Timescale

Timescale, Inc., now doing business as Tiger Data, was founded in 2015 by Ajay Kulkarni and Michael J. Freedman. The pair initially built an Internet-of-Things platform called iobeam/Sensoriant. Sensors they did not control produced continuously growing machine data, and available databases forced a choice between reliable relational transactions and time-series ingest/scale. Instead of maintaining separate operational and time-series systems, they extended PostgreSQL. That internal solution became TimescaleDB, announced publicly in 2017: a C-language PostgreSQL extension that preserves SQL, joins, transactions, tooling and the Postgres ecosystem while adding automatic time/space partitioning through hypertables, time-bucketing, retention policies, compression/columnar storage and continuously refreshed aggregates. Kulkarni is co-founder/CEO, an MIT-trained computer scientist and MIT Sloan MBA whose prior startup was acquired by GroupMe/Skype/Microsoft; he also worked at Microsoft, Citigroup and startups. Freedman is co-founder/CTO and the Robert E. Kahn Professor of Computer Science at Princeton, with MIT S.B./M.Eng. degrees and an NYU Courant Ph.D.; his distributed-systems work includes CoralCDN and software-defined-networking research, and he previously co-founded Illuminics Systems. The company first appears in a 2015 Form D for approximately $3.66 million. In January 2018 it announced that cumulative funding had reached $16 million: a $12.4 million Series A led by Benchmark, with NEA, Two Sigma Ventures and angels, on top of the earlier seed. Peter Fenton of Benchmark joined the board. Icon Ventures led a $15 million Series A-1 in January 2019, with Benchmark, NEA and Two Sigma participating; cumulative funding was $31 million. Redpoint Ventures led a $40 million Series B in May 2021 with all existing institutional investors; Satish Dharmaraj joined the board and the company reported $70 million+ cumulative funding. Tiger Global led a $110 million Series C announced February 22, 2022, joined by Benchmark, NEA, Redpoint, Icon and Two Sigma. The round valued the company at more than $1 billion and brought public funding to approximately $180-$181 million. The current company page rounds this to $180 million. Raw legacy database rows omit the 2018/2019 rounds and sum just three SEC filings to $153.1 million; they are not a complete funding history. The company shifted from an open-source extension plus support/enterprise software to a cloud-first business, while retaining self-managed editions. Tiger Cloud is a fully managed, unforked PostgreSQL service on AWS that combines ordinary relational/transactional data with time-series, events, real-time analytics and AI/search. Its differentiators include hypertables, continuous aggregates, Hypercore hybrid row/columnar storage, up to 95% claimed compression, hot/cold tiering, compute-local caching, read replicas, high availability, PostgreSQL extensions and deep observability. Customers pay by hourly compute and average storage consumption, with Performance, Scale and custom Enterprise tiers; current starting cloud prices are approximately $30/month for Performance and $36/month for Scale, before usage. Enterprise adds negotiated capacity/support/security such as HIPAA, SAML SSO and cross-region controls. A 30-day trial/$1,000 credits program and open-source downloads support product-led adoption. TimescaleDB remains the core open-source/source-available extension. Apache-2.0 covers the open-source portion, while additional Community functionality is governed by the Timescale License, which permits broad internal/commercial self-hosted use but restricts offering the licensed interfaces as a competing database-as-a-service. That mixed license introduced sustainable-cloud economics but also creates user confusion, non-OSI/source-available criticism, compatibility differences between third-party managed Postgres services and Tiger Cloud, and vendor-lock-in risk around TSL-only capabilities such as some compression/hyperfunctions. The company expanded beyond time series into general operational Postgres and AI. Its Postgres Search Stack includes vector search with HNSW and Streaming DiskANN, BM25 keyword search, hybrid retrieval and SQL-native embedding/vectorization pipelines. It positions this as structured memory and retrieval for AI agents without a separate vector database/search cluster. In April 2024 Timescale acquired collaborative SQL editor PopSQL for an undisclosed price, bringing its team and browser/desktop query, dashboard and collaboration technology in-house. PopSQL later moved into limited support; a planned September 1, 2026 shutdown was postponed in August 2026, with no new features and only critical security/emergency fixes. That sequence creates product-sunsetting and customer-portability concerns. On June 17, 2025 Timescale rebranded the company and managed service as Tiger Data/Tiger Cloud while retaining the TimescaleDB product name. The rebrand reflected a move from a time-series specialist toward a broader high-performance PostgreSQL platform for transactional, analytical and agentic workloads, but sacrificed established Timescale brand recognition and risks confusion with TigerGraph, TigerBeetle and other similarly named database products. The 2025 announcement reported more than 2,000 customers across 25 countries, more than 3 million active databases, mid-eight-figure ARR and greater than 100% year-over-year growth. Those are company claims, not audited financial statements. It named customers including Lucid Motors, Hugging Face, Mistral, Financial Times, Barclays, European Space Agency and Schneider Electric; prior releases listed Akamai, Cisco WebEx, Comcast, DigitalOcean, GE, IBM, Microsoft, Pfizer, Samsung, Uber and Walmart among users/community organizations. In 2022 it reported more than 500 paying customers, 7x community growth and 20x revenue growth over two years. In 2026 the strategy re-emphasized the physical world and machine data. Tiger Data launched TimescaleDB Enterprise, a commercially licensed self-managed product for on-premises, edge and private-cloud environments in data centers, energy/utilities, oilfield services and smart manufacturing, while Tiger Cloud centralizes operational analytics. It also developed data-lake integration through Iceberg/S3, broader streaming ingestion and an operational data layer for AI data centers. Cloud security claims include SOC 2 Type II, GDPR/CCPA programs, HIPAA for Enterprise, TLS 1.2+, mTLS internally, encrypted volumes/backups with per-service AWS KMS keys, PITR, MFA, VPC peering/PrivateLink, Transit Gateway, IP allowlists, external penetration testing, 24/7 operations and a 99.9% uptime SLA. These controls do not eliminate the consequences of a database outage, corruption, failed backup/restore, cross-tenant flaw, compromised credentials/extensions, supply-chain bug or operator error. Machine, financial, industrial, crypto and AI-agent data can be mission-critical and sensitive; customers must verify region, subprocessor, retention, key-management, support-access, RPO/RTO and contractual SLA terms. Self-hosters bear upgrade, replication, capacity, vacuum, indexing, compression/tiering and recovery complexity, and database extensions can lag new PostgreSQL versions. Hybrid row/columnar behavior and continuous aggregates require workload-aware schema/query design; reported performance/compression benchmarks are vendor-run and may not generalize. Competition spans time-series databases (InfluxDB, QuestDB, VictoriaMetrics, ClickHouse, Amazon Timestream, kdb+), operational/analytical Postgres and distributed SQL (vanilla PostgreSQL, Aurora, AlloyDB, Neon, Supabase, Crunchy Data, Citus, CockroachDB), analytics warehouses (Snowflake/Databricks) and AI search/vector systems (Pinecone, Weaviate, Elasticsearch/OpenSearch, pgvector providers). Tiger Data's strongest moat is remaining inside familiar PostgreSQL while serving relational and high-ingest time-aware data together. Its challenges are hyperscaler bundling, open-source alternatives, rapidly changing AI/search demand, cloud cost competition, complex licensing, execution across cloud and enterprise self-managed products, and a last priced funding round from the 2021-2022 peak. No material company-specific litigation, regulator enforcement or confirmed breach was found in manual review. Priority diligence gaps include current audited ARR/growth/NRR/margin and burn/runway, post-2025 headcount, current cloud/customer concentration, renewal versus consumption expansion, detailed outage history/SLA credits, independent performance/security validation, TimescaleDB Enterprise adoption, PopSQL migration obligations, license-change risk and any secondary financing/valuation changes since 2022.

