Blog

Anthropic’s unreleased Claude Mythos Preview model has triggered urgent cybersecurity reviews across India’s government, banking, and IT sectors after the AI system autonomously uncovered thousands of previously unknown vulnerabilities in critical open-source infrastructure. The discovery has forced India to confront a stark reality: the same AI capabilities that could strengthen national defense could also empower adversaries to breach India’s digital backbone within hours.

What Mythos Found

Announced on April 7, 2026, Mythos Preview demonstrated unprecedented autonomous vulnerability discovery, identifying:

Vulnerability	Age	System	Impact
TCP SACK denial-of-service bug	27 years	OpenBSD	Remote service disruption
H.264 codec flaw	16 years	FFmpeg	Remote code execution
CVE-2026-4747	17 years	FreeBSD NFS server	Full system compromise
Linux kernel privilege escalation	Unknown	Linux kernel	Chained exploit weaponization

Critically, Mythos didn’t just find these bugs—it wrote complete working exploits without human guidance, including chaining multiple vulnerabilities together. Anthropic’s red team confirmed the model can exploit zero-day vulnerabilities across every major operating system and reverse-engineer exploits for closed-source software.

India’s Structural Exposure

India’s digital infrastructure faces disproportionate risk because its core systems run on the same open-source software Mythos is scanning:

• UPI transactions powering India’s $1 trillion digital payments ecosystem
• Aadhaar verifications at ration shops serving 1.4 billion citizens
• IRCTC tatkal bookings handling millions of railway reservations daily

“As the Siasat Daily noted, every critical Indian platform runs on layers of the same open-source software that Mythos is now scanning”. Legacy public platforms like Aadhaar and GST face “elevated exposure, particularly where uneven security standards persist”.

Banks and Industry Sound the Alarm

HDFC Bank, India’s largest private-sector lender, confirmed direct engagement with Anthropic: “We are engaged with the Data Security Council of India to evaluate risks and impact. We can confirm being in touch with the Anthropic team”. The Data Security Council of India has begun industry-wide consultations, while SaaS and deep-tech firms assess heightened risks extending into SCADA and IoT infrastructure.

The urgency isn’t limited to India. U.S. Treasury Secretary Scott Bessent convened emergency meetings with CEOs of Bank of America, Citi, and Wells Fargo. Government officials in the United States, Canada, and Britain have all met with banking leaders to discuss Mythos-induced threats.

Global Response: Project Glasswing

Anthropic launched Project Glasswing on April 7, assembling Amazon Web Services, Apple, Google, Microsoft, NVIDIA, Cisco, CrowdStrike, JPMorganChase, the Linux Foundation, and Palo Alto Networks to “secure the world’s most critical software”. The initiative includes:

• $100 million in usage credits for Mythos Preview
• $4 million in donations to open-source security bodies
• Defensive AI for vulnerability detection and penetration testing

No Indian firm has been included in the initial consortium despite India being one of Anthropic’s largest markets.

What This Means for India’s IT Sector

Analysts at Motilal Oswal warn that while disruption to India’s $260 billion IT services sector may not be immediate, the pressure to evolve is intensifying as AI compresses vulnerability discovery from months to hours. Kotak Institutional Equities flagged disruption risks to the sector’s growth trajectory.

The situation creates a paradox: India depends on American-built AI models for national security audits while facing exclusion from the very coalition designed to address these risks. Experts are calling for India to:

1. Formally seek inclusion in Project Glasswing
2. Accelerate indigenous AI-driven cybersecurity tool development
3. Implement Zero Trust architecture immediately

As Newton Cheng, Anthropic’s Frontier Red Team Cyber Lead, warned: “We do not plan to make Claude Mythos Preview generally available… However, given the rate of AI progress, it will not be long before such capabilities proliferate, potentially beyond actors who are committed to deploying them safely”.

For India, the message is clear: the window to prepare is closing rapidly. Without indigenous AI security capabilities and formal participation in global cybersecurity initiatives, the country’s digital sovereignty faces unprecedented risk from AI-driven cyber threats.

Washington, April 15, 2026 – In a stark escalation of the U.S.-Iran standoff, President Donald Trump issued a dire warning to Tehran on Tuesday, stating that failure to secure a nuclear deal before the April 21 ceasefire expires would bring “severe consequences.” Speaking amid a tightening U.S. naval blockade of Iranian ports, Trump emphasized that Iran would never possess nuclear weapons, declaring, “Iran will not have a nuclear weapon, and we are going to get the dust back; either we will get it back from them, or we will take it.”

The ominous rhetoric follows the dramatic collapse of peace talks in Islamabad over the weekend. Led by U.S. Vice President JD Vance, the marathon negotiations foundered on Iran’s refusal to abandon its nuclear ambitions. Iranian officials decried U.S. “overreach,” while Vance called the impasse “bad news” for Tehran. With no agreement in sight, Trump activated the naval blockade on Monday, halting all maritime traffic to Iran’s coastline. U.S. Central Command reported zero successful transits in the first 24 hours, as over 10,000 American personnel aboard a dozen warships enforced the measures. Iran branded the action “piracy” and promised retaliation in its territorial waters.

This blockade marks a pivotal shift in the year-long conflict, which intensified after alleged Israeli strikes on Tehran and mutual accusations of nuclear escalation. Trump’s administration has framed the operation as essential to prevent Iran from enriching uranium to weapons-grade levels, citing intelligence showing Tehran’s program advancing rapidly despite international sanctions.

Compounding the pressure, Mossad Director David Barnea delivered a provocative message at a Holocaust Remembrance Day ceremony in Israel. “Our commitment will only be complete once this extremist regime is replaced,” Barnea asserted, signaling Israel’s unwavering resolve. He described the campaign against Iran as multifaceted, estimating regime change could take about a year from the war’s outset—a timeline echoed in reports from The Jerusalem Post. Barnea’s comments underscore Israel’s doctrine of preemptive action, blending airstrikes with covert operations to dismantle Iran’s nuclear infrastructure and proxy networks.

Amid the brinkmanship, glimmers of diplomacy persist. U.S. officials are considering another round of talks before the ceasefire lapses, with Turkey mediating to narrow gaps. Yet optimism is tempered; Iranian state media dismissed further negotiations as futile under the blockade, while hardliners in Tehran rally support for defiance.

The stakes could not be higher. Economically, the blockade has spiked global oil prices by 15% overnight, disrupting 20% of the world’s supply through the Strait of Hormuz. Militarily, U.S. forces stand ready for contingencies, with B-52 bombers repositioned in Diego Garcia. Analysts warn of a potential wider war, drawing in Hezbollah, Houthis, and Gulf allies.

Trump’s base cheers the hardline stance, viewing it as fulfilling campaign pledges to curb Iran’s influence. Critics, however, decry the risks of miscalculation, likening the scenario to the 1979 tanker war. As the April 21 deadline looms, the world watches whether Tehran blinks—or ignites a regional inferno.