges that Degnan shared confidential information with Factory's biggest competitor, Cognition."

Chris Degnan

  • Description: Former Snowflake CRO; partner at RPT Partners; Iconiq go-to-market adviser; now Cognition CRO.
  • Why mentioned: Subject of the spying allegation; disputes it.
  • Quote: "Degnan was Snowflake's first sales hire and later spent 11 years as the data company's chief revenue officer."

Al Gore

  • Description: Former U.S. Vice President, climate activist, and investor.
  • Why mentioned: Less concerned about data centers than expected.
  • Quote: "Al Gore is less worried about data centers than you might expect, even as opposition to them grows around the world."

Lila Preston

  • Description: Gore's investing partner.
  • Why mentioned: Co-interviewee in StrictlyVC Download.
  • Quote: "We talked with Gore and his investing partner, Lila Preston, for the newest edition of StrictlyVC Download."

Sam Altman

  • Description: CEO of OpenAI.
  • Why mentioned: Won't take OpenAI public until safety claims are stronger; subject of the Amodei rivalry.
  • Quote: "an IPO could create pressure to satisfy Wall Street at a moment when the company is trying to put safety and alignment ahead of further capability gains."

Dario Amodei

  • Description: Anthropic CEO and ex-OpenAI executive.
  • Why mentioned: Origin and intensity of OpenAI/Anthropic rivalry.
  • Quote: "Dario Amodei's 2020 split from OpenAI after he became convinced Sam Altman and Greg Brockman were putting commercial priorities ahead of safety."

Greg Brockman

  • Description: OpenAI co-founder.
  • Why mentioned: Named by Amodei as prioritizing commercial goals over safety.
  • Quote: "Sam Altman and Greg Brockman were putting commercial priorities ahead of safety."

Bill Draper

  • Description: Pioneering VC, co-founder of Draper & Johnson, Sutter Hill investor; died at 98.
  • Why mentioned: Obituary recognizing early Silicon Valley investing legacy.
  • Quote: "His early bets included Activision, Skype, and OpenTable."

Bonita Stewart and Jackson Georges Jr.

  • Description: Former Google executives; founders of BAG Ventures.
  • Why mentioned: Launched a debut AI-focused fund.
  • Quote: "founded by former Google executives Bonita Stewart and Jackson Georges Jr., closed an $11.3 million debut fund."

Bob Berg

  • Description: Chair of IF4Peace.
  • Why mentioned: Sponsor message seeking founders/investors for peace-tech systems.
  • Quote: "IF4Peace is bringing together technology, systems thinking and leadership to tackle one of humanity's oldest challenges."

5. Operating Insights

Treat your cap table and advisor relationships as a competitive-intelligence surface

The Factory/Cognition dispute shows how board advisors and investors with go-to-market access can sit adjacent to your top competitor. Grinberg said he "trusted and believed" assurances that Degnan wasn't interested in the competitor, which then reversed.

"Degnan had earlier acknowledged having a 'casual' conversation with a Cognition executive but had assured Grinberg that he had no interest in working for the competitor... 'made too much money' and was 'too lazy to go work for Cognition.'"

Takeaway: Formalize information barriers, limit what advisors see, and put conflict-disclosure and exclusivity expectations in writing, especially when your advisor also advises other portfolio companies.

Use employee tenders as a liquidity valve in a closed IPO window

With IPOs delayed, ElevenLabs shows the playbook: lead with institutional crossover investors to provide employee liquidity at a step-up valuation.

"completed a $300 million employee tender offer at a $22 billion valuation, double its $11 billion valuation in February. The tender was led by Wellington and T. Rowe Price."

Takeaway: If you're a late-stage operator, plan for tenders as a retention tool rather than assuming a near-term IPO.

Prepare for regulatory diligence before it arrives

The FTC's inquiry targets documents and executive testimony on consumer risk, even where voluntary safety accords exist.

"plans to issue subpoena-like demands for documents and executive testimony about potential consumer risks, just as the industry is pledging to police itself under a voluntary White House safety accord."

Takeaway: If you build or deploy frontier or consumer-facing AI, document safety processes and risk assessments now, since voluntary commitments don't preclude enforcement.


6. Overlooked Insights

Public-market IPO returns have collapsed, which may be an earlier signal than the headline narrative

"U.S. companies that went public this year are now trading less than 1% above their IPO prices on average, down from 24% at the end of June."

A drop from +24% to under +1% in a single quarter suggests post-IPO performance is deteriorating quickly, which could further chill the window for private marks.

Amazon's data-capture approach has no consumer opt-out, a potential privacy and regulatory flashpoint

"Customers will not be able to see or delete the images or opt out, short of not ordering from Amazon."

Combined with the FTC's consumer-risk focus, this kind of ambient AI training-data collection from private property may become a litigation and regulatory target, and an opening for privacy-preserving alternatives.