Founding story

Kulkarni and Freedman were building the iobeam IoT platform when continuously arriving sensor data exposed a gap between relational correctness and time-series scale. They chose extensible PostgreSQL, built the missing time/partition/aggregate layer and spun the database into TimescaleDB/Tiger Data.

Business model

Open-core/source-available database company monetizing a consumption-priced managed PostgreSQL cloud, custom enterprise cloud contracts, commercially licensed self-managed TimescaleDB Enterprise and support/services.

Hourly compute plus average GB storage/tiered storage and enterprise support/security commitments in Tiger Cloud; license/subscription/support contracts for TimescaleDB Enterprise; free Apache/TSL community software drives adoption.

Traction

2021: 2m+ active databases, 150 employees and $70m+ funding. 2022: 500+ paying customers, 7x community and 20x revenue growth over two years. Jun 2025: 2,000+ customers in 25 countries, 3m active databases, mid-eight-figure ARR and >100% YoY growth. All are company-reported; current audited figures undisclosed.

Latest developments

In 2026 Tiger Data launched TimescaleDB Enterprise for self-managed on-prem/edge/private-cloud machine data, advanced the Postgres search/data-lake stack and released TimescaleDB 2.29.x. PopSQL's planned September shutdown was postponed but the acquired product remains limited-support with no feature roadmap.