Daily Funding Roundup:
Apr 10, 2026
Friday closed a wild week. TSMC posted blockbuster Q1 earnings (+35% YoY revenue) showing that AI chip demand survived the Iran war. OpenAI launched a $100/month ChatGPT Pro tier targeting developers. Shanghai biotech Oricell raised $110M pre-IPO for CAR-T liver cancer therapy. A Marimo zero-day was weaponized within 9 hours of disclosure. The S&P closed the week up 3.6%, the biggest weekly gain since November.
Rounds
Shanghai-based clinical-stage biotech Oricell Therapeutics closed a $110M pre-IPO round to advance its lead asset Ori-C101, a CAR-T cell therapy targeting hepatocellular carcinoma (liver cancer). If approved, Ori-C101 would be the first globally-approved CAR-T therapy for solid tumors, a category where CAR-T has historically failed due to tumor microenvironment challenges. The round follows a $70M Series C in January, signaling that institutional investors believe Oricell is approaching a regulatory inflection point. The company is targeting a Hong Kong or U.S. listing as the next step. Total raised: $270M.
News & Signals
TSMC posts blockbuster Q1: $35.6B revenue, +35% YoY on AI chip demand
TSMC reported Q1 2026 revenue of NT$1.13 trillion ($35.6 billion), a 35% year-over-year increase that beat consensus estimates. March alone was up 45.2% YoY and 30.7% from February, the strongest monthly reading of the quarter. The Iran war did not dent AI chip demand: TSMC's advanced node business (3nm and below) remained the growth driver as Nvidia, Apple, and AMD continue to consume capacity. CEO C.C. Wei reaffirmed full-year guidance for 30% USD revenue growth, with AI representing the entirety of the growth above prior expectations. TSMC shares surged past NT$2,000 on the report, validating the broader AI infrastructure thesis: the demand signal at the foundry level remains unequivocal even as macro uncertainty whipsaws public markets.
OpenAI launches $100/month ChatGPT Pro tier to challenge Anthropic's Claude Code
OpenAI announced a new $100/month ChatGPT Pro subscription tier (sitting between the $20 Plus and $200 Pro plans) specifically targeting developers who hit Codex usage limits. The new tier offers 5x more Codex usage than Plus (10x through May 31 as a promotional offer) and unlimited access to Instant and Thinking models. This is a direct competitive response to Anthropic's Claude $100/month plan, which has dominated the developer subscription market for over a year. The strategic implication: OpenAI is segmenting its consumer product to capture the developer wallet that Anthropic has been monopolizing. With both companies generating $25B+ in annualized revenue, the subscription pricing wars are now happening at meaningful scale.
Markets close shaky week with biggest gain since November: S&P +3.6%
Despite Thursday's wobble (oil swung between $97 and $100, ceasefire doubts emerged), the S&P 500 finished the week up 3.6%, its biggest weekly gain since late November 2025. The Dow's 1,325-point single-day surge on Wednesday was the largest one-day percentage gain since April 2025. Friday's session was muted as investors took profits ahead of the weekend, with stocks wavering and oil holding steady. The ceasefire is technically still in effect but Iran has periodically restricted Strait of Hormuz traffic in response to Israeli strikes on Lebanon, complicating the recovery. The week's net narrative: relief rally was real, the path forward remains uncertain, and the IPO window has reopened a crack but is not fully open.
Marimo zero-day exploited within 9 hours of disclosure: AI infrastructure security wakeup call
Attackers began exploiting CVE-2026-39987, a critical unauthenticated remote code execution vulnerability in the open-source Python notebook tool Marimo, roughly nine hours after the bug was publicly disclosed. The speed of weaponization (9 hours from disclosure to active exploitation) is consistent with the new pattern of AI-assisted vulnerability discovery and exploitation. Marimo is widely used in machine learning workflows and data science teams, meaning the affected attack surface includes thousands of enterprise data infrastructure deployments. The incident is being interpreted as the practical validation of Anthropic's earlier Mythos warning: AI tools are now finding and exploiting vulnerabilities faster than defenders can patch. AI infrastructure startups (Sentry, Snyk, Wiz, Shield AI for cyber) are seeing increased inbound interest as a result.
VC Mood on X
Friday's mood was contemplative after a week that swung from war-deadline dread to ceasefire euphoria to Friday cautious optimism. TSMC's earnings provided the cleanest validation that the underlying AI demand signal is intact: 35% YoY growth at a $35B revenue base is not the kind of number you can fake or pull forward. "TSMC is the unstoppable indicator," one growth investor posted. "If TSMC keeps printing 30%+ growth, every AI infrastructure thesis is validated regardless of what happens in geopolitics." Several VCs treated the TSMC report as the final word on the question that had haunted markets all week: is the AI capex cycle still on track? Answer: yes, decisively.
OpenAI's $100/month tier provoked the most strategic discussion. The framing as a 'response to Anthropic's Claude' suggests OpenAI sees Anthropic as the primary competitive threat in the developer segment, not Google or Meta. Several enterprise software investors noted that the developer subscription market is consolidating into a duopoly: OpenAI vs Anthropic for code-focused developers, with Google Gemini largely sidelined despite its model quality. The implications cascade to coding tools (Cursor, Windsurf, Codex) which build on top of these foundation models: their margins depend on which side of the duopoly they pick.
The Marimo zero-day was the most discussed security story. "Nine hours from disclosure to weaponization. We are now in the era where defenders need AI agents running 24/7 just to keep up," one cybersecurity investor posted. Several VCs noted that this validates the Anthropic Mythos warning from late March (which had been dismissed by some as overblown). The practical implication for portfolio companies: every startup with public-facing infrastructure now needs continuous vulnerability scanning, and the incident response window has shrunk from days to hours. Cybersecurity startups that can demonstrate AI-powered defense (Sentry, Wiz, Snyk, depthfirst) are seeing increased inbound interest as a result.
Rounds and signals sourced from SEC filings, press releases, and verified news reports. All amounts in USD unless noted. Reporting reflects information available at time of publication.