Newsroom
News, media kit and spokespeople
Covering LLM costs, AI FinOps or agent reliability? Find our releases, logos, approved boilerplate and the people who can speak to the numbers.
About PALM SOUQ · approved 14 Oct 2026
PALM SOUQ PRIVATE LIMITED is the per-tenant cost governance and circuit breaker gateway for AI SaaS companies. Sitting in the request path between applications and model providers such as OpenAI, Anthropic, Google Gemini and AWS Bedrock, PALM SOUQ attributes every token and prompt cache hit to the customer that caused it, enforces budgets and model fallbacks per tier, and stops runaway agent loops in under 50 ms while adding less than 5 ms of latency. Incorporated in 2026 and headquartered in Chennai, PALM SOUQ PRIVATE LIMITED (CIN U46309TN2026PTC193667) is a private limited company incorporated in India and operates on 280+ edge points of presence. Learn more at palmsouqltd.com.
Founded
2026
Team
34 people
Edge PoPs
280+
Press releases
Announcements
Open a release to read it in full. All releases are datelined Chennai, India.
ProductPALM SOUQ launches Prompt Cache Tracker to meter cache reads and writes per tenantMargin reports now show what prompt caching saved, and cost, for every customer.
CHENNAI, India. PALM SOUQ today made its Prompt Cache Tracker generally available on the Growth and Scale plans. The tracker meters cache writes and cache reads separately for every tenant, using each provider's own cache pricing from the PALM SOUQ pricing registry.
Prompt caching can reduce input costs sharply when a long system prompt or document is reused, but it changes the arithmetic of a request: a cache write is priced differently from a cache read, and the savings land on whichever tenant reuses the prefix. Without per-tenant accounting, finance teams either overstate or understate the gross margin of their heaviest customers.
"Caching made our customers' token bills cheaper and their margin reports wrong," said Ananya Iyer, Head of FinOps Research at PALM SOUQ. "The tracker puts the cache line items back where they belong, on the tenant that caused them."
Media contact: press@palmsouqltd.com
CompliancePALM SOUQ adds HIPAA Business Associate Agreements and India data residencyHealthcare AI companies can route PHI through the gateway under Zero Payload Retention.
CHENNAI, India. PALM SOUQ now signs HIPAA Business Associate Agreements with customers on its Growth and Scale plans, and lets those customers pin request metadata to the United States, the European Union or India.
Because PALM SOUQ streams prompts and completions through memory without writing them to disk, the data the gateway stores is limited to token counts, cost, latency and the identifiers customers choose to send. The BAA formalizes that design for covered entities and their business associates.
"Healthcare and legal AI teams asked us the same question in every security review: what do you keep?" said Kabir Sethi, Head of Security and Compliance. "The honest answer is metadata only, and now it is in a contract."
Media contact: press@palmsouqltd.com
ResearchPALM SOUQ publishes the LLM Price Index, a free sourced reference for token and cache pricesInput, output, cached input and cache write prices for 30+ models, with sources.
CHENNAI, India. PALM SOUQ today released the LLM Price Index, a public reference of list prices for more than 30 foundation models across OpenAI, Anthropic, Google, DeepSeek, Mistral, Amazon and open-weight hosts. Each entry records input, output, cached input and cache write prices per million tokens, along with the source and the date it was last reviewed.
The index is the same pricing registry that powers cost attribution inside the PALM SOUQ gateway, maintained by the company's FinOps Research team. It is free to use under a Creative Commons Attribution license.
Media contact: press@palmsouqltd.com
NetworkPALM SOUQ gateway goes global on 280+ edge locationsUnder 5 ms of added latency and a 99.999% availability target, worldwide.
CHENNAI, India. PALM SOUQ has moved its gateway data plane onto 280+ edge points of presence, so per-tenant metering, quotas and circuit breakers run close to customers' applications wherever they are deployed.
Regional control planes in US East, US West, EU Central, India South and Singapore keep tenant budgets consistent across locations. Customer SDKs continue to fail open: if a gateway node cannot be reached, requests go directly to the model provider.
"A cost governor that adds noticeable latency will be removed by the next engineer who profiles the app," said Meera Krishnamurthy, CTO. "We treat the 5 ms budget as part of the product."
Media contact: press@palmsouqltd.com
CompliancePALM SOUQ completes SOC 2 Type II examinationIndependent report covers security, availability and confidentiality controls.
CHENNAI, India. PALM SOUQ today announced that it has received its SOC 2 Type II report, covering the Security, Availability and Confidentiality trust services criteria. The report is available to customers and prospects under NDA through the PALM SOUQ Trust Center.
Media contact: press@palmsouqltd.com
ProductPALM SOUQ ships native usage sync with Stripe, Metronome, Orb and LagoMetered token usage becomes billable overage events without a new data pipeline.
CHENNAI, India. PALM SOUQ now streams per-tenant usage to Stripe Billing, Metronome, Orb and Lago, so AI SaaS companies can launch usage-based token overages on top of flat subscriptions using the billing system they already run.
"Most teams we meet know which customers are unprofitable but cannot bill for it," said Vineet Saresar, founder and CEO. "Billing sync closes that loop in an afternoon."
Media contact: press@palmsouqltd.com
Media kit
Logos and brand colors
Use the mark on dark backgrounds where possible. Do not recolor, stretch or add effects to it.
Canvas
#010102
Lavender (accent)
#5e6ad2
Ink
#f7f8f8
Surface
#0f1011
Always write the name as PALM SOUQ, one word, capital T and capital G. The legal entity is PALM SOUQ PRIVATE LIMITED
Spokespeople
Who can speak to what
Interviews are available in English and Hindi, in person in Chennai or remotely.
| Name | Role | Topics |
|---|---|---|
| Vineet Saresar | Founder & CEO | AI SaaS unit economics, flat-rate vs usage pricing, the Indian infrastructure startup ecosystem |
| Meera Krishnamurthy | CTO | Edge proxies, agent loop detection, latency budgets, rate limiting at scale |
| Ananya Iyer | Head of FinOps Research | LLM price changes, prompt caching economics, gross margin per customer |
| Kabir Sethi | Head of Security & Compliance | Zero Payload Retention, SOC 2, HIPAA and India DPDP Act for AI infrastructure |
Press contact
We reply to media requests within one business day. For data requests on LLM pricing, the LLM Price Index is free to cite with attribution.
Don't wait for a surprise $10,000 OpenAI invoice.
Route your agent traffic through PALM SOUQ in 2 minutes. Change one base URL, pass a tenant header, and every token is attributed and governed.
14-day free production trial · 100k requests included · no card required