Akamai vs Microsoft Azure

a16z Leads Series A in Private Network Startup doxxnet

Andreessen Horowitz (a16z) announced it is leading a Series A investment in doxxnet, a startup building a private, parallel internet infrastructure for individuals and AI agents. doxxnet offers a software-defined global network with non-Internet domains, encrypted peer-to-peer communication, and DNS-level threat protection, all without requiring personal information like email or phone. The company is founded by Barrett Lyon, known for Prolexic (acquired by Akamai), BitGravity, and Defense.net. The funding will support the development of its 'Agentic Defined Networking' concept, which allows users and their agents to configure and control the network. The announcement highlights growing privacy needs driven by AI agents that operate with access to user data.

• a16z is leading a Series A investment in doxxnet.
• doxxnet provides a private, parallel internet infrastructure for users and AI agents.
• doxxnet was founded by Barrett Lyon, who previously founded Prolexic, BitGravity, and Defense.net.

Akamai Announces $11.6 Billion Multi-year Agreement with Anthropic to Support Growing Demand

Akamai today announced a significantly expanded relationship with Anthropic for $11.6 billion of contractual commitment over seven years.

AI companies adopt risky 'take-or-pay' contracts for compute

The AI industry is increasingly relying on take-or-pay contracts to secure computing power, a model borrowed from the energy sector. These agreements require AI companies to pay for GPUs, server time, and other inputs regardless of usage, creating significant financial risks. The payments are backed by expected future revenue, but there are concerns about whether AI revenues will materialize in time to meet payment schedules. Off-balance-sheet commitments are growing, with estimates ranging from $1.65 trillion to $3 trillion. Major players like Anthropic and OpenAI have committed billions, while companies like CoreWeave and Alphabet rely heavily on such contracts. Financial analysts and regulators warn that the entire system could collapse if companies fail to secure enough capital. The Bank of International Settlements has highlighted potential systemic risks, particularly through special purpose vehicles. The article notes that major banks and investors are closely monitoring this trend, with experts predicting a 'wave of disputes' as the sector matures.

• Take-or-pay contracts obligate AI companies to pay for compute regardless of usage, similar to energy sector agreements.
• Anthropic committed $750 billion for compute through 2030; OpenAI could spend over $500 billion in the next 10 years.
• Off-balance-sheet commitments in tech sector estimated at $1.65 trillion to $3 trillion.

Zendesk Names Rachita Sundar as CFO

Zendesk announced the appointment of Rachita Sundar as Chief Financial Officer, effective September 28, 2026. Sundar, who brings over 20 years of experience in finance and operations at technology companies including Microsoft Azure, HubSpot, and Qualtrics, succeeds Julie Swinney, who transitions to Chief Administrative Officer and executive sponsor of Zendesk's Employee Service business. At Qualtrics, Sundar helped finance and close the $6.75 billion acquisition of Press Ganey Forsta. CEO Tom Eggemeier highlighted her financial rigor and operational expertise as Zendesk scales toward a multi-billion-dollar AI business.

• Rachita Sundar appointed CFO of Zendesk, effective September 28, 2026.
• Sundar has over 20 years of experience at Microsoft Azure, HubSpot, and Qualtrics.
• At Qualtrics, Sundar helped finance and close the $6.75 billion acquisition of Press Ganey Forsta.

Anthropic launches cheaper AI model Sonnet 5.5

Anthropic has released Claude Sonnet 5.5, the second model in the Claude 5.5 series, positioned as a complementary workhorse to the more powerful Opus 5.5 for well-defined daily tasks like coding, bug fixing, and document creation. Priced unchanged at $2 per million input tokens and $10 per million output tokens, it runs over 30% faster than Sonnet 5 and costs up to 30% less per task due to reduced token usage. While not advancing frontier capabilities, it shows near-Opus performance on several benchmarks, including GDPval-AA v2.1 and OSWorld 2.1, and outperforms Opus 5.5 on agentic coding tasks, achieving 70.6% on Terminal-Bench 4.0 versus 10.3% for Sonnet 5. Due to significant cyber capabilities, it is the first Sonnet model subject to the same cybersecurity safeguards as Opus 5. Available via Claude, AWS, Google Cloud, and Microsoft Azure, an updated Haiku model is expected in the coming weeks.

• Anthropic released Claude Sonnet 5.5 on September 28, 2026, priced at $2 per million input tokens and $10 per million output tokens.
• Sonnet 5.5 is over 30% faster than Sonnet 5, and cost per task can drop by up to 30% due to reduced token usage.
• The model outperforms Opus 5.5 on agentic coding benchmarks and reaches 70.6% on Terminal-Bench 4.0, versus 10.3% for Sonnet 5.

Anthropic Launches Claude Opus 5.5 Despite Slowdown Call

Anthropic released Claude Opus 5.5, its flagship AI model and the first in the 5.5 family, on September 22, 2026. It achieves the highest-ever score (58) on the Artificial Analysis Intelligence Index and leads six out of ten benchmarks, including outperforming OpenAI's GPT-6 Astra on Terminal-Bench 4.0. Priced at $4 per million input tokens and $20 per million output tokens (20% lower), it promises a 40% cost reduction for typical workloads, though savings depend on config. It is 30% faster, now defaults in Claude Code and the Claude app, and matches Claude Fable 5.1 on most tasks. Early tests show efficiency gains but also more concurrency issues and security test failures. Anthropic cites external safety evaluations, and OpenAI responded with GPT-6 Sol and Luna, priced ~50% lower and offering up to 90% cache discounts.

• Claude Opus 5.5 scores 58 on the Artificial Analysis Intelligence Index, the highest recorded, and leads six out of ten benchmarks.
• Pricing reduced to $4 per million input tokens, $20 per million output tokens, and cache reads at $0.20 per million tokens, promising a 40% cost reduction per task (though configuration may affect savings).
• Opus 5.5 is 30% faster, supports multi-agent scaling up to 100 parallel agents, and is now default in Claude Code and the Claude app.