Best Data Provider For Complete Historical Low-float Stock Data, Including Delisted Tickers And Extended Hours?

Hi everyone. I’ve been building a historical research dataset for low-float/small-cap stocks and I’m running into a much bigger data-coverage problem than I expected.

The general universe I’m looking at is:

• Stocks with fewer than roughly 20M shares in float.
• Lower-market-cap names.
• At least 1M shares of daily volume.
• Dates from 2023 through 2025.
• Premarket, regular-session, and ideally postmarket one-minute data.
• Delisted, renamed, merged, and inactive tickers included.

So far, I’ve been able to build a fairly large list of eligible ticker-days and recover point-in-time information for a lot of old and delisted names. I’ve tested data from Alpaca SIP and Massive/Polygon, but the historical minute data is much more incomplete than I expected.

The issue is not just whether a ticker exists in the database. I need enough complete data to reconstruct:

• Premarket high and volume.
• Regular-session volume and VWAP.
• Intraday highs and lows.
• Technical setups around PMH breaks and pullbacks.
• Whether a theoretical entry reached a stop or 2R target first.
• Missing trading minutes versus minutes where no trades actually occurred.

In my broad initial universe, only around 1,600 out of roughly 293,000 eligible ticker-days had completely stored paths under the original strict definition. I’ve improved coverage using multiple providers, but there are still thousands of older, renamed, or archive-risk ticker-days where the data is partial or completely missing.

I’ve also found that mixing feeds can create problems. For example, bars from a direct-feed composite may look valid on their own but have very different volume from consolidated SIP data because off-exchange and FINRA TRF prints may be missing. Since my study uses volume, VWAP, liquidity, and dollar-volume features, I can’t just patch missing days with a different feed unless the data is reasonably comparable.

Has anyone built something similar and found a provider that offers:

• Historical one-minute bars going back several years.
• Full premarket and postmarket coverage.
• Raw/unadjusted price and volume.
• Consolidated SIP or TAQ-derived data, including FINRA/TRF volume.
• Delisted and inactive securities.
• Historical ticker changes and point-in-time symbol mapping.
• A flat-file or bulk-download option so I can save the data once and reuse it.
• Coverage where an empty response can be distinguished from a genuinely inactive/no-trade day.

I’m open to a paid provider or one-time historical download. I’ve looked at Alpaca, Massive/Polygon, Databento, and IBKR so far. Databento’s direct feeds are interesting, but the feeds visible to my account were missing several exchanges/reporting facilities, so the volume was not comparable to SIP. IBKR also appears to filter some historical trades and may not retain data for securities that are no longer trading.

I’d especially appreciate feedback from anyone who has used AlgoSeek, Tick Data, NYSE TAQ, LSEG Tick History, or another provider for old low-float and delisted stocks.

I’m not looking for real-time alerts or a trading platform. I’m trying to build one clean historical research dataset that I can download, validate, and keep locally.